The UK’s proposed 87% emissions cut by 2040 sounds like a climate target. It is better understood as a physical-economy target.

DESNZ’s provisional 2025 emissions statistics put UK territorial greenhouse gas emissions at about 367 MtCO2e in 2025, around 54% below 1990. The Climate Change Committee’s Seventh Carbon Budget advice recommends a five-year cap of 535 MtCO2e over 2038-42, including international aviation and shipping. That is roughly 107 MtCO2e a year and 87% below 1990 levels.

So the gap from here is not marginal. It is not another round of symbolic tightening. It means cutting roughly another 260 MtCO2e a year from today’s economy.

That implies a different physical economy. Petrol and diesel stop being the default for road transport. Gas stops being the default for domestic heat. Electricity stops being merely one sector and becomes the delivery system for the rest of the transition. Industry has to decarbonise through process change, not just lower output. Land use, agriculture and aviation become unavoidable rather than peripheral.

This is why the cost debate is often misleading. The CCC’s supplementary analysis presents the transition as having a relatively small net additional cost in macro terms: around £4 billion a year on average between 2025 and 2050, or about 0.2% of GDP. Carbon Brief’s analysis of the CCC pathway summarises the same modelling as around £108 billion of net cost out to 2050 and roughly £0.7 trillion of capital investment.

Those figures are useful, but they should not be treated as hard outcomes. They are modelled assumptions about technology costs, fuel savings, deployment rates, consumer adoption and policy delivery. They tell us the shape of the pathway, not a guaranteed invoice.

The better way to read them is this: low net cost, high system change. The UK would be swapping fossil-fuel operating expenditure for upfront capital investment. That can be a good economic trade, but only if the country can finance, permit, build, connect and maintain the assets on time.

Where the UK stands

The UK’s historical performance is real, but not fully transferable. Emissions have fallen sharply since 1990, yet the long-term reduction has been driven heavily by changes in electricity generation: coal decline, cleaner power, renewables, lower electricity demand and industrial restructuring.

The remaining emissions sit in harder-to-shift parts of the economy. DESNZ’s 2025 sector breakdown has domestic transport at 31% of UK net greenhouse gas emissions, buildings and product uses at 22%, agriculture at 13%, industry at 11%, electricity supply at 10%, fuel supply at 7%, waste at 6% and LULUCF close to zero.

That is the important shift. Power is still critical, but it is no longer the whole emissions story. Even eliminating the remaining electricity-sector emissions would not deliver the 2040 target by itself. The next phase is more distributed, more consumer-facing and more dependent on infrastructure that has to be built before the emissions savings appear.

The cost is not one bill. It is a change in the balance sheet.

A lot of arguments about net zero collapse because they treat cost as a single bill. That is the wrong accounting frame. The CCC’s estimate is a net additional cost against a counterfactual where the UK does not make further transition investments. In that comparison, low-carbon investment is partly offset by lower fossil-fuel use, lower vehicle running costs and lower heating costs later.

That does not make the transition cheap in any practical sense. It means the economic challenge is timing, financing and distribution. The investment comes first. The operating savings come later, unevenly and only if the assets are installed, connected and used. That is why the headline 0.2% of GDP number can be simultaneously plausible and politically insufficient.

The CCC/Carbon Brief estimates are useful as ballparks: around £0.7 trillion of capital investment over 2025-2050, net cost of around £108 billion over the period, average net additional cost around £4 billion a year, and public funding needs estimated by Carbon Brief at around £6 billion a year from 2030. But these are pathway assumptions, not settled facts. They depend on deployment, technology cost curves, policy design and whether Britain avoids paying more than necessary through delay, fragmentation and weak delivery.

That distinction matters. A low net cost is not the same as an easy transition. It means the system has to move capital from one set of assets to another. The politics is not only how much the UK spends. It is who pays first, who saves later, who bears the disruption, and whether the state can make the sequence credible enough for households and firms to act.

Transport: oil has to stop being the default

DESNZ’s provisional 2025 data show domestic transport emitted 112.9 MtCO2e, accounted for 31% of UK net emissions and increased by 2.3% from 2024, largely due to higher petrol and diesel use in road transport. That alone should temper any easy story that the UK is already on a smooth glide path.

The CCC pathway is not just a story about higher EV sales. It is a vehicle-stock story. Carbon Brief’s readout of the CCC pathway has around 80% of cars, 74% of vans and 63% of HGVs electric by 2040. Surface transport emissions fall by 86% from 2023 to 2040, with most of the savings coming from cars and vans.

The delta is therefore physical and logistical: public charging, home charging, depot charging, local distribution networks, HGV infrastructure, the used-car market, battery supply chains and fleet finance. The cost is not simply the public charger bill. It is the accelerated replacement of the road-transport capital stock.

The optimistic version is that EVs become cheaper to buy, much cheaper to run and increasingly normal in the used market. The harder version is that upfront costs, charging access, local grid constraints and fleet economics become bottlenecks. The emissions pathway assumes the first story wins fast enough. Policy has to make that assumption real.

Buildings: gas has to stop being the default

DESNZ’s 2025 figures put buildings and product uses at 80.4 MtCO2e, around 22% of UK emissions. DESNZ notes that the sector is primarily fuel combustion in buildings, largely gas and other fuels for heating and cooking. Buildings emissions have only fallen 27% since 1990. That is the contrast: the power sector collapsed coal; the building stock still contains millions of gas boilers.

Carbon Brief’s summary of the CCC pathway has existing homes with low-carbon heating rising from 8% in 2023 to 68% by 2040, while homes with heat pumps rise from around 1% to 52%. Heat-pump installations rise from around 60,000 in 2023 to nearly 450,000 in 2030 and 1.5 million by 2035.

This is where the transition becomes most household-facing. The cost is not just the heat pump. It is installation, hot water tanks, pipework, radiator upgrades where needed, insulation, installer capacity, landlord compliance and the financing model. The CCC supplementary analysis assumes heat-pump costs fall by about one-third between 2025 and 2050, but the practical cost still varies enormously by property type, insulation level and ancillary works.

This is politically different from the power sector. Closing a coal plant is visible to the system but remote from the household. Replacing a boiler enters the kitchen, the mortgage, the landlord relationship and the energy bill. It can still be the right direction, but it is a very different kind of state capacity.

Power: no longer just a sector

DESNZ’s 2025 release has electricity supply at 37.4 MtCO2e, only 10% of UK emissions, with coal emissions in the sector falling to zero after the closure of the UK’s last coal-fired power station in September 2024. This is the UK’s biggest climate success story. But it also creates a misunderstanding: because power decarbonisation delivered so much of the past reduction, it is tempting to assume the next phase can work the same way.

It cannot. Electricity is no longer just a sector to clean up. It becomes the delivery system for transport, heat and parts of industry. Carbon Brief’s CCC pathway analysis has electricity demand more than doubling from 274 TWh in 2023 to 692 TWh in 2050, driven by electric vehicles, heat pumps, industry and hydrogen production. Renewables reach around 80% of generation by 2040.

That means power-sector cost is not just the cost of more turbines or panels. It is the cost of a larger grid, faster connections, more storage, more demand flexibility, more balancing capability, and a market design that can deliver investment without making the system unnecessarily expensive.

This is the enabling sector. If grids, connections, planning and flexibility lag, transport and heating cannot decarbonise at the assumed rate. The electricity system is not merely one line in the carbon budget. It is the platform on which the rest of the target depends.

Industry: the difference between decarbonisation and disappearance

DESNZ’s 2025 statistics show industry emitted 41.1 MtCO2e, around 11% of UK emissions. The sector fell 11.6% from 2024 to 2025, but DESNZ attributes much of the decline to blast furnace closures in iron and steel and lower gas use across industries. That distinction matters. Lower industrial emissions can mean cleaner production, or it can mean less domestic production.

The CCC pathway assumes something more demanding. Carbon Brief’s analysis says industry emissions fall by 78% by 2040 while output is broadly maintained, with electrification providing the largest share of 2040 abatement.

The delta is process change: electrified industrial heat, electric arc furnaces, cleaner chemicals and cement routes, CCS where unavoidable, CO2 transport and storage, and power prices that do not simply push production offshore. This is where the carbon target becomes an industrial strategy test.

The problem is not whether industrial emissions can fall. They can. The question is whether they fall because the UK produces the same goods more cleanly, or because production, jobs and supply chains move elsewhere. Only the first version is a serious decarbonisation success.

Agriculture and land: the residual becomes central

DESNZ’s 2025 sector figures put agriculture at 13% of emissions. It is easy to treat that as a smaller category while power, cars and buildings dominate the debate. But as those sectors fall, agricultural methane, fertiliser, peat, trees and land-use choices become proportionally more important.

The CCC pathway relies on peatland restoration and management, woodland creation, energy crops, agroforestry and hedgerows, alongside changes in agricultural practice and demand. Carbon Brief’s readout notes that agriculture falls more slowly than power or transport and becomes a much larger share of residual emissions by 2040.

The cost here is less like a single infrastructure budget and more like a land-use settlement: payments to farmers, restoration programmes, monitoring, enforcement, land competition, food-system incentives and rural politics. It is slower than building a wind farm or buying a car. It is biological, local and contested.

This is also where the political language of net zero becomes least useful. You cannot just “deploy” land in the way you deploy solar capacity. Land has owners, uses, communities, traditions and trade-offs. The carbon budget makes that unavoidable.

Aviation: the awkward residual

Aviation is the awkward sector because it does not have a straightforward full decarbonisation route on the same timeline as cars or power. Carbon Brief’s analysis of the CCC pathway has aviation emissions falling only 17% to 29.5 MtCO2e by 2040, at which point aviation becomes the highest-emitting sector in the pathway.

That means aviation increasingly becomes a question of who pays for residual emissions. Sustainable aviation fuels, efficiency improvements, demand restraint and engineered removals all matter, but the broad logic is clear: if the sector cannot decarbonise at the same pace, it either pays for removals, limits demand growth, or consumes more of the remaining carbon budget.

This is a reminder that the 87% target is not a pure technology story. Some sectors electrify. Some become more efficient. Some need land or removals. Some face demand questions. The politics differ by sector, which is exactly why the target cannot be treated as one national spending line.

The delivery risk: infrastructure credibility

The UK can make a technically credible case for the 87% target. But technical credibility is not delivery credibility. The final test is institutional: can the UK finance, permit, build, connect and maintain the assets required, while keeping policy stable long enough for households and firms to invest?

This is where HS2 matters. Not because a railway and a carbon budget are the same project, but because HS2 has changed the public context in which any long-term infrastructure promise is heard. It showed how a national project can become narrower, later and more expensive at the same time. That matters for net zero because the 87% target depends on many infrastructure programmes moving together: grids, generation, charging, home heat, industrial clusters, ports, CO2 networks, storage, planning and skills.

The Guardian reported and Reuters also summarised the latest HS2 reset as a project that was originally approved around a £32.7 billion budget and a 2026 opening date for the original network, but now faces a reported government range of £87.7-£102.7 billion, with London-Birmingham services potentially delayed until 2036-2039. The Institute for Government’s HS2 explainer sets out how repeated scope changes, inflation, ground conditions, design choices and governance problems contributed to the escalation.

That comparison is not about saying net zero will become HS2. It is about recognising the same institutional risk: optimistic early assumptions, unstable scope, weak delivery control, slow planning, fragmented accountability and political reset after political reset. A carbon target can survive on paper while the practical programme beneath it drifts.

The right lesson is not to avoid ambition. It is that ambition without delivery architecture is fragile. The 87% target requires long-term commitment, but also ruthless practical discipline: credible cost ranges, transparent contingencies, stable market signals, local planning capacity, procurement competence, network coordination and a clear answer to who pays first.

That is also why this cannot be treated as a “throw money at it” problem. Britain already has deep infrastructure constraints: slow grid connections, planning bottlenecks, overloaded local authorities, high construction costs, weak productivity in delivery, and a political habit of changing course mid-project. More money helps only if the system can convert it into assets. Otherwise the result is higher bills, delayed delivery and declining trust.

The project is therefore both ambitious and expensive, but the alternative should not be described as costless. Fossil-fuel import exposure, volatile energy bills, ageing infrastructure, climate damages and lost industrial opportunities also carry costs. The honest question is not whether the UK pays. It is which costs it chooses, when it pays them, and whether it gets durable assets in return.

That is the proper test of the 87% target. Not whether the number can be legislated. Not whether the CCC pathway is technically plausible. Not whether the spreadsheet says 0.2% of GDP. The test is whether Britain can do what recent infrastructure history has made people doubt: commit for decades, build competently, control costs, and keep public consent through the messy middle.

Sources

· DESNZ, provisional 2025 UK greenhouse gas emissions statistics

· Climate Change Committee, The Seventh Carbon Budget

· Climate Change Committee, supplementary analysis of the Seventh Carbon Budget

· Carbon Brief, analysis of the CCC Seventh Carbon Budget pathway

· The Guardian, report on revised HS2 costs and timetable

· Reuters, report on revised HS2 costs and timetable

· Institute for Government, HS2 costs explainer

The margin is cleaner. The base is not.

The most useful way to read the IEA Global Energy Review 2026 is not as a victory lap, but as a change in the shape of the problem. Global energy demand grew by 1.3% in 2025, slower than in 2024. Solar PV was the largest single contributor to incremental global energy demand, and low-emissions sources together contributed nearly 60% of total demand growth. That is a meaningful shift. Clean technologies are no longer a decorative part of the energy balance; they are visibly affecting the margin.

But that does not mean the fossil system is being displaced cleanly or uniformly. The same IEA review says oil, natural gas and coal demand all grew in 2025, albeit more slowly than in 2024. The distinction matters. Meeting a large share of incremental demand is not the same thing as reducing the existing fossil stock. It means the next layer of growth is becoming cleaner, while much of the installed system remains fossil-based.

This is also where caution is needed with displacement language. In an annual global report, we cannot observe a simple one-for-one substitution where every new solar MWh cleanly removes a specific coal or gas MWh. Dispatch, demand growth, weather, hydro conditions, nuclear availability, fuel prices, curtailment and grid constraints all affect what actually gets backed down. The IEA’s direction of travel is credible, but the language should be treated as system-level contribution, not mechanical replacement.

The transition is now strong enough to shape the margin. It is not yet strong enough to make the fossil base disappear.

Solar has moved the bottleneck

Solar PV is the report’s obvious headline. The IEA estimates that solar generation rose by around 600 TWh in 2025, the largest annual increase ever recorded for any electricity source outside post-crisis rebound years. It also says solar alone met around 70% of global electricity generation growth. Ember’s own global electricity review broadly reinforces the power-sector direction, finding that renewables overtook coal in the global electricity mix in 2025 and that solar posted record growth.

That does not make solar a finished story. It makes solar a system story. The IEA reports record renewable capacity additions of around 800 GW, with solar providing about three-quarters of the total. China alone added nearly 370 GW of solar PV and 117 GW of wind. At that scale, the constraint increasingly shifts from the cost of modules to the ability of grids and markets to absorb the output.

A better way to read this is that solar has solved one problem and exposed the next. Cheap generation is not the same as useful generation. A solar project still needs a connection, an offtake route, a market structure, a profile value, and enough flexibility around it to prevent the value of its own output from being cannibalised. The IEA itself notes that renewable expansion is taking place despite grid-connection delays, supply-chain strains, financial pressures and policy shifts.

This is where the investment lens becomes more interesting than the capacity chart. If modules are abundant, the scarce asset is not necessarily the panel. It may be the grid slot, the land position, the permit, the storage pairing, the contracted offtake, or the ability to inject power when the system actually needs it.

Solar is no longer just a cost curve. It is a test of whether power systems can digest abundance.

Electrification shifts the burden onto infrastructure

The IEA argues that 2025 confirmed the “Age of Electricity”. Global electricity demand grew by around 3%, more than twice the rate of overall energy demand, adding roughly 800 TWh. Buildings, industry, appliances, EVs and data centres all contributed. In the United States, data centres accounted for around half of electricity-demand growth.

The usual reading is that this is good for decarbonisation because more end uses can be served by cleaner power. That is directionally true, but incomplete. Electrification also moves the centre of gravity of the transition onto networks, substations, transformers, connection queues, capacity adequacy, planning systems and consumer bills. The power system becomes the bottleneck through which more of the economy must pass.

This is not just a technology question. It is a balance-sheet question. Fuel-cost exposure can decline in some places, but the system may replace it with long-duration infrastructure commitments: grids, storage, balancing, firm capacity, interconnectors, and capacity payments. That is not an argument against electrification. It is an argument against pretending the cost simply vanishes because the fuel input changes.

Data centres are a good example. Globally, they are still not the whole electricity-demand story. Locally, they can be large enough to shape connection queues, procurement strategies and grid-reinforcement needs. A data centre does not stress the global average. It stresses the substation.

Fossil fuels are changing job description

The fossil fuel story in the IEA review is slower growth, not collapse. Oil demand increased by 0.65 mb/d, gas demand rose by around 1%, and coal demand grew by around 0.4%. Those are not explosive numbers, but they are still positive. The Global Carbon Project also projected a record year for fossil CO2 emissions in 2025, with fossil emissions rising by around 1.1%, although its scope and timing differ from the IEA’s energy-related CO2 estimate.

The important shift is functional. Fossil fuels are losing some of their easier growth markets, but they remain embedded in harder-to-displace roles. Oil demand growth is increasingly tied to petrochemicals and aviation rather than road transport. Gas remains tied to heating, industrial use and flexible power generation. Coal weakens in some systems, but still reappears when gas prices rise or security politics harden.

The oil section makes this clear. The IEA notes that road-fuel demand in advanced economies was broadly flat as efficiency, hybrids and electrification offset activity growth. In China, gasoline and diesel demand were virtually unchanged despite reported GDP growth, while petrochemical feedstock demand accounted for almost all of China’s oil-demand increase. That is not a simple “fossils are over” story. It is a story about where fossil demand becomes stickier.

Gas shows the same pattern. The IEA says buildings became the largest driver of gas-demand growth in 2025 because colder weather lifted heating needs. In Europe, weaker wind and hydro output pushed gas use higher in power generation even as high gas prices weighed on industry. Gas is not just a transition fuel in a slogan. It is also the fuel the system reaches for when weather, flexibility and reliability collide.

The easy fossil demand is being attacked. The sticky fossil demand remains.

Emissions language needs caution

This is the section where the rhetoric should be most disciplined. The IEA says global energy-related CO2 emissions rose by around 0.4% in 2025, reaching a new high of nearly 38.4 Gt. The Global Carbon Project, using a different but overlapping fossil-carbon accounting frame, projected fossil CO2 emissions at around 38.1 Gt in 2025 and also at a record high. The two numbers are not identical because the scopes, timing and methodologies differ, but the basic direction is consistent: emissions did not fall globally.

The IEA also estimates that clean technologies deployed since 2019 avoided more than 35 EJ of annual fossil-fuel demand and around 3 Gt of CO2 emissions annually. That is an important modelling result, but it should be described carefully. “Avoided” emissions are counterfactual estimates. They depend on assumptions about what would have happened without the clean technology deployment: what plants would have run, what vehicles would have been sold, what heating systems would have been installed, and how demand would have been met.

Counterfactual estimates are still useful. Energy modelling has to ask what demand and emissions might have looked like under a different technology mix.

A fair reading is that clean technology deployment is changing the shape of fossil demand, rather than removing the fossil system evenly. It can reduce fuel burn in some sectors and periods, especially where renewables displace generation or EVs reduce road-fuel use. But firm demand, heating needs, industrial use and security margins remain more stubborn, and in stressed conditions they can still pull gas or coal back into the mix.

Clean technologies are bending the system. The harder test is whether that translates into durable reductions in absolute fossil demand, measured emissions, reliability risk and total system cost.

That is the distinction worth keeping. The counterfactual helps explain direction. The observed system shows how much of that direction has actually converted into structural displacement.

China is building transition and insurance at the same time

China remains the decisive swing factor. The IEA says China’s energy-demand growth slowed to 1.7%, while rapid renewables and nuclear growth helped push down coal use in electricity generation. It also estimates that China’s emissions fell by around 0.5%, helped by renewables, slower energy-intensive industry and electric vehicles.

But China is not running a tidy substitution model. It is building clean capacity and energy-security insurance at the same time. The IEA notes that China commissioned almost 80 GW of coal power plants in 2025, following a wave of approvals after the 2021 power shortages. The report frames much of this as capacity intended to meet peak demand and support energy-security goals.

That matters because it says something about how serious energy strategy actually behaves. China is not choosing between clean deployment and optionality. It is buying both. Solar, wind, nuclear, batteries, grids, coal capacity and domestic industrial supply chains are all part of the portfolio.

Western commentary often tries to force China into either “green leader” or “coal villain”. The more useful reading is that China is building the future energy system while retaining a large insurance layer from the old one. That may be inelegant, but it is strategically coherent.

Advanced economies are lower-fossil, not post-fossil

One of the more awkward IEA findings is that emissions from advanced economies grew faster than those from emerging and developing economies for the first time since the 1990s. Advanced-economy emissions rose by 0.5%, driven by colder weather, higher gas consumption and gas-to-coal switching in the United States.

The point is not that the long-term trend has reversed. It is that advanced economies are not post-fossil systems. They are lower-fossil systems under normal conditions, but their emissions profile still responds to weather, fuel spreads, hydro availability, wind output, peak demand and backup requirements.

The United States shows one version of the issue. Strong electricity demand, higher gas prices and slower coal retirements pushed coal-fired generation higher. Europe shows another. Solar output reached new highs, but weaker wind and hydro still left more room for gas-fired generation.

This is a system-completeness issue. Average annual clean generation does not tell you how the system behaves at the margin. A power system can reduce fossil generation across the year and still lean on fossil capacity when demand rises, renewable output is weak, or dispatchable capacity becomes scarce.

That is the harder benchmark for advanced economies now: not headline clean generation shares, but whether lower fossil dependence holds under adverse operating conditions.

Weather is not a footnote

The 2025 energy story is impossible to read properly without weather. The IEA repeatedly points to lower cooling demand after the extreme heat of 2024, colder winters in advanced economies, weak hydro and wind in parts of Europe, and India’s early and strong monsoon. These were not small footnotes; they materially changed fuel demand, electricity demand and emissions.

The IEA estimates that weather-related factors, including temperature variations and shortfalls in hydropower and wind, pushed fossil-fuel combustion emissions up by around 90 Mt CO2 in 2025. At the same time, India’s emissions outcome was helped by monsoon conditions that lowered cooling demand and boosted hydro.

That should make everyone more cautious with annual narratives. A good monsoon can make India look cleaner. A cold winter and weak renewables output can make advanced economies look dirtier. Neither should be mistaken for the full structural trend.

For high-renewables power systems, weather is not external noise. It is the operating environment. The relevant question is not only how the system performs across the annual average, but how it performs in the bad sequence: cold, dark, low-wind, low-hydro, high-demand, high-price conditions.

Storage is where solar volume becomes system value

Battery storage is one of the most important system signals in the report. The IEA says battery storage was the fastest-growing power technology in 2025, with around 108 GW of new capacity deployed, up 40% from 2024. Installed capacity is now eleven times higher than in 2021, and around 80% of new battery capacity was utility-scale.

This matters because storage is the bridge between solar abundance and system usefulness. As solar penetration rises, the problem shifts from producing electricity to shifting it, storing it, pricing it and delivering it when it has value.

This is where deployment charts meet market design. Solar can depress its own capture price if too much output arrives at the same time. Storage can defend value by shifting output, reducing curtailment, relieving congestion and providing ancillary services. The more solar succeeds, the less optional flexibility becomes.

The investment question therefore moves from “what is growing?” to “what becomes scarce after growth?” If solar modules are abundant, the scarce assets are the ones that make solar useful: batteries, grid access, interconnection, flexible demand, firm capacity and market rules that reward availability when it matters.

Nuclear is back in the conversation before it is back in construction

Nuclear generation reached a record high in 2025, but the construction picture is much more concentrated than the rhetoric suggests. The IEA says 3 GW of new nuclear capacity came online, offset by 3 GW of retirements, leaving global nuclear capacity at around 420 GW. It also notes that construction started on 12.2 GW of nuclear capacity in 2025, with nine starts in China and one in Russia.

That does not make nuclear irrelevant. It makes delivery the central question. Nuclear is strategically attractive because it can provide firm low-carbon power, but in many Western markets the gap between policy enthusiasm and actual construction remains large. Respectability has improved faster than build rates.

The point is not to dismiss nuclear. It is to avoid treating rhetorical revival as capacity delivery. Energy systems are not decarbonised by conference panels. They are decarbonised by assets that are financed, permitted, built, connected and operated.

The real conclusion: from abundance to scarcity

The IEA review gives the abundance story: record solar, record renewable capacity additions, fast battery growth, rising electrification and meaningful avoided fossil demand relative to counterfactual baselines. That story is real. But the more useful strategic reading is the scarcity story underneath it.

The next phase of the transition is less about proving that clean technologies can exist at scale. They already do. It is about the scarce things that determine whether scale becomes value: grids, land, permits, connections, flexibility, storage duration, firm capacity, planning consent, public tolerance and market designs that pay for reliability.

That is why the IEA review should not be read as either triumph or failure. It is a report about a transition that is materially advancing while becoming more operationally complicated. Clean energy is increasingly taking the margin. Fossil fuels still dominate parts of the base. Emissions are lower than they would likely have been without clean technology deployment, but they are not yet falling globally.

The bottleneck has moved. It is no longer just technology cost. It is system usefulness.

Source note

This article is based primarily on the IEA Global Energy Review 2026, with cross-checks from Ember’s Global Electricity Review 2026 and the Global Carbon Project’s 2025 carbon-budget release.

Executive verdict. Spain has genuinely weakened the link between gas and the day-ahead electricity price. That is the strongest version of the renewables argument: high wind and solar output mean gas is less often the visible marginal price-setter, and Spain’s headline wholesale price has been far below more gas-exposed European markets. But that is not the same as saying Spain has eliminated gas from the cost of operating the system or from the final bill. In 2025, technical restrictions rose sharply, were overwhelmingly CCGT-led on the upward side, and added materially to the final market price. Early 2026 evidence points in the same direction: the restrictions and reinforced-operation layer appears elevated again, even as PVPC design and tax cuts cushioned household bills.

1. The argument Spain is now asked to carry

Spain has become one of Europe’s most cited examples of renewable-led price decoupling. The simplified story is attractive: Spain built substantial wind and solar capacity, gas was pushed out of the merit order, and wholesale electricity prices stayed low even as gas markets were stressed by the Middle East conflict. The political force of that claim is obvious. It suggests that the route to affordable power is not to redesign markets around fossil-fuel scarcity, but to reduce the number of hours in which fossil fuels are needed at all.

That claim contains an important truth. Analyses of early 2026 put Spain’s wholesale price far below more gas-exposed markets and estimate that gas set, or strongly influenced, the Spanish day-ahead price in only a small minority of hours. The most defensible version is therefore not that renewables have made the whole system cheap in every dimension. It is that renewables have materially reduced gas marginality in the visible energy market.

The problem is that public debate often compresses a layered electricity system into a single market number. Spain’s day-ahead market can clear cheaply while the system operator subsequently brings gas-fired units back through technical restrictions to maintain grid security. This does not invalidate the renewable price effect. It changes what that effect proves.

2. The price stack: one system, four different questions

The day-ahead price is the right benchmark for one question: what did the energy market clear at? It is not a complete benchmark for another: what did it cost to operate the system securely and what did households actually pay?

This sequencing is not a technical footnote. Electricity is delivered through a physical network with voltage limits, congestion boundaries, reserve requirements, ramping constraints and real-time balancing needs. A market schedule can be economically efficient in the abstract and technically inadequate in operation. Technical restrictions are the mechanism used to turn the abstract schedule into a viable one.

3. What technical restrictions are doing in Spain

Technical restrictions are not an accounting fiction or a hidden subsidy. They are a formal security mechanism. When the day-ahead schedule creates a local network constraint, voltage-control issue, reserve shortfall, minimum-generation requirement or other operational problem, the system operator can modify dispatch. Some units are moved upward; others are moved downward. The result is more expensive than the pure merit-order schedule, but more compatible with the physical grid.

In Spain, the upward side of that redispatch is heavily associated with combined-cycle gas turbines. That is the crucial nuance: gas may be absent from the day-ahead margin while still being required later as an operational resource. The former tells us about price formation. The latter tells us about the cost of making the schedule deliverable.

4. The hard evidence from Spain: 2025 and 2026 so far

The key point is not that Spain has no renewable price benefit. It clearly does. The point is that the technical-cost layer did not fall with day-ahead gas exposure. It rose.

Modo Energy’s dataset-level analysis reinforces the same point. Between January 2025 and March 2026, it estimates that average daily CCGT generation scheduled in the day-ahead market was 598 MW, while the provisional viable schedule after technical restrictions was 2,770 MW. That does not mean the day-ahead price is fraudulent. It means the day-ahead price is incomplete.

The 2026-so-far

The 2026 evidence is not yet a clean full-year official dataset, so it should be used carefully. But the direction is important: the technical-restriction and reinforced-operation layer is not obviously falling. It appears elevated on a run-rate basis.

This is where the Iran-war point becomes more nuanced. Spain can be partly protected from gas in the day-ahead energy price while still paying gas-linked costs when CCGTs are scheduled through technical restrictions or reinforced operation. Higher gas prices matter less often in the headline market, but still matter when gas units are called for system security.

5. What household bills actually show

Household bills show consumer insulation, not proof that underlying gas-system costs disappeared. Spain’s regulated-bill and cross-country tariff evidence indicates that consumers were cushioned relative to many European peers. But the bill outcome combines several forces: renewables, the PVPC formula, taxes, regulated charges and supplier/tariff design.

The bill verdict is therefore precise: Spain did not show a clean war-driven household bill spike, but that does not mean gas-linked costs fell. Bills were cushioned by renewable wholesale insulation plus policy and tariff design, while technical restrictions continued to carry gas-linked operating costs.

6. Europe confirms the same lesson

Spain was not the only country where final household bills were shielded. HEPI shows that several capitals saw flat or falling residential electricity prices in April 2026. But Spain belongs to the more insulated group because its all-in household tariff was below the European capital average and materially below gas-exposed markets.

The European comparison changes the claim in two ways. First, Spain is not uniquely immune: Portugal, France and the Nordics also show protection for different structural reasons. Second, Spain’s low household tariff does not by itself prove that renewables alone lowered final bills. Cross-country tariffs bundle energy, network, tax and policy choices. The cleaner statement is that renewables reduced Spain’s wholesale gas pass-through, while final bills were also shaped by policy and tariff design.

7. The wider European comparison: day-ahead prices are not full-system prices

The Spain debate is part of a broader European pattern. Every power system has a first-layer energy price and a second-layer system-operation channel. The names differ by market design, but the physical problem is common: a cheap or single national day-ahead price can understate congestion, local voltage requirements, reserves, redispatch and balancing costs.

The table below is not a harmonised ranking. It is a diagnostic map. Spain reports technical restrictions as part of the final electricity-market price. Great Britain reports balancing and constraint costs through NESO and recovers them through BSUoS. Germany reports congestion-management and redispatch costs. Italy uses the Dispatching Services Market and uplift. France, the Netherlands, Belgium and the Nordics place comparable costs in different reporting and recovery channels.

The point is not to make Spain look uniquely bad. It is to avoid confusing cheap energy with cheap reliability. International price comparisons need a second line below the headline day-ahead number: what had to be done after the market cleared, who was brought on or turned down, and where was that cost recovered?

8. What the Spanish case actually proves

Spain proves less than the victory-lap version claims, but more than the sceptical version allows. It proves that renewable penetration can materially reduce the frequency with which gas sets the day-ahead energy price. That is economically meaningful: it lowers wholesale prices in high-renewable hours and weakens the gas-price transmission channel into the energy component of electricity costs.

It does not prove that renewable-heavy systems become cheap automatically. As low-marginal-cost generation suppresses energy prices, the scarcity value of other attributes increases: flexibility, dispatchability, location, voltage support, frequency response, reserve provision, congestion relief and controllability. These are not optional extras. They are the operating conditions under which cheap energy becomes reliable electricity.

9. Final verdict: the cheapest MWh is not always the MWh that keeps the lights on

Spain’s renewables story is real, but it should be stated precisely. The country has materially weakened the link between gas and the day-ahead electricity price. That is valuable: it reduces fossil-fuel pass-through, supports lower wholesale prices in high-renewable hours and helps explain why Spain’s consumer tariff position looks better than many European peers.

But the stronger claim - that Spain has removed gas from the cost of electricity - does not hold. The system still needs CCGTs for technical restrictions, voltage control and security redispatch. In 2025, those costs rose sharply. In 2026 so far, the available figures suggest they remain elevated and may be running above previous years depending on the scope used. Final bills have been protected, but partly because taxes and tariff design absorbed some of the shock.

The most bulletproof formulation is therefore: Spain has partly decoupled the price of energy from gas; it has not fully decoupled the cost of reliability from gas. Renewables explain the first part. Technical restrictions, CCGT redispatch, taxes and tariff design explain why the final-bill story is more complicated.

Substack pull quote. Spain’s cheap day-ahead power is real, but it is not the whole price. Gas has been pushed out of the visible merit order in many hours, while still being called back through technical restrictions when system physics requires it.

Method note: evidence quality and comparability

Sources

• Red Eléctrica - Average final price

• Red Eléctrica - Energy and prices scheduled for security reasons

• Red Eléctrica - Components of final price and energy at closing

• Modo Energy - Has Spain’s power system decoupled from natural gas?

• Cinco Días - REE reports reinforced-operation cost to 30 April 2026

• HEPI - Household Energy Price Index price data

• La Moncloa / Royal Decree-Law 7/2026 context

• OMIE - Day-ahead and intraday market rules

• NESO - Annual Balancing Costs Report

• SMARD / Bundesnetzagentur - Congestion management

• Terna - Dispatching Services Market / dispatching fees

• RTE - Annual Electricity Review

• IEA - The Netherlands Energy Policy Review

• Elia Open Data - balancing energy prices

• Nord Pool - 2024 market statistics

• Svenska kraftnät - Balancing Market Outlook

The July 2025 REMA decision retained a single GB-wide wholesale price and moved the reform programme toward Reformed National Pricing rather than zonal pricing. DESNZ presented this not as doing nothing, but as a package: stronger upfront locational investment signals, reform of network and connection charging, strategic spatial planning, and changes to balancing and constraint management. In the government’s own language, RNP is meant to give investors a clearer view of the “relative system value” of different locations before final investment decisions, while preserving national wholesale pricing and investor certainty.

That framing is important. The government is not saying location does not matter. It is saying location should be handled through planning, charges and support-scheme design rather than through different wholesale prices across the country. NESO describes RNP as a programme that keeps a single national wholesale price while tackling three problems: stronger locational investment signals, improved balancing and dispatch, and better constraint management.

So the real question is not whether the UK has gone back to the status quo. It has not. The sharper question is whether RNP is a coherent substitute for locational pricing, or whether it is an administrative workaround forced by a market design that has already centralised too much investment risk through national-price CfDs.

1. The awkward chronology: nodal died first, then zonal died for a similar reason

The uncomfortable part of REMA is the sequence. In the 2024 second consultation, DESNZ explicitly narrowed the locational-pricing options. It continued to assess zonal pricing but discounted nodal pricing on grounds of investor confidence, deliverability and decarbonisation risk. The document recognised that nodal pricing could deliver sharper signals and potentially higher benefits, but argued that greater revenue uncertainty and delivery risk could erode those benefits. It also estimated that zonal pricing might take at least five years to implement, while nodal pricing could take up to a decade because it would require central dispatch.

The same consultation also showed why locational pricing remained attractive in the first place. DESNZ stated that locational pricing would embed the locational value of energy into the wholesale price, reflecting supply, demand and network capacity in each location. It cited commissioned modelling suggesting zonal pricing could reduce system running costs by roughly £5bn-£15bn over 2030-2050, with potential consumer benefits of roughly £25bn-£60bn over the same period, while warning that those figures did not capture all costs, including cost-of-capital impacts.

That is where the process critique bites. If nodal was excluded because granular locational pricing could increase revenue uncertainty and raise financing costs, why did zonal remain the central reform candidate for so long before being rejected on a related logic? Zonal is less granular than nodal. It is easier to explain, easier to hedge, and less disruptive. But it is still locational wholesale pricing. It still transfers some locational price and volume risk from consumers and the system operator toward generators and flexible assets.

The government’s best defence is that zonal was a plausible middle ground worth testing: coarse enough to preserve investability, sharp enough to reveal congestion. That defence is not absurd. But the final decision makes the earlier process look exposed. REMA rejected the technically pure version of locational pricing, spent years testing the politically softer version, then concluded that even the softer version threatened the financing model. In that sense, the “dead horse” was not just zonal pricing. It was the attempt to graft locational wholesale risk onto a system increasingly built around national-price investment certainty.

2. CfDs solved one problem and made another harder

The UK’s Contracts for Difference model was designed to solve a real problem: low-carbon projects are capital-intensive, and investors need revenue confidence to build at scale. The CfD scheme, established in 2014, is the government’s main mechanism for supporting low-carbon electricity generation; Low Carbon Contracts Company says it has facilitated investment in 39GW of renewable and low-carbon generation projects including nuclear.

Mechanically, a CfD gives a generator a guaranteed strike price against a market reference price. If the market price is below the strike price, the generator receives a top-up; if the market price is above, it pays back. LCCC describes the scheme as protecting generators from market price volatility and giving investors certainty on returns, thereby reducing financing costs. The scheme’s market reference prices are then used to calculate payments.

That architecture has been extremely useful for procuring renewables at low financing cost. But it also centralises the investment model around a national reference price. A wind farm can be made financeable on the basis of a national price and government-backed revenue stabilisation even if its true system value is lower because it sits behind a constrained boundary. The project looks cheap in the auction; the locational cost can reappear later as curtailment, replacement generation, balancing actions, accelerated network reinforcement, or socialised charges.

This is the deeper lock-in. The more the UK relies on national-price CfDs to mobilise cheap capital, the harder it becomes to expose assets to sharper locational price risk later. The CfD is not just a support scheme; it becomes a political economy. Developers, lenders, advisers, supply chains and auctions all organise around the assumption that the state is suppressing a major slice of merchant risk. Once that becomes the financing norm, locational wholesale reform looks less like market optimisation and more like a retrospective risk reallocation.

3. The system cost is no longer invisible

The old national-pricing model becomes harder to defend as constraints rise. NESO’s 2025 Annual Balancing Costs Report says wind curtailment is a major driver of balancing costs because a large share of GB wind capacity is connected in Scotland, currently a constrained region of the network. When wind generation is high, NESO has to turn down wind and turn on replacement energy elsewhere to keep the system balanced. In 2024/25, wind curtailment rose to 13% of hypothetical wind outturn.

The same report says overall balancing costs totalled £2.7bn in 2024/25, up 10% from 2023/24, with the increase attributed to higher constraint costs and volumes. This is exactly the kind of problem locational pricing is meant to expose: not all MWh are equal if the grid cannot move them to where demand is.

RNP tries to fix that without changing the national wholesale price. This is the planning-led pivot: use the Strategic Spatial Energy Plan, grid buildout, connection reform, TNUoS and charging reform, and balancing-market changes to steer the system. But that approach substitutes administrative foresight for continuous market revelation. It may work for big, slow infrastructure choices. It is weaker for real-time operation, where the value of electricity can change by location every few minutes.

4. RNP is planning-led, not price-led

The centrepiece is the Strategic Spatial Energy Plan. DESNZ’s 2026 RNP Delivery Plan says the SSEP will be the first long-term spatial strategy for Great Britain’s energy generation and storage infrastructure, guiding what capacity is built, where it is located, and when it comes forward. NESO describes the first SSEP as a GB-wide plan mapping potential zonal locations, quantities and types of electricity and hydrogen generation and storage, while making clear it will not identify specific projects.

This is not the same as doing nothing. Planning can coordinate transmission corridors, offshore leasing, connection queues, anticipatory network investment and broad technology volumes. It can avoid obvious absurdities: building large volumes of generation in places where grid capacity is years behind, then paying those assets to switch off.

But planning is not equivalent to prices. A plan can tell the market where government expects value to be. A locational price tells the market where value actually is under current physical conditions. That difference matters for batteries, demand response, interconnectors, electrolysers and flexible loads. These assets are not just investment objects; they are operational resources. They need sharp signals about where to charge, discharge, consume, produce or withhold. RNP can approximate some of this through constraint management and balancing reform, but it is not the same instrument.

That is the central trade-off. Planning can be efficient for big, slow and lumpy decisions: transmission corridors, offshore leasing, connection sequencing, broad technology volumes and strategic investment support. It is less efficient for revealing changing operational value. Prices expose congestion continuously. Plans approximate it administratively. The question is whether the UK can make that approximation sharp enough without recreating the investment-risk problem it was trying to avoid.

5. The US comparison: nodal pricing is workable, but only inside the right architecture

The US example weakens any claim that nodal pricing is inherently unworkable. In organised US markets, locational marginal pricing is a normal part of wholesale market design. ISO New England describes LMP as a way for wholesale electricity prices to reflect the value of energy at different locations, accounting for load, generation, physical transmission limits and losses. It calculates prices at over 1,000 pricing nodes, with LMPs calculated every five minutes.

The point is not that investors face raw nodal volatility with no mitigation. US markets have day-ahead markets, congestion hedges, market monitoring, capacity and ancillary-service revenues in many regions, and long-term bilateral contracting. PJM’s Financial Transmission Rights are explicitly designed as financial instruments related to transmission congestion; PJM says they allow market participants to hedge congestion charges, with revenues based on day-ahead congestion price differences between specified source and destination points.

So the correct lesson is subtle. The US shows nodal pricing can work where the market architecture, contracts, hedging instruments and investor expectations are built around locational risk. It does not prove GB can bolt nodal pricing onto a CfD-heavy national-price system without disruption. In GB, the main renewables investment model has deliberately suppressed wholesale revenue volatility. Moving to nodal late would mean redesigning CfDs, reference prices, supplier hedging, consumer protection, network charging, basis-risk allocation and possibly dispatch arrangements. That is not impossible. But it is not a minor market tweak.

6. Is the UK drifting back toward a centrally planned system?

Not in the old caricatured sense. The UK is not abolishing markets, and RNP still relies on private capital, auctions, contracts and competitive delivery. But it is fair to say the system is moving toward a state-coordinated investment regime with market dispatch at the edges.

That is the irony. CfDs were introduced to reduce investment risk and lower financing costs. They succeeded at that. But by centralising revenue certainty around a national price, they made later spatial market reform more politically and financially difficult. REMA then faced a choice: expose investors to locational price risk and risk higher capital costs, or retain the national price and replace some price signals with planning, charging and administrative levers. It chose the second path.

That choice may be rational under the UK’s current constraints. But it should not be oversold as market-led reform. It is a planning-led response to a market design that has already socialised too much locational risk. The system still uses markets, but increasingly under a strategic state map: where to build, when to connect, how to lease seabed, how to charge networks, how to structure auctions and how to manage constraints.

7. The pathway out: tight, but real

The UK is not trapped forever. But the pathway cannot be a sudden leap from national pricing plus CfDs to full nodal pricing. The credible route is staged, asymmetric and focused first on new and flexible assets rather than retroactive exposure of sunk investments.

1. First, protect sunk assets but reform future CfDs. Existing projects financed under national-price assumptions should not be abruptly exposed to new locational risk. But future CfD rounds can become more locationally intelligent. That could mean SSEP-aligned auction pots, locational eligibility rules, project caps in constrained areas, stronger connection-readiness requirements, or reference-price designs that make future support schemes compatible with deeper locational reform. The principle is simple: do not award cheap capital support to projects whose low bid relies on pushing system costs elsewhere.

2. Second, make network and connection charges sharper but bankable. If wholesale prices remain national, then TNUoS, connection charging and access arrangements must carry more locational information. The signal has to be strong enough to influence siting but stable enough to be financeable. Wildly volatile charges would recreate the same investment-risk problem by another route.

3. Third, expose flexible assets before exposing everything. Batteries, electrolysers, data centres, industrial demand response and other flexible loads are better candidates for granular local signals than legacy CfD renewables. These assets can move, time-shift or respond. Giving them stronger local price or constraint signals can capture some nodal benefits without rewriting the entire investment regime overnight.

4. Fourth, fix balancing and dispatch visibility. Shorter settlement periods, better visibility of market positions, lower barriers for smaller flexible assets to enter the Balancing Mechanism, and more transparent constraint-management products can reduce the gap between national wholesale pricing and physical system needs.

5. Fifth, build a hedging layer before reopening locational pricing. If GB ever reopens zonal or nodal reform, it should not be framed as raw locational exposure. The US lesson is that locational pricing needs financial transmission rights, congestion hedges, liquid forward products and clear treatment of legacy contracts. The debate should be about risk allocation plus hedging design, not just zones versus national price.

6. Sixth, make any deeper reform prospective and testable. A realistic transition would expose new assets first, let flexible demand opt into sharper signals, and use explicit tests: constraint-cost reduction, liquidity, investment premia, CfD bid impacts, consumer distributional effects and network-delivery progress. If those tests fail, pause. If they work, widen exposure.

This avoids a false binary. The answer is not “abolish CfDs and go nodal tomorrow.” The answer is to stop treating national-price CfDs as if they can indefinitely substitute for locational discipline. Planning can coordinate. Charges can steer. CfDs can de-risk. But someone still has to reveal where electricity is actually valuable on the system.

8. Conclusion: not square one, but a narrower corridor

REMA’s final decision is defensible only if RNP becomes a serious spatial and operational reform programme, not a polite label for national pricing with better PowerPoints. The government is right that cost of capital matters. In a capital-intensive transition, a small financing premium can swallow modelled efficiency gains. But that does not erase the underlying problem: the national-price/CfD model has weakened locational discipline, and the system cost is now visible in constraints, curtailment and balancing actions.

The critique, then, is not that REMA should have blindly adopted nodal pricing. It is that REMA’s treatment of nodal and zonal pricing exposed a deeper inconsistency. Nodal was rejected because locational risk threatened investment confidence. Zonal was then kept alive as the workable middle ground, before being rejected because locational risk still threatened investment confidence. That sequence tells us something important: the obstacle was not only the granularity of the pricing model. It was the UK’s prior commitment to a national-price, CfD-centred investment regime.

There is still a path out. But it is no longer wide. The UK has to move gradually: make future CfDs more locationally aware, sharpen network and connection charges, expose flexible assets first, reform balancing and dispatch, build hedging products, and only then consider deeper locational pricing for new investment. The country is not fully locked in. But it has narrowed its own reform corridor.

Sources and further reading

DESNZ, REMA Summer Update 2025

DESNZ, RNP Delivery Plan 2026

NESO, Reformed National Pricing programme

NESO, Strategic Spatial Energy Planning

DESNZ, REMA Second Consultation 2024

Low Carbon Contracts Company, Contracts for Difference overview

Low Carbon Contracts Company, About the CfD scheme

Low Carbon Contracts Company, Market Reference Prices

NESO, Annual Balancing Costs Report 2025

ISO New England, LMP FAQ

PJM, Financial Transmission Rights fact sheet

Reuters, UK rules out zonal energy pricing

The moving parts

For more than a decade, the global solar story has been told through one clean line: panels keep getting cheaper.

That line is still broadly true. But the latest price move out of China tells a more interesting story than a simple reversal in the solar cost curve. Solar is not suddenly becoming expensive. What may be ending is a very specific version of cheap solar: modules pushed ever lower by Chinese overcapacity, brutal producer competition and manufacturers selling into losses.

According to the Financial Times, solar module prices have climbed from around 9 cents per watt at the end of December to 11.4 cents per watt by mid-April, based on BloombergNEF, InfoLink and Bloomberg data. That is a sharp move in percentage terms. But it is still cheap in historical terms: the same report notes that panels were around $2 per watt in 2010.

So the story is not that solar has stopped being cheap. It is that the old floor may have been too low to survive.

That distinction matters. A module price rebound from an unsustainably low base does not invalidate the solar buildout. It does, however, challenge one of the lazy assumptions sitting under parts of the energy debate: that Chinese solar deflation would simply continue, indefinitely, regardless of what it was doing to producer balance sheets or to China’s own industrial-policy priorities.

What actually happened

The immediate news is straightforward. China appears to be leaning against a destructive price war in its solar manufacturing sector. This is happening through a mix of export-rebate changes, investment guidance, industry pressure and a broader official campaign against overcapacity and so-called destructive competition.

One concrete lever is export tax policy. Reuters reported that China would eliminate VAT export rebates for photovoltaic products from 1 April 2026, after previously cutting rebates from 13% to 9% in late 2024. For battery products, the same policy path phases rebates down before eliminating them. The stated logic was to moderate price competition, address overcapacity and reduce trade-dispute risk.

Another lever is investment discipline. China had already finalised stricter solar manufacturing investment guidelines in late 2024, including a higher minimum capital ratio for solar PV projects and tighter efficiency and energy-consumption standards. These rules do not by themselves delete excess factories, but they signal a shift away from the old assumption that more capacity is always better.

The price response now looks visible. The FT’s reported move from 9 cents to 11.4 cents per watt is not a return to expensive solar. It is a rebound from a level where too many producers were bleeding.

The overcapacity machine

China’s solar dominance did not happen by accident. It came from industrial policy, manufacturing scale, supply-chain control, low-cost production, intense competition and a willingness to tolerate margins that many Western industrial strategies would find politically or financially impossible.

That system produced an extraordinary result for the world. Solar modules became cheaper at a speed few energy technologies have matched. It also created a structural problem. Capacity kept expanding even as prices fell. Producers competed on cost, scale and market share until the industry started to resemble a transfer mechanism: cheaper modules for the world, losses absorbed inside Chinese corporate and local-industrial systems.

The oversupply problem is not marginal. Reuters reported that Chinese factories had enough capacity to meet expected global solar demand for 2025 nearly twice over. That is not simply normal competition. It is an industry with too much productive capacity chasing too little profitable demand.

Once that happens, price cuts stop looking like pure productivity gains. Some of them are real learning, scale and process improvements. But some of them are also distress pricing. The difference matters because only the first is durable. The second depends on how long producers, banks, local governments and Beijing are prepared to absorb pain.

The losses became too visible

The financial damage is now hard to ignore. Reuters reported fresh first-quarter 2026 losses among major Chinese solar manufacturers. JinkoSolar posted a quarterly net loss of 463 million yuan, Trina Solar lost 234 million yuan, and Longi Green Energy’s loss widened to 1.92 billion yuan. Improved module prices and storage revenues helped some firms narrow losses, but the sector was still not healthy.

This is the hidden accounting behind part of the cheap panel story. Falling module prices have been good for solar deployment. They have helped developers, installers, households and governments trying to build low-carbon electricity systems. But those gains were not free. They were partly paid for by collapsing margins and balance-sheet damage in the manufacturing base that supplied the world.

That is why the current move is so interesting. A price rebound is usually treated as bad news for clean energy. In one narrow sense, it is: developers would always rather pay less for modules. But in another sense it is a sign that the most destructive part of the market may be being corrected. Ultra-cheap inputs are not always healthy if they come from an industrial structure that is burning its own producers alive.

Beijing is not abandoning solar. It is disciplining it.

The easy mistake is to read this as China losing faith in solar. That is not what is happening. Solar remains central to China’s industrial strategy, export machine and domestic power buildout. Beijing is not walking away from the sector. It is trying to stop the sector from destroying itself through an endless volume-and-price race.

The policy direction is therefore not anti-solar. It is anti-chaos. It is about pushing the sector away from pure competition on scale and price, and toward something closer to quality, consolidation and value. That is also why the rebate and investment changes matter. They are not just technical tax details. They are signals that the state is no longer as willing to underwrite the same level of export-led price destruction.

This is classic industrial-policy tension. The same forces that make a national champion internationally dominant can also become destabilising when too many firms chase the same market with too much subsidised or policy-supported capacity. At first, the result looks like strategic success. Later, it starts looking like capital misallocation.

Solar remains strategic. The price war is what became expendable.

Why silver and materials matter too

Policy is not the only factor behind the price rebound. The FT also points to higher silver costs as a contributor. Silver is used in solar cells, especially in conductive paste, and material-price pressure can still move module economics even when manufacturing technology is improving.

That should not be overstated. Solar has repeatedly found ways to reduce material intensity, improve cell efficiency and push down cost per watt. But it is a useful reminder that clean-energy cost curves are not magic. They are supply chains. They depend on metals, chemicals, factories, logistics, trade policy, financing costs and grid conditions. When one of those layers turns, the curve can flatten or reverse for a period.

The more sophisticated view is not that solar’s learning curve is dead. It is that the learning curve lives inside an industrial system. It can be helped by scale and innovation, but it can also be constrained by materials, policy choices and the need for producers to remain solvent.

This probably does not kill solar demand

The obvious question is whether higher Chinese module prices will damage solar deployment. The answer is: not much by itself, at least not in the way a headline might suggest.

A move from 9 cents to 11.4 cents per watt matters. It can affect procurement, margins and some marginal projects. But modules are only one part of the installed cost stack. In Europe, the UK and other mature markets, the final economics also depend on grid connection, planning, land, labour, inverters, transformers, financing, developer margin, taxes, capture prices and curtailment risk.

This is where the solar debate often becomes too thin. A cheap panel is not the same thing as a cheap delivered system. The module is the visible commodity component. The project is a much larger piece of infrastructure. Once a market has weak grids, long connection queues, limited flexibility or rising curtailment, a few cents per watt on modules is not always the binding constraint.

That is why this should not be framed as the end of solar competitiveness. AP’s reporting on Africa made the same basic point in a different context: China’s rebate changes may raise solar costs and complicate timelines, but solar can remain cost-effective against alternatives such as diesel, especially where fuel-import dependence is high.

For Europe, the effect is likely to be less dramatic than the politics. A modest module price rebound may trim project economics, but it does not suddenly make solar uncompetitive. The larger problems are often system-side: permitting, grids, flexibility, storage, interconnection, cannibalisation and the value of solar output at the hours when solar is abundant.

The world got used to Chinese solar deflation

The deeper lesson is that the rest of the world became used to an unusually convenient arrangement. China overbuilt. Chinese producers fought for share. Prices fell. Global buyers benefited. Governments got cheaper clean-energy deployment. Developers got lower capex. Consumers, in many cases, got access to cheaper systems than domestic industrial policy alone could have provided.

But the losses had to sit somewhere. They sat inside Chinese firms, local government ecosystems, creditor relationships and the wider political economy of China’s manufacturing push. In practice, the rest of the world enjoyed something that looked very close to an implicit industrial subsidy, even when the formal mechanism was more complicated than a single subsidy line.

That is why the new price floor matters. If Beijing no longer wants unlimited producer destruction, one of the major forces behind global solar deflation becomes less reliable. Solar can still get cheaper through efficiency, scale, better manufacturing and design improvements. But the world may get less cheapness from Chinese firms being forced to sell into losses.

This is a more important shift than the module price itself. The market is not just repricing panels. It is repricing the assumption that China will indefinitely absorb the industrial pain required to keep pushing panel costs lower for everyone else.

The Western industrial-policy problem

For Europe and the US, there is a slightly awkward implication. Western policymakers often say they want domestic clean-tech manufacturing. They also want the lowest possible deployment cost. Those two goals can conflict when global prices are set by a Chinese manufacturing base operating at a scale and margin tolerance that domestic producers cannot match.

If Chinese module prices stabilise above the destructive floor, non-Chinese manufacturing may become a little less impossible. But the gap does not disappear. China still has scale, supply-chain depth, process knowledge and manufacturing ecosystems that cannot be replicated by press release. A higher Chinese price floor helps at the margin; it does not create a Western solar industry by itself.

This is where the politics can become confused. If policymakers want domestic production, they need to accept that it may cost more than importing from the cheapest global supplier. If they want the cheapest deployment, they need to accept dependence on a concentrated supply chain. Pretending there is no trade-off is not a strategy.

The same point applies to energy security. Cheap Chinese modules helped accelerate decarbonisation. They also deepened dependence on a single dominant manufacturing system. A more stable Chinese price floor does not remove that dependence. It just makes the industrial structure behind cheap solar more visible.

What to watch next

There are a few things worth watching.

• First, whether module prices keep rising or settle around a more stable floor. A short rebound is different from a structural change in pricing discipline.

• Second, whether consolidation actually happens. China’s policy signals matter, but excess capacity only really clears if weaker producers exit, merge or stop adding capacity.

• Third, whether higher module prices improve manufacturer profitability or are absorbed by higher input costs, including silver and other materials.

• Fourth, whether Europe and the US use the repricing window to build credible manufacturing strategies, or simply complain about dependence while continuing to import the cheapest equipment.

• Fifth, whether solar deployment bottlenecks shift further away from module cost and toward grids, storage, flexibility, permitting and power-market design.

The real conclusion

The headline is that solar panel prices are rising. That is true. But the more useful conclusion is that the economics behind cheap solar are being reset.

For years, the global solar market benefited from a neat equation: China overbuilt, Chinese producers competed aggressively, module prices fell, and everyone else got cheaper panels. That equation is not disappearing, but it is becoming less one-sided.

Solar remains one of the most important and cost-competitive energy technologies in the world. Even after the rebound, module prices are still extremely low by historical standards. The point is not that cheap solar is over. The point is that a particular version of cheap solar may be weakening: the version sustained by Chinese producers selling into losses and Beijing tolerating the resulting industrial damage.

Cheap panels are not over. The idea that they can keep getting cheaper through endless Chinese producer losses is what looks less secure.

That is the story. Not solar’s failure. Not the death of the cost curve. Something more interesting: the moment when the world’s cheapest clean-energy component starts to reveal the industrial system that made it so cheap in the first place.

Sources

Financial Times - Solar panel prices rise after China clampdown on producer price war

Reuters - China to scrap export tax rebates for photovoltaic and battery products

Reuters - Chinese solar manufacturers post fresh losses

Reuters - China solar makers say war-induced demand will not fix overcapacity

Reuters - China finalises stricter investment guidelines for solar manufacturing

Associated Press - Africa’s solar boom faces higher costs as China cuts export subsidies

Wood Mackenzie - The impact of China’s 2024 PV manufacturing guidelines

pv magazine - PV module prices continue to rise unabated

BYD’s fast-charging story is powerful because it attacks one of the remaining psychological advantages of petrol cars: refuelling time. The company’s official claim is that its Super e-Platform can achieve 1 megawatt of charging power and add 400 km of range in five minutes. BYD’s own announcement presents this as refuelling-time equivalence rather than just a charging upgrade; Reuters also reported the 1,000 kW and 400 km in five minutes claim at launch.

That does not make the claim irrelevant. It makes it system-sensitive. The wrong reaction is to dismiss it as impossible simply because most public chargers today cannot do it. The equally wrong reaction is to treat it as proof that EV adoption has been solved. The more precise interpretation is that BYD is trying to shift the adoption debate from “can an EV go far enough?” to “can an EV refuel quickly enough?”

The more grounded version of the story comes from follow-up reporting. Reuters says BYD is pushing flash charging to target hesitant consumers, with claims of charging from 20% to 97% in under 12 minutes and plans for 20,000 flash-charging stations in China plus 6,000 internationally within a year. That is still aggressive. But it is also more useful than the slogan because it raises the right questions: over what state-of-charge window, at what temperature, with what battery degradation, and at what type of charging site?

2. The physics check: 1 MW is not magic, but it is a large local load

A 1 MW charger running for five minutes delivers about 83.3 kWh before losses:

1,000 kW x 5/60 hours = 83.3 kWh

If that produces 400 km of claimed range, the implied energy intensity is roughly 20.8 kWh per 100 km before losses and test-cycle adjustments. For a large EV under favourable assumptions, that is not absurd. But it shows why the headline is not just about charger branding. It requires the vehicle to accept a very high power pulse and the site to deliver it.

This calculation is useful because it identifies the true shift. The problem is not whether enough energy exists. It is whether enough power can be delivered locally, repeatedly, safely and economically.

3. The vehicle stack: high voltage, high current, thermal control and degradation risk

BYD says the platform combines flash-charging batteries, a 1,000V-class architecture, silicon carbide power chips and high-speed drive components. The voltage point matters because higher voltage reduces current for the same power. Lower current helps with resistive losses and thermal stress, although BYD’s own headline still implies extremely high current as well as high voltage.

But fast charging is not only a power-electronics problem. It is also a battery-health problem. Extreme fast charging raises issues around heat generation, lithium plating risk, cell-to-cell thermal gradients, pack cooling and repeated-cycle degradation. Studies of fast charging repeatedly show that temperature and current distribution are not second-order details (Nature Communications; Applied Energy). Even small internal thermal gradients can accelerate uneven degradation, and fast-charge strategies for LFP cells show strong interactions between charging profile, temperature and safety/degradation pathways.

The question is not whether a fresh pack can accept a very high-power charge once. The investable question is whether it can do so repeatedly across seasons, state-of-charge windows, battery ages and warranty conditions without unacceptable degradation.

4. Europe’s current charging baseline: good coverage, different power class

Europe is not infrastructure-empty. The IEA Global EV Outlook 2025 reports that public charging points have doubled globally since 2022 and that Europe has stronger highway fast-charging coverage than many other markets. The IEA executive summary says over three-quarters of European highways have a fast-charging station at least every 50 km.

The European Alternative Fuels Observatory provides the live infrastructure baseline for the EU, with recharging infrastructure data updated monthly. ICCT reported that Europe had roughly 1.14 million public charging points by end-2025, more than four times the 2020 level.

The problem is not absence; it is class. The EU Alternative Fuels Infrastructure Regulation requires car and van charging pools on the TEN-T core network at least every 60 km by end-2025, with at least 400 kW total output and at least one 150 kW charging point. By end-2027, those pools must rise to 600 kW total output with at least two 150 kW points. That is meaningful policy. But compare it with BYD’s claim: AFIR’s early passenger-car requirement is a few hundred kilowatts per pool; BYD is talking about 1 MW to one vehicle.

This is why the European question cannot be answered by counting plugs alone. Plug count matters for access. But BYD-style charging depends on power density, site capacity and the ability to allocate power dynamically across bays.

5. The grid reality: the charger becomes a distribution-network project

A single 1 MW charge point is already a major local load. A hub with eight such bays could theoretically draw 8 MW before site losses, lighting, cooling, retail loads or stationary battery charging. At that point, the project is not simply a retail installation. It becomes a medium-voltage connection and distribution-planning problem.

European grid institutions are already warning that the grid is becoming the bottleneck for electrification. The European Court of Auditors’ 2025 review on making the EU electricity grid fit for net zero highlights delays and the scale of grid expansion required. Eurelectric’s work on distribution-network connection queues describes rising capacity constraints and delays caused by the rapid growth of connection requests across renewables, transport electrification, heating and industrial decarbonisation.

That points to the likely European architecture: not isolated 1 MW posts everywhere, but engineered charging hubs with medium-voltage grid connection, transformer capacity, dynamic power sharing and often on-site stationary storage. Eurelectric's Distribution Grids Handbook notes that assets such as EV charging stations are increasingly connected at distribution level because of consumer proximity and capacity size. The fast-charging site therefore becomes part of local network planning, not just transport infrastructure.

6. The economics: utilisation decides whether the asset is investable

Ultra-fast charging economics are defined by a simple tension: the driver values very high peak power because it saves time, while the operator has to finance equipment, land, civil works, grid connection, transformers, software, maintenance and electricity costs for an asset whose utilisation may be uncertain.

Research on DC fast charging repeatedly finds that utilisation and tariffs are decisive. A recent NREL-linked corridor fast-charging economics study found low-utilisation corridor stations to be roughly six times more expensive per kWh than the average and stations subject to demand charges around 40% more expensive than those without them. A separate NREL assessment of DC fast-charging viability also identifies demand charges and electricity retail prices as major factors affecting station profitability. Earlier work by Muratori and co-authors found that DC fast-charging electricity costs vary greatly and that demand charges create high average electricity costs at low utilisation, with costs falling as utilisation rises.

In a nutshell, the faster the charger, the stronger the consumer proposition. But the faster the charger, the more exposed the operator becomes to peak-capacity costs if the site is not heavily used.

7. Standards and compatibility: the risk of a fast charger that few cars can use

Megawatt charging is not a fantasy category. It already exists as an industry direction for heavy-duty vehicles. SAE J3271 is a system-level standard covering charging equipment and control elements from the point of utility interconnection to the vehicle interface. CharIN frames the Megawatt Charging System (MCS) primarily around trucks and buses, where very large battery packs and tight operational dwell times require much higher power.

For passenger cars, the strategic issue is interoperability. Europe’s current public charging ecosystem has been built around CCS and a regulatory push toward open access, payment transparency and minimum service standards. BYD’s system can help accelerate the technical frontier, but investors will hesitate if a 1 MW site risks becoming a brand-specific or narrow-fleet asset.

A European megawatt passenger-charging network would need compatibility rules before it needs marketing slogans: connectors, communication protocols, payment, cybersecurity, safety certification, warranty treatment and transparent power-rating disclosure.

8. Why the rollout would be uneven across Europe

The right unit of analysis is not “Europe”. It is corridor by corridor, distribution-network zone by distribution-network zone. Europe has high EV penetration and strong charging coverage in some markets, but grid availability, permitting speed, land access, charger utilisation and consumer behaviour vary sharply.

The first viable sites would likely be motorway corridors, airport/taxi hubs, fleet depots, high-income urban charging hubs, retail parks with high traffic, and routes with limited home-charging access. The slower sites would be rural locations with low throughput, constrained distribution networks, difficult land-use planning or weak EV uptake.

This matters for policy. A purely market-led rollout would prioritise high-utilisation nodes. That is efficient, but it can leave coverage gaps. A universal-coverage strategy can fill those gaps, but it may require public support where utilisation is too low to recover fixed costs. This is the same logic that appears in corridor fast-charging economics: excluding stations that fail cost parity can improve commercial returns but reduce geographic coverage.

9. What it would take in Europe

Assuming BYD has the vehicle technology, Europe would need a package rather than a single intervention. The package has five parts.

1. First, grid-ready hub planning. Governments and DSOs would need to identify motorway and urban nodes where several MW of capacity can be made available ahead of demand. This means grid-capacity mapping, anticipatory investment and standardised connection processes.

2. Second, battery-buffered charging where the grid is constrained. Stationary batteries can allow a site to draw more steadily from the grid while delivering short 1 MW bursts to vehicles. This reduces peak pressure but adds capex, land requirements and battery-management complexity.

3. Third, tariff design that recognises flexibility. If charging operators are penalised mainly for peak load, ultra-fast charging becomes difficult to finance. Tariffs need to distinguish unmanaged peaks from flexible, buffered or grid-supportive operation.

4. Fourth, modular and interoperable charger design. The early fleet will be mixed. Most European EVs will not accept 1 MW. Sites should therefore allocate power dynamically across 150 kW, 350-400 kW and future megawatt-capable vehicles rather than strand capacity in dedicated posts.

5. Fifth, a clear investment model. The likely bankable model combines anchor demand from high-throughput users, retail co-location, public risk-sharing for socially useful low-utilisation corridors and transparent standards that reduce stranded-asset risk.

10. Claims, opportunities, constraints: the balanced read

The opportunity is real. If five- to twelve-minute charging becomes repeatable, reliable and widely available, it compresses one of the clearest behavioural disadvantages of EVs. It would be especially important for drivers without home charging, high-mileage users, taxis, fleets and long-distance travel. It would also change how consumers compare EVs with petrol cars: not just on total cost, but on convenience.

But the constraint is equally real. Europe does not merely need faster chargers. It needs grid-ready charging hubs. That means local power capacity, medium-voltage connections, transformers, land, planning consent, dynamic power management, stationary storage in constrained areas, transparent standards, and tariffs that do not punish the very peak-power capability the consumer values.

BYD may be moving the bottleneck. Not from “EVs cannot charge fast enough” to “problem solved”, but from the battery pack to the built environment. The European test is therefore not only technological. It is infrastructural, financial and regulatory.

That is what makes the story interesting. The five-minute EV is not fantasy. But it is not yet a European mass-market condition. It is a system challenge waiting to be built.

Sources and reference links

Key sources used in this draft. Links are embedded for Substack conversion and further checking.

1. BYD Super e-Platform announcement

2. Reuters on BYD 1,000 kW charging platform

3. Reuters on BYD flash charging and overseas rollout

4. EUR-Lex AFIR summary

5. IEA Global EV Outlook 2025 - charging infrastructure

6. IEA Global EV Outlook 2025 - executive summary

7. European Alternative Fuels Observatory

8. ICCT on European public charging points

9. Reuters on Spark Alliance ultra-fast charging network

10. Reuters on ACEA charging gap

11. European Court of Auditors grid review 2025

12. Eurelectric on distribution connection queues

13. Eurelectric Distribution Grids Handbook

14. ACEA European EV Charging Infrastructure Masterplan

15. NREL corridor fast-charging economics study

16. NREL assessment of DC fast charging viability

17. Muratori et al. on DC fast-charging electricity rates

18. NREL future demand for EV charging infrastructure

19. Nature Communications battery thermal gradients

20. Applied Energy fast charging degradation and safety

21. SAE J3271 Megawatt Charging System

22. CharIN Megawatt Charging System

BloombergNEF’s latest gas-plant cost signal is striking: US combined-cycle gas plant project costs rose to about $2,157/kW in 2025, up from less than $1,500/kW in 2023. Bloomberg’s reporting also notes that US utilities filed for almost 24GW of gas-fired capacity in 2025, a sharp jump from 2023, while build times increased by around 23%.

At first glance, this looks like the obvious market response to rising electricity demand. Data centres, new factories, electrification, coal retirements and stretched grids all make firm capacity more valuable. Gas is dispatchable, familiar to utilities, and easier to underwrite than many still-emerging alternatives. In that narrow sense, more gas is not surprising.

But that is not the full story. The more interesting point is that the supposedly practical fallback is now showing constraint symptoms of its own: turbine queues, labour pressure, interconnection delays, pipeline questions, financing strain and ratepayer politics. Gas may still be useful. It may even be necessary in specific nodes. But usefulness is not the same as cheapness, speed or clean cost allocation.

1. The headline cost increase is real enough to matter, but it is not a single-cause story

The 66% figure should not be read as pure turbine inflation. A higher average $/kW can reflect project mix, location, size, permitting complexity, EPC scope, grid interconnection, pipeline-related works, emissions controls and financing assumptions. In other words, the average plant being filed in 2025 may not be directly comparable to the average plant filed in 2023.

Still, the direction is hard to dismiss. Independent market signals point the same way. Reuters reported that planned US gas-fired capacity more than tripled to 252GW in 2025, while combined-cycle capital costs have moved above $2,400/kW and turbine lead times have extended beyond five years. Wood Mackenzie has also described a severe gas-turbine supply-demand imbalance, with turbine prices expected to rise sharply through 2027 as developers try to secure equipment for projects tied to electrification and data-centre demand.

2. Gas still answers the reliability question better than most single alternatives

There is a reason utilities are reaching for gas. Solar and wind may provide low-cost energy, but they do not by themselves provide the same dependable capacity value at the hour and node of need. Four-hour batteries are valuable for peak shifting, reserves and short-duration flexibility, but they do not fully replace multi-day firm generation. Nuclear uprates, geothermal, long-duration storage, demand response and transmission can all help, but each has its own scale, timing and permitting constraints.

That matters because US electricity demand is no longer flat. The EIA expects US electricity use to rise for four consecutive years for the first time since 2007, with the strongest four-year growth period since 2000, driven heavily by large computing centres. In its 2026 outlook, EIA also describes data-centre load as the dominant driver of long-term US electricity growth after a decade-plus demand plateau.

The data-centre numbers explain why the system is reaching for firm capacity. Lawrence Berkeley National Laboratory estimated that US data-centre electricity consumption rose from about 58TWh in 2014 to about 176TWh in 2023, and could reach roughly 325TWh to 580TWh by 2028. That would be a very large new load wedge for a system that had spent much of the previous decade planning around modest demand growth.

3. The bottleneck has moved from electrons to deliverable capacity

The United States is not simply short of annual generation. It is short of deliverable capacity in the places and timelines where load is arriving. That distinction matters. Energy can be abundant in aggregate while reliability remains scarce at specific nodes, hours and interconnection points.

The turbine market makes the point clearly. According to Reuters, developers are ordering earlier, reserving equipment before final project certainty, using different turbine classes and considering gas engines as a bridge because large-frame turbine supply is tight. GE Vernova’s recent results tell the same story from the supplier side: strong demand for gas turbines and grid infrastructure lifted its outlook, with backlog reaching $163bn and expectations of at least 110GW of turbine backlog and reservations by year-end.

That is why “gas is fast” now needs qualification. Gas can be faster than a new nuclear plant or a major greenfield transmission build. But that does not make it unconstrained. If equipment lead times push delivery toward the early 2030s, then gas is not just the practical option; it is another infrastructure supply chain competing for scarce equipment, labour and grid access.

4. The geography matters more than the national capacity number

A national 24GW figure hides different local stories. New gas is most valuable where load growth is concentrated, transmission is constrained, coal retirements are material, or data centres require high-availability power. A plant replacing retiring capacity is not the same as a plant built around new data-centre load, and neither is the same as defensive utility reliability planning.

The current US map is not evenly distributed. Bloomberg’s project map points to clustering across the Midwest, Southeast, Texas-adjacent regions and parts of the East. Recent company updates reinforce that the demand shock is localised: NextEra has discussed large gas-fired data-centre power projects in Pennsylvania and Texas that could total nearly 10GW, while CenterPoint has cited more than 12GW of committed industrial load and plans to deliver 8GW of electric service to Greater Houston projects by 2029.

Texas shows the interaction between load, transmission and local reliability. Reuters recently reported that ERCOT supply margins are expected to remain squeezed until grid expansions arrive, with major transmission additions not expected until after 2030. That creates a near-term window in which gas looks attractive not because it is universally optimal, but because it can be sited near demand or near gas infrastructure while transmission catches up.

A better way to classify the gas buildout

5. The real political fault line is cost allocation

Higher gas capex becomes politically explosive when the cost is socialised. In a regulated utility model, capital invested in plant and equipment can enter the rate base. Customers then pay for depreciation and an allowed return through rates, subject to regulatory approval. NARUC materials describe the core ratemaking logic: a return is provided on rate base, and the cost of capital is multiplied by rate base to determine the revenue requirement recovered from customers.

That is why the type of gas project matters. If a plant replaces retiring system capacity, broad cost recovery may be defensible. If a plant is built primarily because a cluster of data centres or industrial loads arrives faster than the grid can absorb, then the allocation question becomes much sharper. Are households and small businesses paying for system reliability, or are they underwriting the backup capacity of a concentrated new load class?

This is not a theoretical problem. Policymakers and regulators are already debating how large loads should interconnect and how costs should be assigned. The Department of Energy has framed data-centre electricity demand as a major near-term system issue, while EPRI estimates that data centres could grow to consume up to 9% of US electricity generation annually.

6. Gas can lower adequacy risk while raising affordability and fuel-risk exposure

New gas can reduce blackout risk. It provides dispatchable capacity, improves reserve margins and gives system operators a controllable resource. That is valuable in a system facing fast load growth, interconnection backlogs and uncertain demand geography.

But reliability risk is not the only risk. New gas also brings fuel-price exposure, pipeline dependence, emissions risk, stranded-asset risk and bill risk. If the plant is rate-based, customers may carry the capital cost. If gas prices rise, customers may also carry fuel adjustment costs. If demand forecasts disappoint or data-centre load proves more flexible or migratory than expected, the system may be left with long-lived assets built around a temporary or overstated demand shock.

This is the trade-off the headline cost number forces into the open: gas may remain the most bankable firm-capacity tool in many US regions, but the full cost of using it is no longer just a fuel-cost question. It is a capital-cost, deliverability and allocation question.

The better question

The gas rush should not be analysed as one national story or as a simple fuel-choice debate. The useful lens is more specific: what constraint is each project solving, how deliverable is it, and who ultimately carries the cost?

That means asking:

• Is the project replacing retiring capacity, serving concentrated new load, relieving a local grid constraint, or being built as defensive planning against uncertain forecasts?

• Is gas still the lowest-regret firm-capacity option once dependable capacity value, build time, fuel risk, emissions exposure and local network constraints are compared properly?

• Can the plant actually be delivered on the timeline implied by the demand forecast, or is the “fast” option already running into its own queue?

• Who created the incremental need, and does the tariff, contract or rate-base structure allocate the cost to the customers driving it?

• Which risks are being reduced ( adequacy, reserve margin, local reliability), and which risks are being shifted onto customer bills, such as capex recovery, fuel volatility and stranded-asset exposure?

Conclusion: the fallback option is being repriced

The BloombergNEF cost signal should not be read as an argument that new gas is irrational. In many US regions, gas will still be part of the practical answer to near-term reliability needs. It is dispatchable, familiar, financeable and often easier to site than alternatives that require large new transmission, long-duration storage or still-scaling technologies.

But the same signal does challenge the simplistic version of the gas argument. Gas is not magically outside the infrastructure economy. It needs turbines, skilled labour, permits, pipelines, interconnection, financing and regulatory approval. When everyone reaches for it at once, those inputs become scarce and their cost rises.

The final lesson is therefore more precise than “build more gas” or “do not build gas”. The US power system is being forced to distinguish between energy, capacity and deliverability. The question is not just how much generation is added, but whether the right kind of firm capacity is built in the right places, on the right timeline, and paid for by the right customers.

Gas can still reduce adequacy risk. But if expensive new gas is built for concentrated demand and recovered broadly through customer bills, reliability risk falls while affordability and allocation risk rise. That is the real analytical fault line.

References and source notes

1. Bloomberg Green / BloombergNEF, “Cost to Build a Natural Gas-Fired Power Plant Surges 66%” (23 April 2026).

2. EnergyNow syndicated summary of BloombergNEF gas-plant cost figures.

3. Reuters, “Power developers adapt gas turbine strategies to mitigate tight supply” (2 March 2026).

4. Wood Mackenzie, “Gas turbine prices soar 195% as market faces supply-demand crisis” (1 April 2026).

5. US EIA, “EIA forecasts strongest four-year growth in U.S. electricity demand since 2000” (13 January 2026).

6. US EIA, Annual Energy Outlook 2026 release (8 April 2026).

7. Lawrence Berkeley National Laboratory / DOE, 2024 United States Data Center Energy Usage Report.

8. Reuters, “GE Vernova lifts 2026 outlook as AI boom fuels power equipment demand” (22 April 2026).

9. Reuters, “NextEra expects agreements on Japan-backed gas-fired data center projects” (23 April 2026).

10. Reuters, “CenterPoint Energy’s quarterly profit rises on surging data center demand” (23 April 2026).

11. Reuters, “Texas power supply margins squeezed until grid expansions kick in” (21 April 2026).

12. NARUC, Revenue Requirements, Rate Base and Cost of Capital.

US Department of Energy, Clean Energy Resources to Meet Data Center Electricity Demand.

Markets are often treated as a lagging reflection of the energy transition, reacting to deployment, policy, and cost trends already in motion. Most of the time, that description is broadly right. But occasionally, markets move first, repricing not just what is being built, but how the system itself is understood.

The divergence across nuclear, wind, and solar equity indices over the past few years is one such moment. What appears at first glance as sector rotation reads more convincingly as a structural reassessment of the transition itself from a narrow focus on cost per megawatt-hour to a broader valuation of system role, policy exposure, and industrial viability.

This piece examines what the indices are signalling, what the underlying physical and financial data reveal, and why the transition narrative may be entering a new phase, one defined not by which technology is cheapest, but by which delivers the most value to the system as a whole.

1. What These Indices Represent (and What They Do Not)

The nuclear, wind, and solar indices tracked by Trading Economics are not measures of generation, deployment, or system cost. They are global public equity baskets composed of listed companies with exposure to these sectors. As such, they reflect capital allocation, investor expectations, financing conditions, and policy exposure rather than physical system performance.

The nuclear index largely tracks uranium producers, fuel-cycle firms, and nuclear infrastructure and services companies, such as Cameco, Uranium Energy Corp, and firms tied to reactor construction and maintenance. The wind index captures turbine manufacturers such as Vestas and Nordex, developers, and supply chain participants. The solar index is heavily exposed to manufacturing and global supply chains, with a notable concentration in Chinese producers including JinkoSolar, Trina Solar, and LONGi Green Energy.

These are therefore heterogeneous business models rather than clean technology proxies. They do not directly indicate which technology is scaling fastest or delivering the lowest cost. Instead, they reveal what markets are willing to fund, and under what constraints.

2. What the Indices Show

Three distinct patterns emerge over the past cycle.

Nuclear equities have undergone a sustained re-rating, with the index rising from the low-to-mid teens in 2023 to above 50 by early 2026. This reflects increasing value placed on firm, policy-backed, long-duration capacity. The nuclear renaissance is being driven by converging forces: data centre operators seeking reliable baseload power have signed large power purchase agreements with nuclear operators, governments have issued supportive executive orders and financing packages, and the IEA reports that nuclear investment has grown 50% over the past five years, with spending on new plants and upgrades expected to exceed $70 billion in 2025. Over 70 GW of new nuclear capacity is under construction globally, one of the highest levels in thirty years.

Wind equities experienced a drawdown driven by cost inflation and project disruptions, followed by a partial recovery. European turbine manufacturers faced a particularly challenging period from 2022 onwards, as supply chain pressures including higher steel and rare earth prices, elevated cost of capital, and protracted permitting timelines compressed margins. Turbine prices in Europe rose roughly 10% between 2021 and 2022, and the weighted average cost of capital in the sector increased from 5.1% to 6.6% by 2024. Recovery has been supported by improved auction design, the EU’s permitting reforms, and record levels of capacity entering auction pipelines in 2024.

Solar equities have shown uneven performance, with strong deployment offset by severe margin compression and industrial overcapacity. China’s solar module manufacturing capacity reached an estimated 1.2 TW by end-2025 — nearly double total global installation demand of around 650 GW. The six leading Chinese manufacturers reported combined losses of $2.8 billion in the first half of 2025, reversing years of record profits, as module prices fell more than 30%. Beijing has since moved to restrict new capacity expansion and encourage industry consolidation.

Taken together, these patterns suggest a shift away from a purely cost-based framework toward a more complex evaluation of system role, reliability, and industrial structure.

3. From Energy Output to System Value

For much of the past decade, the transition narrative has been framed around the levelised cost of energy (LCOE). This implicitly treats electricity as homogeneous; a kilowatt-hour is a kilowatt-hour regardless of when, where, or how reliably it is delivered. In reality, system value depends on more than energy volume.

A more complete representation is:

V = f(E, A, T, R)

where energy (E), availability (A), timing (T), and reliability (R) jointly determine system value. As variable renewable penetration increases, these additional dimensions become more important, and markets appear to be reweighting them accordingly.

This reweighting helps explain the nuclear re-rating. Nuclear plants deliver high availability factors (typically above 90%), generate around the clock regardless of weather, and provide inertia and frequency response services that support grid stability. These attributes become increasingly valuable in systems with high shares of intermittent generation. The IEA’s World Energy Outlook 2025 notes that nuclear is “increasingly seen as a critical part of a secure and sustainable energy mix” as demand surges and the need for reliable, low-emissions baseload grows.

4. The Physical System: Expansion and Constraint

Global electricity generation is expanding across all major sources. According to the IEA’s Global Energy Review 2025, generation grew by over 1,200 TWh in 2024, a 4% increase that significantly exceeded the 2.6% average of the prior thirteen years. Renewables and nuclear together accounted for more than 80% of that growth, up from two-thirds in 2023. Solar PV alone added roughly 480 TWh, far exceeding any other single source.

The IEA’s Electricity 2025 report forecasts that renewables will surpass coal as the largest source of global electricity generation in 2025 or 2026, a historic milestone. Nuclear power generation is on track to reach a new record high in 2025, driven by restarts in Japan, new reactors in China, India and Korea, and robust output in France and the United States. The system is additive rather than substitutive: fossil fuels remain material even as clean energy scales.

However, this expansion introduces integration challenges. The constraint is shifting from generation capacity to system balancing, flexibility, and grid infrastructure. The IEA’s Electricity Mid-Year Update 2025 highlights the growing importance of managing periods when combined wind and solar output is very low, so-called “Dunkelflaute” events, which led to extremely high wholesale prices in Northern Europe in late 2024 and early 2025.

5. Regional Evidence

In Europe, high renewable penetration has led to increased curtailment and more frequent negative pricing events. Aurora Energy Research reports that technical curtailment exceeded 10 TWh across Europe in 2024, with over 1,000 GW of renewable capacity awaiting grid connection approval. Negative price hours surged in 2025, with Spain, the Netherlands, and Germany each recording more than 500 hours of sub-zero wholesale prices. In Germany, where renewables accounted for 59% of electricity generation in 2024, grid congestion management costs reached €2.9 billion, and roughly 3.5% of renewable output could not reach consumers. These are symptoms of a system where generation has outpaced the infrastructure needed to deliver and absorb it.

In the United States, the picture is shaped by surging electricity demand. The US Energy Information Administration projects total consumption rising to over 4,260 billion kWh by 2026, driven largely by data centres and AI workloads. S&P Global estimates US data centre grid-power demand will nearly triple to 134 GW by 2030. This has elevated the value of firm, dispatchable capacity and driven hyperscalers, Amazon, Google, Meta, and Microsoft, to sign multi-gigawatt nuclear power purchase agreements. The nuclear sector has responded with reactor restart plans, SMR development, and executive orders aimed at quadrupling US nuclear generating capacity by 2050.

In China, rapid solar deployment is accompanied by grid constraints and continued dominance in manufacturing. China’s installed solar capacity surpassed 1,230 GW by early 2026, yet manufacturing capacity of 1.2 TW vastly exceeds global demand. The government’s mid-2025 subsidy reforms led to a sharp decline in new installations, and Beijing has moved to restrict new manufacturing capacity and encourage consolidation. Meanwhile, China’s nuclear programme continues to lead globally, with six reactor construction starts in 2024 alone, and the country accounts for nearly half of all nuclear capacity under construction worldwide.

Across regions, the system is scaling, but integration is becoming the binding constraint.

6. Capital Allocation and Investment Trends

Global energy investment reached a record $3.3 trillion in 2025, according to the IEA’s World Energy Investment 2025 report. Clean energy technologies, renewables, nuclear, grids, storage, low-emissions fuels, efficiency, and electrification attracted $2.2 trillion, roughly twice the $1.1 trillion directed to fossil fuels. Electricity sector investment alone reached $1.5 trillion, some 50% higher than total spending on bringing oil, gas, and coal to market.

Solar PV remains the single largest investment category at $450 billion. But the composition of clean energy investment is broadening. Nuclear investment has grown substantially, with spending expected to exceed $70 billion. Grid investment, at around $400 billion, remains insufficient relative to the surge in generation capacity, a persistent bottleneck flagged by the IEA and industry analysts alike. Battery storage investment reached $66 billion in 2025, reflecting the growing need for system flexibility.

China now accounts for almost a third of global clean energy spending, nearly as much as the US and EU combined. A decade ago, fossil fuel investments were 30% higher than electricity investments; today those positions are reversed. The IEA describes this as the onset of the “Age of Electricity”, with electricity consumption growth driven by electrification, cooling, data centres, and artificial intelligence.

7. The Political Economy of the Transition

Energy has re-emerged as a security issue, with greater emphasis on domestic capacity and resilience. The geopolitical landscape has shifted markedly since Russia’s invasion of Ukraine, which exposed Europe’s vulnerability to gas supply disruption and accelerated investment in renewables and energy efficiency across the continent.

Industrial policy is now central to the transition. The US Inflation Reduction Act provided hundreds of billions in tax incentives for clean energy, reshaping the investment landscape and spurring record deployment of solar, storage, and EVs. In May 2025, the Trump administration issued executive orders aimed at expanding nuclear capacity, while the One Big Beautiful Bill Act signed in July 2025 shortened eligibility for solar and wind tax credits. The EU’s Net-Zero Industry Act and Critical Raw Materials Act seek to build domestic manufacturing capacity and reduce dependence on Chinese supply chains for key technologies.

Trade dynamics are shaping outcomes. The EU opened an investigation into Chinese wind turbine suppliers under the Foreign Subsidies Regulation in 2024. Anti-circumvention rules in the US target Chinese-affiliated solar production routed through Southeast Asia. China controls roughly 80% of global neodymium supply, a critical input for wind turbine magnets, creating strategic vulnerability for Western manufacturers. Chinese solar manufacturers’ overcapacity crisis has triggered a $7 billion state-backed consolidation fund, while simultaneously driving module prices to levels that undercut competitors worldwide.

The transition is evolving from technology diffusion into strategic industrial competition. Wind and solar outcomes are now closely tied to policy design, trade dynamics, and supply chain geography.

8. Implications

Three shifts follow from this analysis.

First, value is moving from cost alone to system functionality. The LCOE framework, while useful, is increasingly insufficient. Investors are pricing in dispatchability, grid services, and policy alignment alongside generation cost. Nuclear’s re-rating is the clearest expression of this shift, but it is also visible in the growing investment in storage, grid infrastructure, and demand-side flexibility.

Second, the focus is shifting from deployment to integration. Adding generation capacity is necessary but no longer sufficient. The binding constraints are now grid connection queues, transmission bottlenecks, curtailment, and the management of temporal mismatches between supply and demand. Europe’s record negative pricing hours and over 1,000 GW of projects awaiting grid approval illustrate the scale of the integration challenge.

Third, outcomes are increasingly shaped by policy frameworks rather than purely market forces. The IRA, the EU’s industrial policy package, China’s capacity management interventions, and US nuclear executive orders are all reshaping the investment landscape. The question is no longer simply which technology is cheapest, but which industrial strategy, regulatory framework, and supply chain architecture can deliver a reliable, affordable, and secure electricity system.

9. A Necessary Caution

These indices are indicators of sentiment and capital allocation. They are not direct measures of system performance or technological superiority. Equity markets are subject to momentum, narrative cycles, and speculative excess, as the sharp corrections in nuclear-adjacent stocks like Oklo and NuScale Power in late 2025 demonstrated, when both fell 50% from their highs despite neither company having built or operated a reactor. The indices should be interpreted as signals rather than definitive conclusions: they reveal where capital is flowing and what investors are pricing in, but they do not settle questions of long-term system design.

10. Conclusion

The divergence across nuclear, wind, and solar equity indices reflects a broader repricing of the energy transition. Markets are increasingly valuing system function, policy alignment, and industrial structure alongside cost.

Global energy investment has reached record levels, with clean energy attracting twice as much capital as fossil fuels. But the composition of that investment is shifting toward grids, storage, nuclear, and system-enabling infrastructure, not just generation capacity. The physical system is expanding, but integration challenges are mounting: curtailment, negative pricing, grid queues, and the mismatch between generation and delivery infrastructure.

The transition is no longer defined solely by how much energy is produced, but by how effectively it is integrated and delivered within the system. The indices are telling us that markets have begun to price this in.

References

1. IEA (2025), World Energy Investment 2025, IEA, Paris. Licence: CC BY 4.0.

2. IEA (2025), Global Energy Review 2025: Electricity, IEA, Paris. Licence: CC BY 4.0.

3. IEA (2025), Electricity 2025, IEA, Paris. Licence: CC BY 4.0.

4. IEA (2025), Electricity Mid-Year Update 2025, IEA, Paris. Licence: CC BY 4.0.

5. IEA (2025), The Path to a New Era for Nuclear Energy, IEA, Paris.

6. IEA (2025), Renewables 2025, IEA, Paris. Licence: CC BY 4.0.

7. IEA (2025), World Energy Outlook 2025, IEA, Paris. Licence: CC BY 4.0.

8. Ember (2025), Wind Sector Challenges Are Blowing Over, European Electricity Review 2025.

9. Aurora Energy Research (2026), European Renewables Market Overview Report.

10. S&P Global (2025), Data Center Grid-Power Demand to Rise 22% in 2025.

11. Perkins Coie (2026), Nuclear Industry Kicks Off 2026 With Major Public and Private Sector Announcements.

12. Slaughter and May (2026), Horizon Scanning: Nuclear Energy.

13. OPIS (2025), Solar Photovoltaic Overcapacity Sparks Market Turmoil in China.

14. Deloitte (2025), 2026 Power and Utilities Industry Outlook.

15. BloombergNEF (2025), Global Cost of Renewables to Continue Falling in 2025.

16. REN21 (2025), Global Status Report 2025: Wind Power.

17. Fortune (2026), Negative Prices for Electricity Are Getting More Common in Europe.

18. Grid Strategies (2025), Power Demand Forecasts Revised Up for Third Year Running.

19. Trading Economics, Energy Indices. Nuclear, Wind, and Solar Energy Index data.

1. The claim and the distinction that matters

The current reporting around Ed Miliband’s expected announcement presents the move as an attempt to “delink” gas and electricity prices. Yet the details reported so far point to something narrower. The government appears to be considering two connected moves: first, pressing older low-carbon projects, especially those under the Renewables Obligation, towards fixed-price contracts similar to Contracts for Difference; and second, tightening the Electricity Generator Levy to raise short-run funds while consultation proceeds on wider reform (Guardian, 2026; Financial Times, 2026).

That is major news. But it does not follow that the UK has found a way to remove gas from the underlying economics of the power system. The crucial distinction is between price formation, generator remuneration, and retail bill pass-through. Those are related layers, but they are not the same layer. Much of the present debate collapses them into one.

The framework in my own recent paper starts from exactly that problem. Marginal-cost pricing is not defended there as a slogan or a perfect institution, but as a mechanism embedded in deeper system requirements: demand-supply balancing, efficient resource allocation, reliability and security, transparency and fair competition, and affordability (Delardas, 2025). If that framework is applied to the current UK discussion, the reported reform looks less like true delinking and more like a contractual insulation layer built around part of the generation stack.

2. Why this is not delinking in the strong sense

To delink in the strong sense, gas would have to stop doing the work that makes it relevant to electricity pricing. It would have to stop being a frequent marginal source in tight hours, stop shaping scarcity in balancing, stop influencing the residual market that clears when intermittent output falls short, and stop feeding through to tariffs indirectly via supplier costs. Nothing in the reporting so far establishes that (Guardian, 2026; Financial Times, 2026).

What is actually being discussed is much closer to a rearrangement of revenue structures. Legacy low-carbon assets would be moved, or nudged, onto fixed-price contracts. Their income would then be tied less to live wholesale prices and more to administratively set strike prices. That changes the remuneration of a block of generation. It does not, by itself, abolish the residual need to price the last flexible unit required to keep the system whole (Delardas, 2025).

This is why the language matters. If the wholesale market remains marginal at the point where system scarcity is resolved, and if gas still provides a significant share of flexible backup when renewables underperform, then gas remains system-relevant even if a larger share of renewable output sits behind fixed contracts. The reform may change who is exposed to volatility. It does not automatically change what the system still needs to be priced.

3. The closest analytical analogue is a green power pool, not a clean break

In the paper, I reviewed proposals such as split markets and green power pools that seek to separate low-carbon generation from the ordinary marginal-pricing environment. Their appeal is easy to understand. If renewable and nuclear output can be financed on long-term fixed terms, while a smaller dispatchable segment addresses the residual balancing problem, then consumers appear less exposed to fossil fuel volatility (Delardas, 2025).

But that appearance is not equivalent to a system-level resolution. Once low-carbon assets are moved into fixed-price contracts, the remaining dispatchable market becomes even more important because it carries a larger share of scarcity, adequacy, and balancing responsibility. If gas remains central within that residual market, then the system has not escaped the gas-power link. It has concentrated it.

This is why I do not regard the reported move as wholesale reform in the strong sense. It is better understood as a hybrid: a larger administered renewables block on one side, and a residual market still solving the hard real-time problem on the other. In the paper, that family of designs is treated as carrying real trade-offs rather than offering a clean exit from marginal pricing (Delardas, 2025).

There is also a practical political reason this path is attractive. It allows ministers to say they are weakening the gas link without immediately rewriting all wholesale settlement, reserve procurement, balancing logic, or retail tariff design. That makes it easier to announce. It does not make it equivalent to full delinking.

4. CfDs are not a free pass around system economics

Contracts for Difference are well-established instruments. The UK Parliamentary Office of Science and Technology summarises them as contracts that support low-carbon build by paying or receiving the difference between a strike price and a reference market price (POST, 2026). Low Carbon Contracts Company explains that the scheme is funded by electricity suppliers through the Supplier Obligation Levy (LCCC, 2026).

That structure is important because it makes clear what CfDs do and do not do. They do not remove the need for a market reference price. They do not remove the need to value balancing or reserves. They do not eliminate the supplier-side cost stack. They provide a contractual rule for redistributing revenue in line with market outcomes.

In a high-price world, that can look very attractive. When the market reference price sits above the strike, CfD-backed generators pay back the difference. That can soften the exposure of suppliers and, eventually, consumers to extreme wholesale spikes (POST, 2026; LCCC, 2026). But the mirror image also matters. When market prices fall below strike, the payment flow reverses. The public or supplier side then supports the generator instead.

So CfDs are not a one-way bill-cutting machine. They are closer to a hedge. In one state of the world, they claw money back. In another state of the world, they create a payment obligation. That asymmetry is central to the current UK debate because the government appears to be considering moving more legacy projects into this framework precisely after a period of renewed fossil-price stress (Guardian, 2026; Financial Times, 2026).

5. Why the LCOE issue matters more than it first appears

A deeper problem follows from how fixed-price contracts are set. In my analysis, I argued that LCOE-based or cost-based remuneration can be useful for project support, but it is an incomplete guide to actual system value. It does not fully capture scarcity, balancing requirements, integration costs, locational conditions, or the dynamic value of flexibility (Delardas, 2025).

That matters here because the more the UK shifts legacy output onto fixed-price contractual terms, the more it risks widening the gap between how a technology is remunerated and how the system actually values it at the margin. In a calm political setting that can remain hidden. In a stressed system, it reappears through other channels: balancing costs, reserve procurement, capacity payments, curtailment, congestion management, or simply a more expensive residual market.

This is one reason I do not read the present move as an escape from gas-linked power economics. It is at most a partial relocation of those economics. A state-backed contract may mute one revenue exposure while leaving the underlying physical dependence on flexibility intact. If anything, the residual flexible fleet may become more visibly system-critical under this structure.

6. Bills are a different layer again

The public debate often leaps straight from headlines about wholesale reform to household bill promises. That leap is analytically weak. I have previously stressed that average consumer tariffs do not simply mirror fluctuating wholesale prices and also include transmission, distribution, and other centrally managed system costs (Delardas, 2025).

Ofgem’s current methodology makes the same point in practical regulatory form. Its 2025 call for input locates the CfD allowance within the default tariff cap’s wholesale cost allowance (Ofgem, 2025), and its 2026 technical approach note states explicitly that the CfD allowance accounts for costs associated with the Supplier Obligation Levy (Ofgem, 2026).

That means CfDs do not directly rewrite retail tariffs. They work through supplier costs. Only then, with timing, regulation, hedging, and tariff design in the way, do they feed into bills. The retail effect is therefore indirect. It can smooth volatility in some periods, but it does not guarantee structurally lower bills in all periods.

This point is even sharper now because Reuters reported that some renewable-policy costs have already been shifted to general taxation as part of recent bill-easing measures (Reuters, 2026). That is a reminder that the UK bill debate is already one of reallocation as much as cost elimination. Charges can move between wholesale allowance, policy support, taxation, and other pass-through channels. They do not disappear merely because the label changes.

7. The reversal risk is real

A high-gas environment makes fixed-price, low-carbon contracts politically attractive because they appear to offer protection against fossil fuel price spikes. But that same design can become politically awkward when market prices normalise or fall. The reason is simple: strike prices that looked prudent when gas was expensive can look burdensome when gas is cheaper and the reference price falls below them.

This is not a theoretical quirk. It is built into the contract logic itself. The same mechanism that generates paybacks in one state of the world generates top-up payments in another (POST, 2026; LCCC, 2026). The political promise, however, is usually sold only in the first state.

That creates a risk of future disillusionment. If the government now signs or compels more legacy assets into high strike-price arrangements in the name of protection from gas, it may later discover that the policy is criticised for imposing its own burden when the market turns. In other words, the policy can reverse. That is a central reason why I do not treat it as delinking. A true delinking would not depend so heavily on which side of the strike the market happened to land.

8. The flexibility problem does not go away

The hardest point in all of this is also the most basic. Intermittent renewables do not abolish the requirement for flexibility. They change its role, timing, and remuneration problem. Proposals that shield low-carbon generation from live price signals can still leave the system needing adequate flexible plant, responsive balancing, and investment incentives that are difficult to replicate through administrative means alone (Delardas, 2025).

This is exactly where the present UK idea may run into strain. If a larger share of low-carbon output moves outside direct merchant exposure, the residual market may become thinner but more important. A thinner market can be more volatile, more exposed to scarcity, and more dependent on a smaller fleet of assets whose economics are harder to stabilise. If those assets are gas, then gas can still matter a great deal even if ministers claim a weaker link overall.

The risk is not simply that gas still sets price in many tight periods. It is that the remaining flexible segment becomes a bottleneck around which the rest of the system must be organised. At that point the appearance of delinking in one layer may coexist with greater dependence in another.

9. What full delinking would really require, and why that is mostly a test rather than a proposal

It is useful to define the strong version of delinking, but only as an analytical test. If one really meant full delinking, the UK would need to stop gas from doing more than setting revenues for some renewable assets. Gas would have to stop determining marginal scarcity in the residual market, stop providing the reference against which fixed-price contracts settle, stop anchoring balancing decisions in tight hours, and stop feeding through indirectly into supplier costs and tariffs.

That would imply something much more radical than the reported policy. It would require a redesign of wholesale settlement, balancing procurement, reserve, and adequacy arrangements, and retail pass-through architecture. It might even require holding gas outside the market in some strategic reserve role, a possibility mentioned in current reporting but still clearly at the edge of the discussion (Guardian, 2026).

I do not set that out here as an advocated destination. Quite the opposite. Under the logic of my paper, the more seriously one tries to imagine full delinking, the more one runs into the underlying constraints of the power system itself. Someone still has to balance supply and demand. Someone still has to respond when weather-dependent output falls short. Someone still has to be remunerated for reserve, adequacy, and ramping. Those constraints are not rhetorical. They are structural (Delardas, 2025).

That is why the strong version of delinking is useful mainly as a revealing counterfactual. It shows how much work the current language is doing. Once the term is defined strictly enough to mean genuine separation from gas-linked system economics, it begins to look either radically invasive, highly distortive, or simply infeasible in clean form. The current proposal is narrower precisely because the constraints are real.

10. The real choice the UK is making

The UK is not choosing between a gas-linked system and a fully delinked one. It is choosing how much volatility to reallocate, how much of the generation stack to move onto administered contractual terms, and how much of the remaining flexibility burden to leave in a residual market still shaped by scarcity.

That can be a defensible political choice. After a fossil shock, governments understandably want less household exposure to extreme price spikes. But it should be described honestly. This is not the abolition of gas-linked electricity pricing. It is a reshaping of who bears that linkage, when they bear it, and through which institutional route it arrives.

The risk is that the policy is sold as a clean victory over the old market logic when it is really a redistribution of that logic across contracts, levies, supplier costs, and residual balancing. If that happens, the next reversal will look like a surprise even though it was embedded from the start.

So the right description is more sober. The reported reform may strengthen bill smoothing in some states of the world. It may improve the revenue stability of some legacy low-carbon assets. It may also intensify the importance of the residual flexible market, create future strike-price overhangs, and increase the role of administrative management. Those are not incidental side effects. They are the core trade-offs.

That is why I do not regard the current package as delinking. It is not a clean break from gas-powered price formation. It is a different contractual overlay on top of a system that still has to obey the same underlying constraints.

References

Delardas, O. (2025). Transforming marginal-cost pricing: A blockchain approach to fundamental energy system dynamics. Utilities Policy. DOI: 10.1016/j.jup.2025.101942.

Ofgem (2025). Energy price cap methodology: Contracts for difference review.

Ofgem (2026). Energy price cap: technical approach to Market Wide Half-Hourly Settlement.

Low Carbon Contracts Company (2026). About the CfD scheme.

HM Government (2023). Electricity Generator Levy: introduction.

Parliamentary Office of Science and Technology (2026). Contracts for Difference and the economics of renewable energy deployment.

Ambrose, J. and Pratley, N. (2026, 17 April). Rachel Reeves to raise windfall tax on low-carbon electricity generators. The Guardian.

Millard, R., Armstrong, A. and Pickard, J. (2026, 20 April). Ed Miliband unveils move to delink gas and electricity prices. Financial Times.

Reuters (2026, 25 February). Most Britons’ energy bills to fall after regulator cuts cap.

The EV-versus-ICE argument is now one of those debates where the volume of commentary has become a substitute for clarity. There is no shortage of information. There is, however, a shortage of coherence. One stream treats the electric vehicle as a moral emblem. Another treats it as a policy artefact. A third reduces the whole matter to range anxiety, charging queues, or the latest quarterly sales wobble. Somewhere underneath all of that, the actual engineering story is often lost.

That is a mistake, because private transport is one of the most consequential technological systems ordinary people live inside. It shapes how labour is organised, how cities expand, how households budget, how states tax energy, how supply chains are built, and how people experience autonomy in daily life. The right starting point, then, is neither political enthusiasm nor political backlash. It is the machine itself.

If one strips away the signalling and returns to first principles, the subject becomes both more interesting and more demanding. The relevant question is not simply whether electric vehicles will ‘win’. It is what problem each architecture is trying to solve, how well it solves it, and which trade-offs are being shifted rather than eliminated. The most useful way into the issue is therefore bottom-up: begin with what makes a good automobile, move through the physical logic of propulsion, then work outward into the historical system that made the internal combustion engine dominant and the conditions under which battery-electric drive is now challenging it. That bottom-up approach is also the best defence against fractured or misleading narratives, because it forces the analysis back onto conversion efficiency, torque delivery, transmission architecture, mass, thermal limits, energy storage, infrastructure, and cost. Those are the terms on which the race is actually being run.

A useful benchmark here is NREL’s FASTSim framework, which models vehicles in terms of mass, drag, rolling resistance, powertrain components, battery or fuel storage, and efficiency behaviour across drive cycles. That is the level at which the discussion needs to be held.

What makes a car good

A good car is not simply one that is fast, fashionable, or low-emission. A good car is one that solves several engineering problems at once without making any one of them intolerable. It must deliver enough wheel torque for useful launch and overtaking, enough sustained power for realistic driving, enough efficiency to keep operating costs and energy demand within reason, enough thermal stability to survive repeated use, enough reliability to justify ownership, and enough ease of refuelling or recharging to make the machine feel like an instrument of agency rather than a source of friction.

In practice, that means the quality of a powertrain cannot be read off a single headline number. Peak horsepower is not the same thing as drivability. Nameplate range is not the same thing as practical mobility. A vehicle is better understood as a system of constrained optimisation. The core trade-offs sit between torque, power, efficiency, mass, durability, thermal management, cost, packaging, and energy replenishment. Every major propulsion architecture - petrol, diesel, hybrid, plug-in hybrid, battery-electric, and fuel-cell electric - is simply a different answer to that optimisation problem.

Once framed that way, a clearer picture emerges. The electric vehicle is not revolutionary because it is new. In one sense it is not new at all. It is revolutionary because it rearranges where the hard constraints lie. The internal combustion car is constrained above all by the inefficiency and narrow operating sweet spot of combustion. The battery-electric car is constrained above all by the mass, cost, charging behaviour, and thermal sensitivity of electrochemical storage. Neither is frictionless. They simply place the friction in different parts of the machine.

How the internal combustion engine works, and why it dominated

The internal combustion engine won the twentieth century not because it was elegant in a pure engineering sense, but because it formed the core of a powerful sociotechnical package. Its operating logic is familiar: a fuel-air mixture is burned in cylinders, pressure acts on pistons, reciprocating motion is converted into crankshaft rotation, and that torque is passed through a clutch or torque converter, gearbox, differential, and driveshafts to the wheels. The machine is mechanically intricate because combustion itself is an awkward prime mover. It produces usable power only within a limited range of engine speeds and loads, and it wastes a great deal of the fuel’s energy as heat.

That waste is not trivial. DOE-backed reviews of the future of the internal combustion engine make the point plainly: conventional light-duty spark-ignition engines typically convert only a minority of the fuel’s chemical energy into useful motion, while decades of research have focused on raising brake thermal efficiency through combustion control, boosting, dilution strategies, exhaust heat management, and aftertreatment integration. Advanced combustion programmes have pursued brake thermal efficiencies as high as 55% in heavy-duty contexts, while more recent work continues to push spark-ignition designs upward. But the historical fact remains: the ICE prevailed for most of automotive history despite its mediocre thermodynamic performance because it was coupled to liquid fuels of exceptional energy density and to a refuelling system that was fast, cheap, and geographically pervasive.

This is why the engine alone does not explain its dominance. The broader automotive system did. Historical work on the automobile industry’s development and studies of technological lock-out show how the ICE became embedded through scale manufacturing, fuel distribution, road building, service networks, consumer familiarity, and policy institutions. Research on early vehicle rivalry is especially useful here because it reminds us that electric vehicles were not absent from the beginning. They were displaced. Early motoring did not begin with a single obvious winner. It settled into one.

The reasons are straightforward. Petrol and diesel carried far more usable onboard energy than the batteries of the period. Refuelling was measured in minutes rather than hours. Long-distance travel was easier. Cold weather, heavy loads, and irregular use were easier to accommodate. Repair skills and parts diffused widely. Once those advantages were reinforced by roads, suburbanisation, and mass production, the ICE car ceased to be just a machine and became the normal form of automobility.

What the EV changes technically

The battery-electric vehicle changes the propulsion chain at a fundamental level. Instead of converting chemical energy into heat, then pressure, then piston motion, then rotational output, it stores energy electrochemically and converts it into torque through power electronics and an electric traction motor. In the simplest form the chain becomes battery to inverter to motor to reduction gear to wheels. A great deal of mechanical mediation disappears.

That simplification matters. Recent reviews of BEV powertrain architecture emphasise that the electric powertrain is built around five tightly coupled subsystems: battery storage, power electronics, the electric machine, the transmission or reduction gear, and the control strategy that coordinates them. The important point is not that the EV has fewer parts in some abstract sense, but that it dispenses with entire classes of components whose job in an ICE car is to manage the engine’s narrow efficient operating band. That is why many BEVs can use a single-speed transmission without feeling compromised in ordinary driving.

The electric machine itself also behaves differently from an engine. It can deliver very high torque from zero or near-zero speed, which is why EVs often feel immediate in a way that conventional cars do not. The U.S. Department of Energy’s EV primer makes this consumer-facing point directly, but the engineering version is more precise: wheel torque is available with less delay because the motor does not need to build revs or wait for a gearbox to put the engine into a favourable part of its torque curve.

There are several major motor families in the EV field - permanent-magnet synchronous machines, induction motors, and reluctance-based designs among them - and the choice between them turns on efficiency, power density, cost, rare-earth dependence, thermal behaviour, and controllability. Recent work on motor efficiency maps is useful because it shows that the EV advantage is not merely about peak values; it is about how efficiently the machine behaves across the torque-speed plane encountered in real driving.

This is where the conversation usually improves. The EV revolution is often described as a battery story, but it is just as much a control and power-conversion story. A modern EV is effectively a carefully managed electromechanical system whose performance depends on inverter behaviour, software calibration, thermal constraints, regenerative braking strategy, and the battery management system’s willingness to deliver or accept power at any given moment. In other words, the EV is not simply ‘a car with batteries’. It is a different architecture of motion.

Torque, power, transmission, and the feeling of the machine

The language of torque and power is often misused in public debate, but it matters because it maps directly onto how the vehicle feels. Torque is rotational force. Power is the rate at which work is done. A car that feels energetic from a standstill usually has strong wheel torque at low speed. A car that continues to pull hard at higher speed needs power. The distinction matters because EVs and ICE vehicles tend to excel in different parts of that envelope.

An electric motor’s ability to provide strong low-speed torque is one reason BEVs feel unusually responsive in urban use. The car moves cleanly, without the elastic delay between throttle input, engine response, and gear selection that characterises many conventional drivetrains. A combustion engine, by contrast, often relies on gearing to multiply torque and on higher engine speed to reach more favourable power output. That dependence on gearing is precisely why transmissions became so central to ICE vehicle design.

This is also why transmission architecture cannot be treated as a side note. Recent research on two-speed BEV transmissions shows that adding gears to an electric drivetrain can improve performance or efficiency in specific applications, especially where the trade-off between launch, gradeability, cruising efficiency, and top speed becomes more demanding. But the fact that such gearboxes are optional in many passenger EVs tells you something important: the electric machine’s usable operating range is much broader than that of a combustion engine.

Efficiency follows the same logic. Comparative efficiency assessments consistently show that battery-electric drivetrains convert a much larger share of input energy into motion than internal combustion vehicles do. That is not a rhetorical advantage. It is a system one. It means less energy is wasted in the conversion step. It is one of the deepest reasons EVs perform so well in stop-start city driving, where conventional engines suffer from idling, transient inefficiencies, and braking losses.

Regenerative braking sharpens this contrast. A combustion vehicle mostly throws away kinetic energy as heat when it slows down. An EV can recover part of that energy back into the battery. The exact amount varies with speed, state of charge, traction limits, and system calibration, but the principle is transformative. It turns deceleration from a pure loss into a partial recovery event.

Where the EV remains constrained

None of that means the electric vehicle has abolished trade-offs. It has relocated them. The limiting component is now the battery pack. The pack carries the vehicle’s energy, but it is costly, heavy, sensitive to temperature, and slower to replenish than a tank of liquid fuel. The engineering challenge therefore shifts from making combustion tolerable to making storage and charging good enough.

That shift is visible in three places. The first is mass. Batteries are heavy, and although the electric drivetrain is efficient enough to compensate impressively in many use cases, the mass penalty does not disappear. The second is charging power and time. A battery can only accept charge at rates the chemistry and thermal system will tolerate, which is why rapid charging is as much a cooling problem as an electrical one. The third is ambient temperature. Recent range studies show how cold and hot conditions reduce usable range through a combination of lower usable battery energy, HVAC demand, and powertrain losses. This is not a niche inconvenience. It is a structural part of the technology.

A further complication is that motorway driving is a harsher environment for EVs than city driving. At high speed, aerodynamic drag dominates, regenerative opportunities shrink, and the battery must supply sustained power rather than short bursts. That is why urban economics often flatter EVs more than intercity ones. It is also why the argument should never be framed as if one duty cycle represents the entire market.

This is where serious analysis parts company with cheerleading. The question is not whether the EV is technically impressive - it plainly is. The question is where its present constraints still bite hardest. Long-distance irregular travel, weak charging geographies, high-load towing, severe temperatures, and ownership models without reliable home charging remain the most important friction points.

Why cars came to mean independence

The political and emotional charge surrounding this issue is easier to understand once one recognises that the car has never been only a transport device. It became a lived symbol of independence because its technical properties allowed a particular kind of freedom to feel normal. Energy was stored onboard in dense form. Range was substantial. Refuelling was fast. Routes were flexible. The driver did not have to negotiate a timetable or rely on shared infrastructure in the same way public transport users did.

Scholarly work on automobility and liberal freedom helps here because it shows that the car’s association with autonomy is not merely cultural ornament. It was stabilised by institutions, landscapes, and habits that made private vehicular mobility feel like an extension of personal choice. Research on Britain’s post-war expansion of automobility tells the same story in historical form: roads, suburbs, commerce, and social expectations adapted to the car until the car came to seem natural.

This matters because the EV does not simply enter a neutral field. It enters a field already structured by that older ideal of freedom. For some drivers, especially those with off-street parking and predictable daily mileage, the EV can strengthen that sense of autonomy. The vehicle is charged at home, mechanical maintenance is lower, and the driver begins each day with an effectively ‘full’ car. For others, especially those without home charging or those whose travel is irregular and long-range, the EV can feel like a partial retreat from the freedom the car historically promised. That perception is not irrational. It is an infrastructural judgement disguised as a cultural one.

Where the real competition now sits

Once the story is rebuilt from engineering realities, the EV-versus-ICE race looks less like a direct technological succession and more like a contest between bottlenecks. The electric drivetrain is superior at converting onboard energy into useful wheel torque with high efficiency, smoothness, and mechanical simplicity. The liquid-fuel system remains superior at storing large amounts of mobile energy compactly and replenishing it quickly through a mature global network. That is the race in one sentence.

This is why the contest is so uneven across segments. In city cars, suburban commuting, ride-hailing fleets, last-mile logistics, and other predictable duty cycles, the EV’s strengths line up well with real use. In heavier, more irregular, colder, more infrastructure-scarce, or more time-critical use cases, the older system retains advantages that are not merely sentimental. They are engineering advantages translated through infrastructure.

This is also why hybrids matter more than simplistic two-camp narratives allow. They exist because the market is often unwilling to choose between electric drivability and liquid-fuel convenience. From a bottom-up point of view, the hybrid is not a philosophical compromise. It is a technically intelligible response to a world in which charging access, purchase cost, and usage diversity still constrain full electrification.

What the latest data say

The market data now reinforce that engineering reading. The IEA’s Global EV Outlook 2025 reports that global electric car sales exceeded 17 million in 2024 and passed a 20% sales share worldwide. China remained the centre of gravity, with over 11 million electric cars sold and electric vehicles accounting for almost half of all car sales. The same IEA market chapter shows how concentrated the transition remains: China dominates scale, Europe remains a major but more policy-sensitive market, and emerging economies are growing quickly from a much smaller base.

The more recent 2025 and early-2026 data show that the story is still advancing, but not smoothly. In Europe, ACEA’s full-year 2025 registration data show battery-electric cars reaching 17.4% of the EU market, while hybrids climbed to 34.5%. That alone is a useful corrective to lazy narratives. Europe is electrifying, but much of the gain is being mediated through hybridisation, not BEVs alone. ACEA’s 2026 running statistics also point to continued movement in the mix during early 2026.

The United States looks more uneven. EIA’s February 2026 assessment found that battery-electric sales weakened after federal tax credits expired late in 2025, while hybrid sales continued to rise. An earlier EIA market note showed that EVs remained much stronger in luxury segments than in the mass market, again pointing to price and use-case constraints rather than any simple technological verdict.

China remains the decisive market, but even there the picture is no longer one-dimensional. China’s National Bureau of Statistics reported that output of new-energy vehicles reached 16.524 million in 2025, up 25.1% year on year. Chinese official releases during 2025 already showed double-digit growth through the year. Yet short-run 2026 reporting also suggests sensitivity to subsidy changes and market saturation dynamics. Reuters’ January 2026 market report described a global slowdown in registrations driven by weaker China and North America numbers, while its April 2026 report on March registrations showed Europe surging, North America falling sharply, and China softer than before.

In other words, the recent data do not say the transition is failing. They say it is becoming more differentiated. The EV market is no longer a single story of linear ascent. It is a set of regional trajectories shaped by engineering fit, industrial scale, subsidy design, trade policy, fuel prices, and charging infrastructure. That is precisely what one would expect if the contest were being decided by real system constraints rather than by slogans.

What deserves attention going forward

If the aim is to understand the evolution of vehicle propulsion systems in a way that is immune to fashion trends, then attention should be directed to a narrower and more demanding set of variables.

The first is duty cycle. Not all cars solve the same mobility problem, and the transition will be segmented accordingly. The second is battery economics: cost per usable kilowatt-hour, pack integration, charging durability, and how much of the pack’s mass and cost can be justified in a given vehicle class. The IEA’s affordability analysis is important here because it shows how sharply the economics diverge by region and vehicle segment.

The third is charging access. The IEA’s charging chapter makes clear that the quality of the user experience depends not merely on the number of public chargers, but on the relation between public and private charging, grid readiness, charge rates, and housing form. A technology can be efficient at the vehicle level and still constrained at the system level.

The fourth is industrial concentration. The EV transition is also a battle over manufacturing scale, battery supply chains, rare-earth exposure, and cost-down learning curves. China matters so much because it has not merely adopted EVs. It has built the industrial machine around them.

The fifth is the continuing evolution of the ICE itself. The combustion engine is not frozen in 2005. It continues to improve through hybridisation, control sophistication, and incremental efficiency gains. That matters because the EV does not compete against a static incumbent.

Above all, the right question is not whether one should be ‘for’ EVs or ‘for’ ICEs. The right question is which architecture solves which mobility problem better once one counts the whole system: the machine, the energy carrier, the replenishment network, the cost structure, and the lived experience of use.

Conclusion

While the electric vehicle is the stronger propulsion architecture, the internal combustion vehicle is still embedded in the stronger mobile-energy system. That is why the race remains open in some segments even as the engineering case for electric drive becomes increasingly hard to deny.

Seen this way, the EV revolution is real, but it is not magical. It does not abolish the old trade-offs so much as move them. Combustion’s weaknesses in efficiency, complexity, and urban drivability are replaced by battery constraints in mass, cost, temperature sensitivity, charging dependence, and industrial supply. The market outcomes of 2025 and early 2026 are entirely consistent with that reading: strong global momentum, powerful Chinese scale, continued European growth, U.S. unevenness, hybrid resilience, and regional volatility where policy or infrastructure do not align with the underlying engineering.

That is also why the conversation is worth having carefully. Once the politics are stripped out, what remains is not a culture war but a remarkable chapter in the history of private transport technology. It is the story of one dominant architecture that organised the twentieth century, and of a rival architecture now advancing not because it is virtuous by declaration, but because in key respects it is technically better. Whether that technical superiority becomes universal market dominance depends not on rhetoric, but on batteries, charging, cost, and the reorganisation of the wider transport system around them.

References

Recent market and policy sources

- International Energy Agency, Global EV Outlook 2025

- International Energy Agency, Executive summary - Global EV Outlook 2025

- International Energy Agency, Trends in electric car markets

- International Energy Agency, Trends in electric car affordability

- International Energy Agency, Electric vehicle charging

- ACEA, New car registrations in 2025; battery-electric 17.4% market share

- ACEA, Reliable figures and statistics (2026 running data)

- U.S. EIA, EV sales fell as hybrid sales continued to grow in 2025

- U.S. EIA, Hybrid vehicle sales continue to rise as electric and plug-in hybrid shares shift

- National Bureau of Statistics of China, Statistical Communique of 2025

- Government of China / CAAM release, NEV production and sales in first 10 months of 2025

- Reuters, January 2026 global EV sales slowdown report

- Reuters, March 2026 EV registrations and Europe’s surge

Engineering, technology, and historical literature

- NREL, FASTSim technical report

- World Electric Vehicle Journal, Powertrain in Battery Electric Vehicles (BEVs)

- World Electric Vehicle Journal, Two-speed transmission efficiency in BEVs

- Energies, Comparative study of electric machine efficiency maps

- OSTI, Effects of Ambient Temperature on Electric Vehicle Range

- U.S. Department of Energy, Electric vehicles and chargers primer

- Canada Energy Regulator, BEVs are far more fuel efficient than ICE vehicles

- OSTI, The Future of the Internal Combustion Engine

- OSTI, Combustion strategies for increased thermal efficiency

- Cambridge, From Revolution to Revolution: a Changing Automotive Industry

- Cambridge, Institutional lock-out of sustainable transport technologies

- Business History Review, Visions of Transportation and early EV/ICE rivalry

- JSTOR, Automobility, Liberalism, and the Ethics of Driving

- JSTOR, The expansion of automobility in urban Britain, c.1955-70

For years, the renewable transition was argued mostly through the language of build-out, falling unit costs, stronger auctions, and growing deployment. That story was not invented. It reflected real progress. The UK Solar Roadmap says the country had 18 GW of installed solar capacity by Q1 2025, while DESNZ’s Energy Trends March 2026 reports that renewables generated a record 152.5 TWh in 2025, equal to a record 52.5% share of UK electricity generation. The same release notes that wind and solar each generated around ten times as much as they did in 2015.

At the policy level, the ambition has been made explicit. The Clean Power 2030 Action Plan frames the next phase as a decisive scaling of clean electricity, targeting 45 to 47 GW of solar by 2030 alongside major growth in wind, batteries, long-duration storage, interconnection, and consumer-led flexibility. That matters because it shows the official story is no longer just “build more renewables”. It is increasingly “build a system that can live with them”.

But analytically and financially, much of the debate was still dominated by plant-level metrics. LCOE, strike prices, WACC, IRRs, auction clearing prices, and investment certainty became the standard vocabulary of the transition. The IEA’s Projected Costs of Generating Electricity is representative here: it documents the extraordinary fall in renewable generation costs, but it also underscores why plant-level costs are not the same as system value. Even as more nuanced approaches such as value-adjusted LCOE emerged, public debate kept circling back to a single, simplified idea: if renewables are cheap, why should the system still look expensive?

After the 2022 crisis, that simplification hardened into a political reflex. If gas still set the price, then the market design itself was said to be the problem. Calls for decoupling, splitting, shielding or bypassing marginal pricing moved quickly into the mainstream. My own paper was written into that debate. When I first wrote it, I presented it as a theoretical comparison and critique of competing market designs. But in retrospect, it proved to be something broader: almost a manual for how the transition’s tensions were likely to surface in practice. In my 2025 research paper, I argued that many reform proposals misdiagnosed the problem by focusing too narrowly on the pricing formula while underestimating the broader system architecture needed to keep electricity affordable, reliable, transparent, and balanced.

That is why the current NESO story matters so much. NESO’s latest summer announcement does not say the grid is collapsing under “too much solar”. In fact, it says Great Britain’s electricity system is expected to remain secure and reliable this summer. But it also says periods of excess electricity are becoming more common, and that the updated Demand Flexibility Service will reward consumers and businesses for increasing electricity use during periods of excess supply. Ofgem’s March 2026 decision adds another layer: it says DFS delivered roughly £700,000 of balancing cost savings in its first year of operation, suggesting that demand-side turn-up can be an efficient balancing option relative to alternatives.

This is the point at which the debate needs to mature. Paying consumers to use electricity is not proof of failure. But neither is it proof that the system has become simple or cheap in any complete sense. It is a signal that the problem is shifting. The issue is no longer only whether the system can generate low-cost electricity. It is whether it can absorb, transport, stabilise, value, and settle that electricity when it arrives in concentrated, weather-driven surges. NESO’s own explanation of constraint payments is useful here: when power cannot be transferred across the network as needed, generators may be asked to reduce output and can be compensated for doing so. In other words, surplus is not the same thing as frictionless abundance.

Why the paper now reads less like theory and more like diagnosis

What the paper tried to do was to pull the debate away from slogans and back toward system mechanics. Its framework rested on five principles: demand-supply balancing, efficient resource allocation, reliability and security, transparency and fair competition, and affordability. Those principles were presented as broad evaluative criteria for market design, but today they read more like a map of the current moment. The live UK discussion around curtailment, surplus, flexibility, and paid demand can be understood almost entirely through that lens.

1. Balancing: surplus power is still a balancing problem

The first principle is balancing. In the paper, I treated balancing markets not as a peripheral layer but as a core institution that makes electricity markets workable. That distinction matters more today than it did when the piece was published. A system can be energy-rich in aggregate and still operationally awkward in specific hours. NESO’s summer outlook makes exactly that point: low summer demand combined with strong wind and solar output can create periods of surplus that require active management. Demand turn-up, exports, storage, and output reduction are therefore not side issues. They are the grid revealing its operating logic in public.

Seen this way, the phrase “paid to use electricity” can mislead in two directions at once. It can sound like either a consumer windfall or a system failure. In reality, it is neither. It is a balancing tool. The important lesson is that once variable generation becomes large enough, balancing stops being a residual function and becomes one of the system's central economic questions.

2. Resource allocation: cheap output is not the same as high system value

The second principle is efficient resource allocation. This is where the older LCOE-heavy debate starts to look inadequate. A low-cost megawatt-hour is not automatically a high-value megawatt-hour. Its value depends on when it arrives, where it arrives, what the network can do with it, and what other assets must be ready when it does not. That was one of the paper’s central criticisms of split-market and fixed-price approaches: by focusing too heavily on average generation cost, they risk obscuring the real cost of system integration.

That critique now looks uncomfortably timely. If the system has to pay consumers to “show up” and absorb surplus, the key issue is no longer whether generation was cheap to produce. The issue is whether the electricity arrived in a form that the system could use naturally. When it does not, the system must either pay demand to absorb it, curtail generation, lean on storage, export it, or procure other balancing actions. None of that means the original generation was pointless. It means the value of generation and the value of the system are not the same thing.

3. Reliability and security: scarcity is moving, not disappearing

The third principle is reliability and security. One of the mistakes embedded in the old politics of the transition was the assumption that if energy became cleaner and more abundant, security would improve automatically. But security is not just a quantity problem. It is also a controllability problem. As renewable penetration rises, scarcity does not disappear. It changes location. It moves away from reliance on fuel alone and toward flexibility, timing, network transfer capability, reserve arrangements, and controllable demand.

This is not an abstract point. NESO’s 2025 Annual Balancing Costs Report explicitly states that decarbonizing the system and adding more intermittent, low-carbon generation change the demands on balancing tools. Likewise, Future Energy Scenarios 2025 stresses that much higher shares of wind and solar require substantial new sources of flexibility and storage to meet demand year-round. The system is therefore not leaving system-management economics behind. It is entering deeper into them.

4. Transparency: the public conversation still hides the system story

The fourth principle is transparency and fair competition. This part of the debate receives too little attention. To an ordinary consumer, “paid to use electricity” sounds simple: a nice deal in one half-hour window. It may indeed be a good deal in that moment. It may even be cheaper for the system than curtailing generation or calling on more expensive balancing actions. But unless the conversation also explains where the avoided costs sit, how the incentives are funded, and what exactly is being optimized, the public is still being shown only the surface of the system.

That matters politically. As the grid becomes more dependent on DFS-like services, locational balancing, curtailment management, storage dispatch, and active constraint resolution, opacity itself becomes a market-design problem. A system can be technically sophisticated and still fail the public if its cost logic is invisible. One of the underappreciated strengths of the five-principle framework is that it forces affordability and transparency to be discussed together rather than separately.

5. Affordability: the system may still save money, but the surplus is not costless

The fifth principle is affordability. This is the point where the argument often slides into overclaiming. It would be wrong to say that paying consumers to use electricity automatically proves their bills must rise overall. Ofgem’s DFS decision suggests the opposite can be true in specific cases: demand-side turn-up may save money relative to alternative balancing options. But it would be just as wrong to treat these payments as proof that the system has become costless. The payment exists because the surplus still has to be managed. The fact that it may be the cheapest available option does not make it free. It makes it an efficient way to handle real system costs.

That distinction matters. The correct system’s reading is not “this proves the transition is failing” and not “this proves the transition is solved”. It is possible that a more renewable-heavy system can lower variable generation costs while simultaneously increasing the importance of balancing, absorption, networks, and settlement design. Affordability, therefore, has to be judged at the whole-system level, not at the level of one cheap or negative-price interval.

Where this seems to be heading

The bigger implication is that the center of the electricity debate is shifting. For years, the fight was framed around the relationship between renewables, gas, and wholesale price formation. That question has not disappeared, but it is no longer sufficient. The next stage of the debate is likely to revolve around system absorption: who provides flexibility, how local congestion is handled, how storage is valued, how much demand can be moved in time, how networks are reinforced, and how all of this is coordinated without collapsing into either inefficiency or opacity.

That is also why my earlier paper now reads less like a defense of one pricing model and more like a broader warning against single-metric thinking. The real issue was never simply whether gas should set the price or whether renewables were cheap on average. The issue was whether the system's architecture could reveal and manage the full cost of matching variable supply to real demand while preserving reliability, transparency, and affordability. The current UK moment does not settle that question. But it does reveal, very clearly, where the pressure points are emerging.

The transition, in other words, is entering a more system-intensive phase. Cheap generation remains essential. Deployment remains essential. Investment remains essential. But the political and economic story cannot stop there. Once renewable output becomes abundant for more hours, the hard question is not just how cheaply power can be produced. It is how coherently it can be absorbed and governed. That is what curtailment, DFS payments, network constraints, and flexibility markets are really about.

Conclusion

The current “paid to use electricity” moment is not a gimmick. It is one of the clearest signs yet that the energy transition is entering a new phase. The question is no longer just whether we can build enough low-carbon generation, or whether that generation is cheap at the plant level. The question is whether the wider system can absorb it cleanly, balance it efficiently, and deliver it as reliable value rather than periodic surplus.

That is why this matters. When the system has to pay demand to appear, curtail output, or lean more heavily on balancing tools to keep conditions stable, the lesson is not that abundance is meaningless. It is that abundance alone is not enough. Power still has to be moved, timed, stabilized, and matched to demand. If it cannot be, the costs and pressures do not disappear. They re-emerge elsewhere in the system.

So the lesson is bigger than one summer headline. The transition is no longer just a story about replacing expensive fossil generation with cheaper renewables. It is now a story about governing abundance: managing mismatches, valuing flexibility, expanding networks, integrating storage, and designing markets that can turn variable output into a stable, affordable system. That is where the next phase of the energy debate now sits.

References

[1] UK Government, Solar roadmap: United Kingdom powered by solar (2025).

[2] DESNZ, Energy Trends: March 2026.

[3] UK Government, Clean Power 2030 Action Plan: A new era of clean electricity (2025).

[4] International Energy Agency, Projected Costs of Generating Electricity 2020.

[5] NESO, Surplus electricity expected on GB system this summer as NESO rolls out new consumer flexibility tool (2026).

[6] Ofgem, DFS-A18 Decision 2026 and related derogation materials (2026).

[7] NESO, What are constraints payments?

[8] NESO, Summer Outlook publication and supporting update (2026).

[9] Orestis Delardas, Transforming marginal-cost pricing: A blockchain approach to fundamental energy system dynamics, Utilities Policy (2025).

[10] NESO, 2025 Annual Balancing Costs Report.

[11] NESO, Future Energy Scenarios 2025: Pathways to Net Zero.

Most energy commentary stops too early. A project wins support, secures a site, clears planning hurdles, and suddenly the whole thing is discussed as if success has already been achieved. But that is not how project finance works. Winning support is not the same thing as carrying a project through to financial close. Strategic value is not the same thing as full-cost coverage. Technical promise is not the same thing as bankability.

TwinHub is a useful case precisely because it forces those distinctions into the open. It was not an ordinary offshore wind project. It was a 32 MW floating offshore wind demonstrator at Wave Hub, around 16 km off Hayle in Cornwall, using existing cable tails into Hayle substation. The TwinHub project material and the Hexicon Wave Hub acquisition release make clear why that matters. The project inherited real electrical infrastructure, which should narrow part of the cost burden relative to a clean-sheet demonstrator. But it also remained a small, novel offshore project trying to prove a dual-turbine floating concept in a market that had become much less forgiving of thin economics.

The core argument of this article is simple. TwinHub should not be read as proof that floating wind is impossible, but neither should it be treated as an unlucky one-off with no wider lesson. It is better understood as a case in which support, infrastructure advantage and strategic relevance were all real, yet still not enough to create a robust investment case once the full chain from revenue to capital structure was examined.

What TwinHub is, and why it is not a generic offshore project

TwinHub matters because it sits between categories. It is too small to look like a mature fixed-bottom offshore wind farm, yet too capital-intensive to be treated like a lightweight pilot. Public project material describes two floating platforms carrying two turbines each, while Hexicon’s Wave Hub acquisition material points to inherited offshore and onshore electrical infrastructure and a 30 MW grid connection expandable to 40 MW. That infrastructure matters because grid and export scope can be a project in their own right. In plain terms, TwinHub did not start from a blank slate.

But that advantage cuts only one way. Small offshore projects struggle to dilute fabrication, marine spread, insurance, installation and financing overheads. Novel concepts widen contingency and make both procurement and lender comfort harder to secure. So the project began with an advantage and a handicap embedded in the same structure. That is the starting point a reader has to understand before looking at a single model output.

Why winning support was never enough

The public support anchor is clear enough: TwinHub won an AR4 CfD at £87.30/MWh in 2012 prices, which is commonly framed at roughly £104/MWh in nominal benchmark terms for this type of reconstruction. That benchmark framing comes from the Frazer-Nash review for DESNZ, which also makes the much more important point: the awarded price was likely insufficient to fully cover project costs and therefore was not representative of project LCOE.

This is where debate often gets lazy. The public discussion tends to collapse three separate claims into one: the project won support; therefore the economics must have worked; therefore later impairment must have been caused by some new external shock. That chain is too compressed. A project can clear support while still being thin on a full-cost basis. It can look strategically sensible while still being a hard carry for capital. And if that is true, later inflation or supply-chain pressure does not need to create the whole problem from nothing. It only needs to press on a structure that was already narrow.

What the model is doing, and what it is not doing

Helios Nexum LITE is not being presented as a native floating-wind simulator. It is being used as a finance engine. The model runs treat annual net delivered energy as a flat annual-yield proxy and then let the workbook perform the financing translation: capital structure, cost of capital, debt sizing, DSCR, payback, LCOE and value. The model inputs rely on external anchors.

That matters because it tells the reader exactly what sort of claim is being made. This article is not pretending to recover Hexicon’s internal underwriting. It is not claiming to know the exact electrical savings attributable to Wave Hub. It is not claiming a perfect floating-wind technical simulation. What it is doing is more disciplined than that: taking a bounded set of public assumptions, loading them transparently into a real project-finance structure, and then asking whether the resulting economics look coherent with the later impairment.

Basic technical and financial assumptions used in the model runs

The table below outlines the basic case setup and shows why the numbers in the KPI section later are not arbitrary. The 32 MW size and Wave Hub infrastructure come from public project material and Hexicon’s acquisition documentation; the support price starts from the AR4 award and its benchmark treatment in Frazer-Nash; the generic offshore-demo cost and performance anchors come from ORE Catapult; the low-output case is tied back to Inderes; and the later impairment context comes from Hexicon’s Q4 2025 report.

The energy cases are built directly from public net-capacity-factor anchors applied to a 32 MW project. At 50% net capacity factor, annual energy is 32 × 8,760 × 50% = 140.2 GWh. At 40%, it falls to 112.1 GWh. The central case uses 53%, which gives 148.6 GWh and is intended as a modest uplift on the ORE Catapult demo benchmark rather than a claim to an undisclosed TwinHub production case.

Capex is scaled in the same way. ORE Catapult’s £6,350/kW demo benchmark implies about £203.2m for 32 MW, while its £5,075/kW pre-commercial benchmark implies about £162.4m. The demo case therefore uses £203.2m as the public demonstrator anchor. The central case applies a limited credit for inherited Wave Hub electrical infrastructure, taking capex down to £195.3m, a reduction of about £7.9m or roughly 4% versus the generic demo benchmark. The downside case instead stresses capex up to £231.2m, around £28.0m above the demo anchor and about 14% higher, to reflect a later-stage cost reset.

Opex is treated similarly. Starting from ORE Catapult’s £120/kW-year demo operating-cost anchor, a 32 MW project gives £3.84m/year. The central case trims that slightly to £3.70m/year, while the downside case lifts it to £4.42m/year, around 15% above the base. So the KPI outputs later in the piece are not built from freehand assumptions.

The discount rates deserve more care than a casual model note usually gives them. The WACC is not a loose narrative assumption but an endogenous output of the capital-structure and cost-of-capital blocks. The model uses a 3.0% risk-free rate, a 6.0% equity risk premium, no separate country-risk premium for the UK, and an unlevered beta of 0.8. Those inputs sit toward the more benign end of what could be defended for a first-of-a-kind UK floating-wind demonstrator, which is deliberate: the result does not rely on punitive financing assumptions. On that basis, the model produces costs of equity of roughly 9.9% in the demo case, 10.1% in the central case and 8.5% in the downside case, because leverage changes the beta rather than the number being moved by hand.

On the debt side, the model runs use pre-tax senior debt costs of 6.5%, 7.0% and 9.0%, which become after-tax debt costs of about 4.9%, 5.25% and 6.75% once the 25% UK corporation-tax rate is applied. Weighted by the corrected debt and equity shares in each case, that yields WACCs of about 8.0%, 8.2% and 8.2%. The important point is that the discount rate now comes from an explicit, inspectable capital-market logic rather than from a vague hurdle-rate gesture.

The debt sizing also deserves a brief note. The leverage ratios the model produces, around 37 to 39% of total uses in the demo and central cases, are well below the debt shares a mature offshore wind project could often support. That is not a modelling choice. It is the DSCR constraint doing its job. When revenue cover is too thin to service a fuller debt stack without breaching the covenant floor, the model simply cannot size more debt. The low gearing is therefore a symptom of the economics, not a separate assumption.

Inflation and escalation assumptions also need to be visible because they shape both price and cost recovery over the life of the asset. The model runs use a base inflation structure of 2.0% for general CPI, fixed opex, variable fees, land lease, merchant-price escalation, PPA indexation and sustaining capex, with insurance escalated at 3.0%. That is deliberately conservative in form rather than dramatic in tone. It means the model is not juicing the case with heroic price growth. Instead it is letting a fairly standard inflation backbone roll through both revenue and costs. That is why the life-average realised price in the corrected runs rises above the day-one £104/MWh support anchor to about £117.7/MWh, yet still remains below modelled LCOE in every case. In other words, even with ordinary indexation doing its work, the project still struggles to clear its full cost of capital.

This is also why the article keeps separating hard fact, benchmark and inference. The corrected workbooks now supersede the earlier shortcut finance-shell table for value metrics, but the route into the numbers still matters. The reader should be able to see not only the output, but how each block of the output was assembled: what comes from TwinHub, what comes from Hexicon, what comes from Frazer-Nash, what comes from ORE Catapult, and where the model itself is simply translating those inputs into financeability rather than pretending to discover hidden project truth.

Figure 1. Support anchor versus modelled LCOE

This is the visual that should reset the reader’s intuition. In all three cases the modelled LCOE sits above the £104/MWh support anchor. The central case is the important one because it is the case most sympathetic to TwinHub’s infrastructure advantage while still staying within a disciplined reconstruction. And even there the model comes out at roughly £163/MWh. That is not a catastrophic miss in the sense of a project that is obviously absurd; it is worse, because it is a structurally thin miss. It means the project does not need a dramatic external shock to become uncomfortable. It is already entering the later inflation and execution environment with too little room.

Figure 2. Capex corridor and the limited value of inherited infrastructure

The capex bridge is where the optimism has to be cut back to size. Wave Hub is real. Existing infrastructure is real. A live grid pathway is real. But they do not turn a small floating-wind demonstration project into a mature low-cost build. The central case comes down from the generic demo benchmark, but only modestly. That is exactly the nuance that gets lost when people hear “existing infrastructure” and assume it solves financeability. It narrows the gap; it does not erase it.

Figure 3. The capital stack problem

This is the point at which project-finance logic becomes tangible. The reader should not focus only on whether debt participates. Debt does participate. The central case still sizes about £76m of senior debt, and the demo case roughly £75m. But total uses remain much larger than debt capacity, leaving equity to carry about £119m in the central case, about £128m in the demo case and roughly £194m in the downside case. That is the real commercial bottleneck. Many discussions stop at “the project could get debt.” The harder question is whether the residual equity cheque is commercially attractive after that debt has done all it reasonably can.

Figure 4. EBITDA is not the same thing as a comfortable carry

Readers often assume that if EBITDA is positive and non-trivial, the project must be broadly fine. That is too crude. TwinHub’s model runs still show meaningful operating cash generation, with life-average EBITDA margins around 74% in the demo case, 76% in the central case and 62% in the downside case. But EBITDA is only the entrance fee to the financing discussion. It still has to support debt sizing, protect equity value and justify the uses stack. A project can look operationally credible and still be financially weak.

Full KPI suite from the model runs

How the reader should interpret those numbers

The most important nuance is that debt metrics and value metrics are not saying exactly the same thing. In the corrected central run, the minimum DSCR lands at 1.50x and the model labels credit as acceptable, yet project NPV is still roughly negative £74m and equity NPV roughly negative £76m. That is not a contradiction. It is the whole point. A project can be capable of supporting some debt and still be a poor value proposition for equity. It can pass a narrow credit lens while failing the broader economic one.

That distinction matters because it is very easy for public debate to flatten everything into a single yes-or-no test. Did it work or not? Was it financeable or not? The model suggests a more serious answer. TwinHub appears to have had enough substance to avoid looking absurd. It still carries revenue, sizes debt, and produces coherent operating cash flow. But the structure remains too narrow to look comfortable. That is why the project is better described as marginal and conditional than as plainly impossible.

Why the impairment becomes easier to understand

Once the corrected runs are laid beside the project narrative, the later impairment stops looking mysterious. Start with a project that already appears thin on full-cost economics. Add the small-scale offshore overhead burden, concept novelty, higher caution in procurement and financing, and then the worsening macro backdrop that the Hexicon Q4 2025 report implicitly points toward when it books the impairment and reverses the FID-linked contingent consideration. The result no longer needs to be dramatised. A project can remain strategically intelligible and still become too thin to carry.

That is why the impairment should be read carefully. The rigorous claim is not that we can reproduce Hexicon’s exact internal deal model. We cannot. The rigorous claim is that the model runs make the later write-down economically legible. The public story and the model story now rhyme in the right way: a real infrastructure advantage, a still-heavy cost stack, some debt capacity, too much equity burden, and a value envelope that was probably never wide enough for later stress to be absorbed comfortably.

What this case says about energy-project discourse more broadly

TwinHub is useful beyond itself because it exposes a common weakness in how projects are discussed. Public debate loves milestones: award, consent, construction start, FID. But the hard work sits in the translation layer between those milestones. How much annual energy is bankable rather than simply plausible? How much of capex is fixed and unavoidable? How much contingency is required because the concept is still novel? How much debt can be supported on a conservative case? How large is the equity cheque left behind? What exit value remains after all that? These are not side questions. They are the project.

That is also why a piece like this has to walk the reader through the assumptions rather than simply drop a table of outputs. Numbers become credible when the route into them is visible. The point is not to pretend to know more than the public record allows. The point is to know exactly how little needs to be assumed before the economics start to look strained.

The conclusion

The fairest reading of TwinHub is therefore neither cheerleading nor dismissal. The project had genuine strategic logic. It had genuine infrastructure support through Wave Hub. It had enough substance to win support and enough operating credibility for a modelled debt package to participate. But that is not the same thing as being comfortably financeable. Once the model runs are taken seriously, the more sober conclusion is that TwinHub was likely living inside a narrow economic envelope from quite early on.

That is a more interesting conclusion than simply saying floating wind failed or support failed. What the case really shows is that support price, technical plausibility and bankability are three different things. A project can possess the first two and still come unstuck on the third. For capital allocation, that narrower claim is the more serious one.

What is happening in Europe right now is not just another row about subsidies. It is the return of an older problem. Following the war around Iran and the disruption of flows through the Strait of Hormuz, oil and gas prices have surged again, Europe’s fossil-fuel import bill has jumped, and Brussels is warning capitals not to turn an energy shock into a fiscal crisis through broad, open-ended support measures. The immediate line from EU and ECB officials is clear: help, if offered, should be temporary, targeted and limited, as reflected in recent remarks from the ECB.

That fiscal warning is understandable. But it also reveals the deeper issue. Europe once learned that energy dependence is a strategic vulnerability, not just a pricing inconvenience. The lesson from earlier crises was that resilience has to be built before the shock through lower structural exposure, not improvised during the shock through emergency transfers. That is the part Europe keeps half-remembering only after stress returns, as the IEA’s historical review of energy security responses makes clear.

Figure 1: Europe’s current energy anxiety is not an isolated episode. The system has repeatedly been pushed back into crisis mode by external fuel and geopolitical shocks.

The first big parallel is the 1970s oil shocks. Those crises taught advanced economies that dependence on externally supplied oil could quickly become an inflation shock, a growth shock and a political shock at the same time. The institutional response was not only rhetorical. It included the creation of the International Energy Agency, the build-out of strategic oil stocks, demand-restraint planning, efficiency improvements and broader fuel diversification. In other words, the response was to treat dependence itself as the problem. The point was never simply to cushion consumers after the fact. It was to reduce the system’s exposure before the next disruption arrived. The IEA’s comparison between the recent crisis and the 1970s is useful here.

The second parallel is the 2006 and 2009 Russia-Ukraine gas crises. Those episodes taught Europe a more specific lesson: interdependence is not the same thing as security. The EU’s own post-crisis assessment of the 2009 gas crisis stressed the need for reverse-flow capability, stronger interconnections, emergency coordination, LNG optionality and the ability to survive the loss of a major supply route. Later Commission documents on European energy security explicitly note that these changes improved resilience relative to 2009, while the Commission’s 2013 infrastructure review noted that reverse-flow projects helped avoid a gas crisis during the February 2012 cold spell.

Then came the 2022 Russia shock, which forced Europe to relearn the same lesson at much larger scale. The response was REPowerEU: save energy, diversify supplies, and accelerate domestic clean energy deployment. That was a serious response, and it did reduce Russian dependence. The European Council’s summary of REPowerEU and the Commission’s own “REPowerEU, three years on” review still frame it in exactly those terms.

But this is where the present moment becomes more interesting. Not all crisis lessons transfer cleanly across fuels or shock types. Europe’s post-2009 gas reforms were mainly about gas route resilience. REPowerEU was heavily focused on reducing Russian gas dependence and accelerating the clean transition. Both matter. But the current shock is hitting through a different channel: Middle East disruption, oil and refined products, shipping chokepoints, and fears around diesel and jet fuel as much as pipeline gas. Recent reporting on EU preparations for prolonged disruption and record U.S. fuel exports to Europe and Asia to replace Middle East supplies captures that shift. Europe can buy more LNG and improve reverse flows; that does not remove exposure to a global oil shock priced through seaborne markets, which is precisely the broader vulnerability discussed in the IEA’s work on sheltering from oil shocks.

Figure 2: Post-2022 Europe reduced Russian gas exposure, but it did not eliminate external dependence. The vulnerability changed form rather than disappearing.

That is why I am not fully convinced the present shock can be managed on the same terms as the previous one. Some of the old tools still matter. Strategic stocks matter. Demand restraint still matters. Efficiency still matters. Diversification still matters. But the current episode is a reminder that Europe solved part of one dependence problem while remaining exposed to another. The lesson from the 1970s, from 2009, and again from 2022 was not simply “respond quickly.” It was that dependence is itself a strategic risk. If that risk is not reduced in advance, the cost does not disappear. It comes back later as inflation, industrial stress, and eventually fiscal strain.

That is the real significance of today’s debate. Europe is once again arguing about how much support states can afford once the energy shock has already landed. But by that stage the underlying failure has already happened. A resilient system is not one that gets better at emergency subsidy design every few years. It is one that reduces the chance that external fuel stress can repeatedly force the same choice between household pain and sovereign balance sheets. The ECB’s recent framing of targeted support and second-round inflation risks sits squarely inside that broader problem.

Energy policy is often narrated through the language of succession. One era is said to have been about security of supply, another about market liberalisation, another about decarbonisation, and the present one about resilience or industrial policy. That sequencing can be useful, but it also risks overstating discontinuity. In practice, energy strategy has rarely moved through the clean replacement of one governing logic with another. More often, new priorities are layered onto older strategic concerns, temporarily reordering them without fully displacing them.

That is why the present moment matters. What is changing now is not that climate has ceased to matter, nor that decarbonisation has somehow been disproved. It is that the climate-centred phase of energy policy is beginning to give way to something older and more constant: a harder strategic grammar organised around dependence, timing, infrastructure, system control, geopolitical exposure, and the political capacity to carry these burdens. The technologies have changed. The strategic problem has not.

The focus here is not “the global energy system” treated as one uniform object. It is the energy-policy trajectory of advanced industrial systems, especially Europe, viewed in interaction with global fuel, equipment, capital and mineral chains. That distinction matters. A war-driven gas shock in Europe, a blackout in Iberia, and disruption around the Strait of Hormuz are not the same event and do not reveal the same weakness. What they do reveal is a recurring strategic truth: when the system is stressed, policymakers are driven back toward questions of dependence, operability, control and exposure.

I. The fuel-security era

The first anchor is the fuel-security era. The International Energy Agency was created in 1974 in direct response to the 1973-1974 oil crisis, when the embargo exposed the vulnerability of advanced economies to imported hydrocarbons. In that regime, the central strategic variables were plain: fuel availability, producer power, import dependence, price shocks, dispatchable capacity, strategic reserves and macroeconomic insulation. Security of supply was not one policy objective among many. It was the architecture within which the rest of the policy had to fit.

This earlier energy regime was built around molecules and the infrastructures that moved them. Strategy meant reducing exposure to fuels one did not control, or at least making that exposure more manageable. Diversification, storage, domestic production, fuel switching, nuclear expansion and efficiency belonged to the same broad logic: not the abolition of dependence, but the reduction of its most dangerous forms. The IEA’s own contemporary framing of energy security still preserves that continuity, even though it now extends beyond oil to gas, electricity and clean-energy supply chains.

II. The market-integration era

The second anchor is the market-integration era that took clearer institutional shape in Europe across the 1990s and early 2000s. The 1996 Electricity Directive did not remove the strategic problem of energy dependence. It reframed it through competition, interconnection and regulatory integration. Security was increasingly to be pursued not only through domestic buffers and fuel strategies, but through market opening, cross-border exchange and more efficient network use.

The European Commission’s 2000 Green Paper on security of supply is especially revealing in this regard. It warned that Europe’s external dependence was rising, framed security of supply as the uninterrupted physical availability of energy products at affordable prices, and situated that concern inside a broader debate that already included sustainability. This was not yet the climate-centred period, but the bridge toward it is already visible. Europe was trying to manage dependence through markets while also confronting the strategic and environmental limits of that model.

That is an important historical anchor because the climate-centred interval did not emerge after a simple hydrocarbon-security world. It arrived after a market-mediated security world, one that had already attempted to reconcile interdependence, affordability, competition and environmental pressure. Seen in that longer sequence, the climate-centred period becomes easier to locate historically: not as the replacement of strategy by climate, but as a third strategic regime layered onto the earlier two.

III. The climate-centred interval

By the climate-centred interval, I do not mean only the years immediately after the Paris Agreement. I mean a broader period, especially across the mid-2010s and early 2020s, in which decarbonisation increasingly appeared capable of functioning as the master organising principle of energy policy. Paris helped codify that direction through its goals of limiting temperature rise, strengthening adaptation, and aligning finance flows with low-emissions, climate-resilient development. But the interval rested on more than diplomacy. It drew its force from an unusual alignment between policy ambition, investor narratives, industrial deployment and improving technology economics.

That interval was real. According to IRENA’s Renewable Power Generation Costs in 2024, utility-scale solar PV costs in 2024 were roughly 90% below 2010 levels, while onshore wind costs were around 70% lower. Ember’s Global Electricity Review 2025 reports that low-carbon sources supplied 40.9% of global electricity generation in 2024, the highest share since the 1940s. These were not cosmetic shifts. They changed the plausibility of the transition narrative.

But the interval should be understood carefully. Its coherence did not emerge by magic, as though technology simply became cheaper and politics followed. It was produced through policy support, industrial scaling, financing adaptation, regulatory commitment and political legitimacy. Market formation drove deployment; deployment drove learning; learning improved economics; improved economics widened coalitions. The interval was historically unusual not because it suspended the old logic of energy strategy, but because it temporarily made climate appear capable of subsuming it. And it could do so partly because the fuel-security and market-integration logics were, for a time, quiet enough not to demand priority, not because they had been resolved.

IV. The structural conditions that never left

To say that the climate-centred interval is giving way is not to say that its achievements were illusory. It is to say that deeper structural conditions, briefly pushed into the background, are now reasserting themselves more visibly. They receded during the interval, not because they were solved, but because the interval's favourable alignment of falling costs, policy momentum, geopolitical stability, and manageable demand growth temporarily reduced the pressure they exerted.

The first of these conditions is timing. Energy systems are not governed only by annual balances and headline capacity additions. They are governed by when supply appears, how quickly assets can be built, how long flexibility can hold, and whether institutions can act before constraints bind. This was true in the fuel-security era, when reserve cover and dispatchable availability mattered. It remains true now, only in a more electrified and operationally sensitive form.

The second is materiality. Energy systems depend on physical networks, transformers, cables, pipelines, mines, ports, inverters, turbines, maintenance cycles and reserve margins. They are never pure stories of prices or targets. They are built environments with lead times, bottlenecks and failure modes.

The third is system control. This includes observability, reactive power management, frequency stability, inertia or its substitutes, protection settings, communications and real-time coordination. It is not enough for a system to be low-carbon on paper. It must also remain controllable under stress.

The fourth is political legitimacy. Energy transitions are not only engineering exercises. They must be financed, governed and accepted through institutions that can distribute cost, risk and benefit in ways that remain politically durable. That broader point also sits at the centre of the framework I have previously discussed in conference work: transitions unfold through competing narratives, coalitions, and implementation dynamics rather than through techno-economics alone.

The fifth is strategic exposure. Dependence does not disappear because a system becomes cleaner. It changes form: from fuel imports and shipping lanes to grids, minerals, manufacturing concentration, interconnection, digital infrastructure and industrial supply chains. These are the harder constants of energy strategy. They are not new, but they become salient in different ways in different regimes.

V. Why those conditions are reasserting themselves now

The current phase feels less forgiving because the transition has advanced far enough for its more difficult layers to become unavoidable. The easier political and industrial task was to show that clean technologies could scale and become economically serious. The harder task is to build the architecture that allows them to operate as a durable strategic system.

That is why grids have moved to the centre of the debate. The European Commission’s grids agenda says electricity consumption in the EU could rise around 60% by 2030, that 40% of distribution grids are already more than 40 years old, and that €584 billion of grid investment is needed by 2030. The significance of that figure is not merely financial. It marks the point at which the transition stops looking like a generation story and starts looking like a system-delivery story.

The same is true of minerals and industrial concentration. The IEA’s Global Critical Minerals Outlook 2025 states that the average market share of the top three refining nations for key energy minerals rose from around 82% in 2020 to 86% in 2024, with about 90% of supply growth coming from the leading supplier. This does not invalidate decarbonisation. It reveals that decarbonisation is not a post-dependence condition. It is a different dependence structure.

Demand growth is adding further pressure. The IEA’s Energy and AI analysis estimates data centres used around 415 TWh in 2024 and projects their electricity consumption could rise to around 945 TWh by 2030. That is not the whole story of future demand growth, but it is a useful illustration of the broader point: the system is not simply trying to decarbonise a static load. It is trying to decarbonise while demand itself becomes more electricity-intensive, more digital and more timing-sensitive.

VI. The pressure points that clarified the system

These structural conditions did not become visible only through slow-moving trends. They were clarified by shocks, not because the shocks created the underlying reality, but because they exposed it to policymakers more brutally than ordinary debate could.

The first major pressure point was the Russia-Ukraine war. The REPowerEU response made explicit what had often been implicit: in Europe, the transition had become inseparable from the geopolitics of dependence. The European Commission says Russian gas imports to the EU fell from 150 bcm in 2021 to 52 bcm in 2024, and the Russian share of EU gas imports dropped from 45% to 19%. The war did not merely accelerate climate policy. It forced a strategic reordering in which decarbonisation, diversification and security of supply became intertwined under the pressure of hard geopolitical exposure.

The second pressure point was the Iberian blackout of 28 April 2025. Here it is important not to collapse the lesson into a crude morality tale. The ENTSO-E blackout page links both the factual and final investigations, while Red Eléctrica’s June 2025 report provides the Spanish operator’s own account. The final expert report describes a sequence involving uncontrolled voltage rise, loss of voltage control, oscillatory phenomena, rapid generation output reductions and disconnections, and cascading events that led to the blackout. In the half-hour before the incident, the system experienced a 0.63 Hz local oscillation and then a 0.2 Hz inter-area oscillation. The report’s management summary makes clear that this was a system-operability event, not a simplistic referendum on one generation technology.

The significance of the Iberian event lies precisely there. It forced system reality into full view: voltage control, reactive power behaviour, protection settings, observability of embedded generation, oscillatory behaviour, and the limits of policy frameworks that are more comfortable counting megawatts installed than analysing operability under stress. A structural break of that kind does not tell us that one technology is good or bad in the abstract. It tells us where the actual frontier of energy policy has moved. It has moved into control, integration and system architecture.

The third pressure point is the current Iran-linked disruption around the Strait of Hormuz. This is a live crisis, so it should be handled cautiously. But the strategic lesson is already plain. The IEA’s Strait of Hormuz note states that in 2025 nearly 15 mb/d of crude oil, almost 34% of global crude oil trade, passed through the Strait. Reuters reported today that the IEA expects Middle East supply disruptions to intensify in April and hit Europe more directly. One does not need to inflate that into a grand claim about the end of the transition to grasp its importance. It is another reminder that the older hydrocarbon security architecture still has the power to force itself back onto the agenda very quickly.

These pressure points are not identical and should not be made to carry one uniform lesson. The Russia-Ukraine war exposed the geopolitical fragility of Europe’s gas dependence. The Iberian blackout exposed the operational fragility of a power system under complex control stress. The Hormuz disruption exposes the enduring leverage of oil chokepoints over the wider global economy. Their common significance lies elsewhere: each forced policymakers to see again that energy policy remains governed by dependence, timing, control and exposure, even when public rhetoric is dominated by transition.

VII. What is actually returning

What is returning, then, is not the old world in any simple sense. Europe is not going back to the 1980s. The technology base has changed too much for that. What is returning is something more fundamental: the primacy of energy strategy over energy narrative.

During the climate-centred interval, it was easier to believe that once costs fell and deployment accelerated, the rest of the strategic problem would gradually be absorbed by the logic of decarbonisation. The present phase is showing the limits of that assumption. Cheap generation is not the same thing as system adequacy. Installed capacity is not the same thing as controllable power. A cleaner mix is not the same thing as a more governable dependence structure.

That is why the current period feels, in one sense, older than the decade that preceded it. Not because nothing changed, but because energy policy is once again being judged less by declared direction and more by whether the system can actually perform under strain. The climate interval now looks less like the final transcendence of energy strategy and more like a period in which climate temporarily became the dominant language through which energy strategy was interpreted.

VIII. Where this leaves the next phase

The next phase of energy policy will not be decided by who can recite the cheapest asset-level technologies, nor by who can most theatrically denounce the transition. It will be decided by who can build a more governable energy architecture.

That means taking timing seriously again: not only build rates, but sequencing, flexibility duration, maintenance cycles, reserve margins, lead times and the pace at which networks and institutions can adapt. It means taking system control seriously: reactive power, protection settings, observability of embedded resources, communication standards, and the discipline of operation under abnormal conditions. It means taking strategic exposure seriously: not just Russian gas or Middle Eastern oil, but also transformers, cables, copper, rare earths, inverters, software and the concentration of industrial capabilities in a small number of jurisdictions. And it means taking political durability seriously, because no energy architecture, however elegant on paper, survives if it cannot distribute costs, risks and gains in a way that remains legitimate.

The deeper mistake in much current debate is to imagine that energy policy is essentially a dispute over preferred technologies. It is not. It is a dispute over which dependence structure can be made more resilient, more governable and more politically durable.

That, to me, is the more revealing way to read the present moment. The climate-centred interval was real. It achieved things that matter and will continue to matter. But it was not the final transcendence of energy strategy. It was a period in which climate, for a time, looked capable of becoming the sovereign grammar of energy policy. What comes after is not a negation of that period. It is the reappearance of the harder subject that never left.

Energy policy began with fuel security, moved into market-mediated interdependence, then entered a climate-centred interval of unusual coherence. It is now moving into a phase where the central question is no longer whether the system can become cleaner in principle, but whether it can remain operable, affordable, controllable and strategically resilient as that transformation deepens.

The significance of the present moment is not that it overturns the transition narrative, nor that it straightforwardly confirms it. What it reveals instead is not the disappearance of earlier energy logics, but their accumulation. The central issue is no longer only the direction of change, but whether the emerging system can operate credibly under constraint. A period of rapid technological and policy advance oriented around decarbonisation now has to mature into one in which fuel security, market integration, infrastructure, system control and strategic exposure are reconciled rather than treated as sequential stages or allowed to crowd one another out.

That is a narrower claim than either triumphalism or pessimism, but probably a more useful one.

The easiest mistake in reading the current U.S. energy situation is to assume there must be one master story. Oil prices. Trump. Offshore wind. Inflation. ESG backlash. Geopolitics. But the structure underneath looks messier and more revealing than that. The United States is not facing one uncertainty regime. It is operating inside a stack of them, only partly connected, each stressing a different part of the system. That means the country is not simply living through a louder version of the same old cycle. It is trying to make long-duration decisions while several different risk systems are active at once.

That distinction matters for energy. A fuel shock, a policy-credibility shock, a geopolitical shock, and a transition-finance shock do not have the same cure. They hit different veto points in the system. So the practical question is not whether uncertainty is high. It is which uncertainty is binding, through what mechanism, and with what strategic consequence.

Read that way, the charts do more than show noise. They show a country whose strategic debate has become structurally unstable. The United States is no longer arguing mainly over one scarcity problem. It is arguing over which risk should dominate strategy: fuel exposure, geopolitical risk, policy durability, transition credibility, or public climate salience.

1. What the data says, not just what it looks like

Three statistical results matter most. First, the strongest pairwise correlation in the merged sample is between economic policy uncertainty and U.S. ESG uncertainty, at about 0.55. That is meaningful, but it is still far from a one-factor world. Second, oil-price uncertainty is only weakly related to headline economic policy uncertainty, at roughly -0.06, and only modestly related to global energy uncertainty, at about 0.33. Third, U.S. climate concern is elevated but only loosely connected to the harder uncertainty metrics: its correlation with economic policy uncertainty is about 0.10, with geopolitical risk about 0.16, and with global energy uncertainty essentially zero.

The trailing data tells a similar story. Over the latest trailing 12-month window, U.S. ESG uncertainty has the highest average standardized reading in the set, at 3.92σ. Economic policy uncertainty follows at 3.31σ. U.S. climate concern is elevated but much lower, at 0.98σ. Geopolitical risk sits around 0.76σ, and oil-price uncertainty, despite its recent violence, averages only 0.64σ over the last year because it is spiky rather than continuously dominant.

The period averages sharpen the point. In the 2024–2026 window, the highest composite channel readings are trade/regulatory uncertainty (2.47), headline policy uncertainty (2.05), and ESG/transition uncertainty (1.99). Geopolitical/security uncertainty is also elevated (1.10), but less central than the policy-and-transition complex. That is why this moment looks less like a pure commodity shock and more like a struggle over rules, credibility, and strategic direction.

The first chart makes that visible. Oil-price uncertainty is the most violent line in the sample, especially at the recent end of the series. But the broader pattern is not simply one of oil shock. The stronger persistent story is that policy uncertainty and transition-facing uncertainty have risen and stayed elevated together. Climate concern is present, but it behaves more like a persistent salience series than a hard volatility channel. Global energy uncertainty helps explain the broader historical backdrop, but because the merged panel ends that series earlier than the others, it is strongest as a regime lens rather than as a very-latest-month indicator.

The stress map adds something that the line chart cannot. It shows that U.S. uncertainty has not advanced in one straight line. Different periods are dominated by different mixes. In some phases oil matters most. In others the policy-and-transition complex becomes more central. That is important because it means the strategic argument keeps changing even when the word uncertainty stays the same. The country is not merely seeing “more” uncertainty. It is rotating through different hierarchies of risk.

The correlation structure is the key analytical result. The strongest relationship in the set is between economic policy uncertainty and U.S. ESG uncertainty. That suggests the transition is increasingly being contested through rules, regulation, finance, and political credibility rather than through technology costs alone. Elsewhere the relationships are weaker, and in a few cases close to zero. Oil uncertainty is not the organizing anchor of the whole system. Climate concern does not unify the field either. What we are seeing instead is a stack of different uncertainty channels pressing on different parts of the economy at the same time.

2. The quick reading

3. Driver map: what each metric is actually picking up

This is where the charts stop being descriptive and start becoming useful. Each series is a different transmission channel of strategic stress. They are not six names for the same thing.

4. What this says about the U.S. situation on the ground

The United States now looks less like an economy organized around one dominant scarcity problem and more like one organized around a contested hierarchy of risks. The fuel channel is still real. Oil spikes still matter. But the strongest persistent signals in the latest data are policy-facing and transition-facing. That points to a system increasingly constrained by credibility, coordination, and institutional durability rather than by fuel alone.

Put differently, the country is no longer only asking whether it can get energy. It is also asking whether it can maintain a stable enough policy, trade, regulatory, and financing framework to build the next system. That is why the correlation between headline economic policy uncertainty and ESG uncertainty matters so much. It suggests the transition is being fought over the rules of capital allocation and state commitments, not just over cost curves or emissions targets.

The weaker correlations elsewhere are just as revealing. Climate concern is present, but it does not unify the strategic field. Geopolitical risk matters, but it is not reducible to oil. The broader picture is a U.S. system in which several veto points are active at once, with no settled agreement on which one should dominate strategy. That is why energy politics feels unstable on the ground. The country is not merely nervous. It is divided over which fragility it fears most.

5. How the external literature and the news fit the pattern

Federal Reserve research shows that uncertainty shocks depress investment and notes that in mid-April 2025, U.S. economic policy uncertainty reached a new peak while trade-policy uncertainty surged even more sharply.

Richmond Fed research shows that economic policy uncertainty rises around elections and rises more when elections are close and polarized, which helps explain why policy durability has become such a central transmission channel.

Brookings argues that the energy transition is unfolding under deep uncertainty about supply, demand, markets, and geopolitics, which is consistent with the rotating stress structure visible in the heatmap.

Columbia’s Climate of Contempt argues that bitter partisanship blocks durable climate and energy policy even when public concern exists, which fits the weak link between climate salience and the harder policy and finance channels.

The SEC’s March 2025 decision to end its defense of climate-disclosure rules is one concrete example of why ESG uncertainty can be read as institutional instability rather than as a soft sentiment variable.

The Conference Board’s 2025 survey reported that 80% of surveyed sustainability executives were reworking ESG strategies in response to policy shifts, reinforcing the idea that transition finance is being repriced through political uncertainty.

Recent Reuters reporting shows that the March 2026 Iran conflict has again pushed energy-price and inflation risk back into the macro picture, illustrating why oil uncertainty remains explosive even if it is not the whole story.

6. The synthesis

What these series reveal is not generic uncertainty but a system with multiple active veto points. Fuel risk, geopolitical risk, policy risk, transition-finance risk, and climate salience are all present, but they do not move as one. That is why U.S. energy politics now feels so unstable: the country has not settled on which risk should dominate strategy.

The U.S. energy system is no longer organized around one dominant scarcity problem. It is organized around a contested hierarchy of risks.

That is also why the piece should be read as more than a story about one technology or one policy fight. The deeper issue is that America’s next energy system is being built in an environment where the risks are plural, the veto points are multiple, and the state has still not decided which fragility it wants to reduce first.

Selected references

Federal Reserve, Costs of Rising Uncertainty (2025)

Richmond Fed, Economic Policy Uncertainty in Election Years (2024)

Brookings, Navigating market and political uncertainties in the age of energy transition (2025)

Columbia Center on Global Energy Policy, Climate of Contempt

SEC, Votes to End Defense of Climate Disclosure Rules (2025)

The Conference Board, Sustainability Under Scrutiny (2025)

PolicyUncertainty.com, methodology and index pages

Reuters, Iran-war energy inflation reporting (2026)

The event is bigger than the technology

The latest US offshore wind reversal matters for a reason that goes well beyond offshore wind itself.

What is revealing here is not simply that one administration dislikes a technology another administration supported. That is common enough in politics. What is more revealing is the mechanism. According to Reuters and the Associated Press, the Trump administration agreed to reimburse TotalEnergies for relinquishing offshore wind leases off New York and North Carolina, with the company then redirecting an equivalent amount of capital into US oil, gas and LNG projects. This is not just permitting friction or rhetorical hostility. It is a more direct act of capital re-routing, which is why the episode should be read as something larger than a dispute over turbines at sea.

At the centre of the story is a bigger argument about how the United States now understands power, growth, resilience, and strategic dependence.

Energy policy has always been statecraft

The first mistake in energy debate is to treat episodes like this as if they break from some normal baseline in which governments leave the sector alone. They do not. The US state has always shaped energy systems, especially in strategic moments. The 1973–74 oil embargo accelerated a broader reorganisation of federal energy policy, and the Department of Energy was created in that period. More recently, the Inflation Reduction Act was described by the US government as the single largest investment in climate and energy in American history.

The serious question, then, is not whether Washington intervenes. It always does. The serious question is how it intervenes, toward what kind of system, and with what degree of durability.

Every major US energy turn presents itself as realism. The 1970s framed realism through fuel vulnerability and geopolitical shock. The IRA framed realism through industrial policy, domestic manufacturing and accelerated clean-energy buildout. The current turn appears to frame realism differently again: through a harder preference for hydrocarbons, conventional supply and exportable fuel. That is the historical continuity worth keeping in view. The state is still choosing an energy pathway. What is changing is the kind of pathway it now wants to privilege.

Why the two camps will read this differently

This is why the two camps will fight so hard over the present case. The pro-wind camp will see political sabotage of a long-duration clean-energy pathway. From that perspective, the damage is not confined to the leases themselves. It is damage to policy credibility. If investors can be brought into strategic infrastructure and then pushed back out once political control changes, the message is that long-horizon energy investment in the United States remains politically fragile.

The anti-wind or pro-fossil camp will see something different. It will read the move as a correction of a distorted policy mix. In that account, Washington had leaned too far toward a politically favoured technology whose economics, delivery path, or institutional demands were weaker than advertised. Redirecting capital toward oil, gas and LNG is then framed as reprioritisation toward assets seen as more grounded in reliability, strategic utility or economic realism.

Both readings are understandable. Neither gets to the heart of the matter.

The real divide is over rival theories of power

The deeper issue is that each side is really defending a different theory of what secures national strength. One side increasingly sees geopolitics through fuel exposure. Its instinct is that structural dependence on globally traded hydrocarbons leaves an economy vulnerable to external shocks, price transmission, and geopolitical instability. In that worldview, resilience comes from electrification, infrastructure coordination, industrial repositioning, domestic clean manufacturing, and a system that becomes less sensitive to fuel volatility over time.

The other side sees geopolitics through physical control and dispatchable supply. Its instinct is that strategic strength still comes from abundant extractable energy, domestic production, export capacity, and the diplomatic or industrial leverage that accompanies those things. In that worldview, conventional energy remains the firmer base for economic growth and geopolitical power.

That is why this is not really about offshore wind. Offshore wind is simply where those two theories now collide. One theory says future strength will come from building an economy around electrified infrastructure and lower fuel exposure, even if that demands more planning, more coordination and more institutional competence. The other says future strength still depends on preserving extractive abundance, dispatchable supply, and a model of energy security that fits more comfortably with older forms of industrial and geopolitical hard power.

Offshore wind was always a wider institutional test

This is also why it is too shallow either to romanticise offshore wind or to treat recent setbacks as proof that the technology was a fantasy all along. Technologies do not carry their economics internally. Their economics depend on the wider system architecture around them: the grid, the balancing regime, the ports, the transmission buildout, the permitting process, the supply chain, the financing logic, the industrial base, and the state capacity holding it together. Offshore wind in the United States has therefore been more than a test of turbines at sea. It has been a test of whether the state can actually deliver the wider architecture that offshore wind requires.

That test has produced mixed signals. Costs have risen, timelines have slipped, permitting has proved contentious, and supply chains and supporting infrastructure have not developed with the smooth inevitability often implied by transition rhetoric. Many people read this as a failed test. That goes too far. A better reading is that offshore wind has exposed the deeper institutional question: not whether the technology can work, but whether the surrounding political and industrial system can support it consistently enough to make it strategic.

A great deal of energy commentary still treats technologies as if they explain themselves. They do not. Their economics and strategic value are functions of the systems in which they sit.

The US is confusing technology preference with dependency preference

This is why the more useful distinction here is not between one technology and another, but between different dependency preferences. Governments do not merely choose technologies. They choose which dependencies they find more tolerable, and which they believe they can govern more competently.

Backing offshore wind is not simply backing turbines. It is backing a theory that the United States can coordinate transmission, ports, marine logistics, long-horizon infrastructure, industrial supply chains, and capital formation well enough to make that pathway durable. Backing oil and gas is not simply backing molecules. It is backing a theory that control over extractive capacity, exportable fuel, dispatchable supply, and conventional energy strength remains the safer basis for geopolitical and economic power.

Why this feels more volatile now

The United States has always had energy-policy swings, but this does feel more volatile than many earlier periods. The main reason is that energy and climate are now more tightly bound up with partisan identity. Pew Research Center found a sharp partisan divide over whether climate policies help or hurt the economy. Columbia’s Center on Global Energy Policy has highlighted how partisanship increasingly constrains climate action. Brookings argued in 2025 that the United States had become a major source of uncertainty in the pace of the global energy transition.

This suggests a two-way relationship: polarisation feeds policy volatility, and policy volatility reinforces polarisation by making every election feel like a contest over the entire development model.

That matters especially in energy because energy is one of the sectors least able to thrive on strategic whiplash. These are long-duration, capital-intensive, infrastructure-heavy systems. They do not just need support. They need credible support. They need manufacturers, developers, and investors to believe that a strategic direction will survive long enough to justify sunk costs and wider industrial mobilisation. That is why uncertainty matters so much. The Richmond Fed notes that economic policy uncertainty rises significantly around elections and is amplified by close, polarised contests. A Federal Reserve note reported that economic policy uncertainty reached a new peak in April 2025, with trade-policy uncertainty rising even more sharply.

Once a state becomes willing not only to change incentives but to unwind already-committed pathways and redirect capital elsewhere, the signal does not stop at one project category. It raises the risk premium on the credibility of the strategy itself.

The uncertainty backdrop is broader than a single policy shock

The current moment also looks more crowded empirically than a simple story of election-driven policy volatility would suggest. The uncertainty series do not point to one dominant stress alone. They show a broader regime in which policy uncertainty, oil-price uncertainty, geopolitical risk, ESG-related uncertainty and climate concern move in overlapping but distinct ways.

The recent indexed comparison is especially revealing here. Oil uncertainty is the most explosive and erratic component. Economic policy uncertainty remains elevated and recurrent. ESG-related uncertainty rises sharply in the recent period. Climate concern, by contrast, is steadier and less violent. The point is not that all these series move together mechanically. They do not. It is that US energy choices are now being made under several layers of uncertainty at once.

The regime heatmap strengthens that point by showing that uncertainty clusters rather than simply rising in a straight line. Different periods are characterised by different combinations of stress. Some are more fuel- and geopolitics-heavy. Others are more policy-heavy. Others show a broader spread across several channels at once. That matters because it suggests the current fight over energy direction is not simply a reaction to one shock. It is taking place within a wider environment of repeated and shifting strategic pressure.

The correlation structure points in the same direction. These uncertainty measures are related, but they are not interchangeable. That is important analytically. It means there is no single variable that can explain the present conflict on its own. Different political camps can point to different risks and still be responding to something real. What divides them is not whether risk exists, but which risks they treat as more fundamental and which dependency structure they consider more governable.

That is why the deeper problem is not just volatility in the narrow sense. It is the absence of a settled hierarchy of priorities. The United States is trying to make long-horizon energy choices in an environment where fuel risk, geopolitical risk, policy risk, sustainability risk and public climate concern all remain active, but where there is no stable consensus on which of these should dominate strategy.

What this says about growth, geopolitics, and the state

This is where the story becomes more than a sector fight. Each energy choice implies a different view of how growth is generated, but those growth models are now being judged under very different kinds of uncertainty.

A system built around electrification, infrastructure coordination, and lower fuel exposure assumes that future growth will come from industrial repositioning, network buildout, manufacturing depth, and a less fuel-sensitive cost base over time. But it also assumes the state can manage complex infrastructure risk, permitting risk and policy durability risk well enough to make that pathway credible.

A system built around hydrocarbons, dispatchable supply, and exportable fuel assumes that future growth will still rest more securely on extractive abundance, conventional energy strength, export leverage, and a more familiar relationship between energy production, hard power, and economic scale. But it remains more exposed to fuel-linked volatility and geopolitical transmission.

That is why the two camps so often talk past one another. They are not simply advocating different technologies. They are backing different models of growth under different assumptions about which uncertainty is more manageable. One side is more willing to absorb infrastructure and coordination risk in exchange for reducing fuel exposure over time. The other is more willing to accept fuel and geopolitical exposure in exchange for preserving dispatchable strength and conventional energy leverage.

Each pathway also implies a different view of the state. The first assumes the state can coordinate complex infrastructure, tolerate long payback horizons, and sustain industrial direction across political cycles. The second assumes the state should lean harder on established energy strengths, minimise dependence on more fragile transition architectures, and preserve forms of energy power it already knows how to wield. That is why the dispute is so intense. It is not only about energy. It is about which model of growth, geopolitical posture, and state competence the country still finds believable.

The deeper diagnosis

What this episode ultimately reveals is that the United States is still unresolved on a basic strategic question: what kind of power it wants energy to provide, and what kind of system it trusts itself to build.

The problem is not just partisan disagreement. It is the absence of a stable hierarchy of risks. Fuel risk, policy risk, geopolitical risk, sustainability risk, and public climate concern are all active at once, but there is no durable consensus on which of these should dominate strategy. That is why the uncertainty environment matters so much. It is not merely elevated. It is plural. Different channels of stress intensify at different times, interact only partially, and point toward different strategic instincts.

That leaves the country in a structurally unstable position. It wants the language of industrial strategy, but often struggles with the durability such a strategy requires. It wants the language of resilience, but remains divided over whether resilience means reducing fuel exposure, preserving conventional energy strength, or simply keeping optionality alive in an unsettled world. It wants growth, but is still split over whether growth should be anchored in infrastructure-led electrification or in extractive abundance and fuel-linked leverage.

That is why repeated policy swings are so revealing. They do not just reshuffle projects. They expose an unresolved struggle over what kind of economy the United States is trying to become, what kind of industrial base it believes it can sustain, and what kind of geopolitical posture it wants energy to support. More than that, they suggest the country is still arguing over a more basic question: not just which energy pathway it prefers, but which uncertainty it fears most.

A more productive direction

A more productive direction begins by dropping the fiction that any pathway is dependency-free. The task is to become more explicit about which dependencies are tolerable, which are strategically dangerous, and which require a level of state coordination that politics may not reliably sustain.

That means drawing a clearer distinction between support and credibility. Supporting a sector rhetorically or financially is not the same as creating a policy regime durable enough for long-duration capital to trust. It also means being more honest about the system architectures each technology assumes. Offshore wind is not just offshore wind. Oil and gas are not just oil and gas. Each sits inside a wider chain of infrastructure, exposure, institutional demand, and geopolitical meaning.

It also means being clearer about risk ranking. A serious strategy cannot treat all uncertainty as equal. It has to decide which exposures are acceptable, which should be reduced first, and which forms of state coordination are actually believable under current political conditions. Until that issue is addressed more directly, the United States may keep mistaking strategic movement for strategic clarity. A state that keeps changing which dependencies it underwrites is not necessarily becoming more resilient. It may simply be getting better at redistributing its fragilities.

Executive thesis

VALCOE is a serious and useful advance over plain LCOE. It corrects the obvious mistake of treating every megawatt-hour as economically identical by adjusting plant-level cost for energy value, capacity value and flexibility value. That makes it a better screening metric for cross-technology competitiveness in evolving power systems. (IEA WEO 2018; IEA/NEA 2020; GEC Model inputs)

Helios Nexum sits one level closer to the asset. It does not ask only what a technology is worth in an average regional system. It asks what a specific solar + BESS project can actually deliver, sell and finance once hourly output, residual-load pressure, curtailment activation, price capture, battery operation and funding structure are all allowed to interact. In that sense it is better understood as a delivered-system-finance framework than as a rival headline metric.

The practical implication is straightforward. VALCOE remains highly useful for broad technology ranking and policy screening. Helios Nexum adds value where decisions become real: site selection, revenue diligence, lender analysis, curtailment risk translation and investment committee judgment. On those margins, it can exceed VALCOE because it endogenises some of the exact frictions that average-system metrics necessarily smooth away.

1. Why these frameworks should not be conflated

The electricity literature has gradually moved from simple plant-cost comparisons toward richer attempts to incorporate system interaction. LCOE remains transparent and intuitive, but it abstracts from timing, adequacy, flexibility, congestion and market value. System LCOE, market-value studies, LACE and VALCOE all emerged because the economic meaning of a megawatt-hour changes with the state of the system and with technology mix. (Ueckerdt et al. 2013; Hirth 2013; Joskow 2011; IEA/NEA 2020)

Helios Nexum belongs to a related but distinct branch of this evolution. Rather than asking how to improve a generic technology-comparison metric, it asks how to translate system stress into project cash-flow reality. That is a different analytical problem. It is closer to bankability analysis than to macro technology ranking.

This distinction also maps cleanly onto the five energy-system fundamentals set out in Delardas’ (2025) framework: demand-supply balancing, efficient resource allocation, reliability and security, transparency and fair competition, and affordability. Any metric that claims relevance to modern power systems should be judged by how well it reflects those fundamentals, not just by how elegant its algebra looks.

2. What VALCOE does well

In the IEA formulation, VALCOE begins with projected LCOE and then adjusts it for three service categories: energy value, capacity value and flexibility value. The purpose is to compare technologies with very different operational characteristics on a more economically meaningful basis than plant cost alone. (IEA WEO 2018; IEA/NEA 2020)

That move is conceptually important. A solar plant that mostly produces in low-price midday hours is not equivalent to a flexible gas turbine or a battery that can move output into scarce periods. By embedding hourly dispatch and simulated market outcomes, VALCOE tries to represent the system value of what each technology contributes rather than merely the engineering cost of producing electricity. (IEA WEO 2018; GEC Model inputs)

For policy and planning, this is a major improvement over plain LCOE. It gives planners a structured way to compare wind, solar, storage and thermal generation in the same frame. The IEA Batteries and Secure Energy Transitions report uses exactly that logic to show why solar PV paired with a four-hour battery can become more competitive than new coal or gas in major markets once value adjustments are applied.

VALCOE therefore remains strong on three dimensions: cross-technology comparability, high-level policy relevance and a recognition that energy value, adequacy and flexibility must all matter in low-carbon systems.

3. Where VALCOE remains structurally limited

Its limitations follow from its purpose. VALCOE is still an average-system competitiveness metric. It is built for planners and policy makers, not for the underwriting of a single project. The IEA itself is explicit that VALCOE does not capture every cost and benefit and can be improved by including additional elements such as network integration costs. (IEA WEO 2018)

That caveat matters more than it first appears to. If a technology looks competitive in a regional average but faces severe local congestion, weak capture prices, charging constraints, curtailed output, imperfect route-to-market arrangements or market-design barriers to monetising flexibility, then a developer and a lender face a different economic problem from the one VALCOE was designed to solve. (IEA WEO 2018; IEA/NEA 2020)

In other words, VALCOE answers: “Is this technology broadly competitive in this system?” It does not fully answer: “Will this specific project earn enough on delivered and monetisable MWh to clear financing thresholds under real operating conditions?” That second question is precisely where Helios Nexum begins.

4. What Helios Nexum is actually doing

Helios Nexum is best understood as a layered 8760 asset model with an explicit system-stress translation mechanism. It starts from hourly production and battery-operation logic, but then moves further than most project models by making curtailment and value capture conditional on grid conditions rather than treating them as simple haircut assumptions.

At its core, the framework links four blocks: hourly renewable production, country-specific grid-stress diagnostics, curtailment activation and project-finance translation. That chain is what differentiates it from both plant-level LCOE and average-system value metrics.

The workbook structure shows this clearly. Yield scenarios are derived from PVGIS time-series logic; residual mix tables combine demand and renewable output; grid parameters contain country-specific calibration values; and the financial statements and CFADS schedules carry the resulting delivered output through to project returns, debt metrics and investment judgment. In the Executive Summary, even LCOE is defined on delivered MWh rather than gross theoretical generation, which is a subtle but crucial choice for a congestion- and curtailment-aware framework.

5. Direct comparison: where Helios Nexum matches VALCOE and where it extends it

The most important extension is that Helios Nexum does not stop at value recognition; it carries that value deterioration or improvement through the exact variables that lenders and equity investors actually care about. In practice, that means curtailment and price capture are not abstract “system effects”; they become delivered MWh, realised revenue, debt service headroom and ultimately a changed view of whether an asset is robust, marginal or uninvestable.

6. The country-specific grid logic is where Helios Nexum most clearly surpasses VALCOE

A major differentiator is the calibration philosophy. The grid logic in Helios Nexum is not imposed as a single universal curve. Fragility thresholds, curvature and stress activation are customised by country or market context. The workbook’s GridParameters sheet assigns jurisdiction-specific values for variables such as the curtailment threshold and convexity parameter, reflecting different grid conditions rather than assuming one generic renewable-penetration response for all systems.

That matters because renewable stress is not homogeneous. Two systems can have the same average renewable share and still behave very differently because their demand shape, thermal flexibility, interconnection, balancing regime, price formation and renewable mix differ. A metric that uses a single broad value correction can miss those path dependencies. A calibrated stress framework can at least attempt to represent them.

The additional methodological note is equally important: the calibration is not merely judgmental. The framework uses country-specific best-fit curves linked to load conditions and regressions against load-tightness behaviour. In plain language, the shape of curtailment or stress response is allowed to reflect the observed character of the underlying system rather than being assumed a priori. This makes the metric less generic but more realistic for actual project screening.

That is one of the clearest respects in which Helios Nexum can surpass VALCOE. VALCOE recognises that value depends on system conditions. Helios Nexum goes further by allowing those conditions to alter the project pathway through calibrated thresholds, activation logic and non-linear absorption dynamics.

7. How the two frameworks compare against energy-system fundamentals

Against demand-supply balancing, Helios Nexum is strong because the balancing problem is visible inside the metric itself. Residual-load pressure, load tightness and storage behaviour can directly alter what is physically deliverable and economically saleable. VALCOE captures balancing more indirectly via value streams. (Delardas 2025)

Against efficient resource allocation, both frameworks add value beyond plain LCOE, but in different ways. VALCOE improves technology ranking at system level. Helios Nexum improves capital allocation at project level by highlighting where headline-competitive technologies may still be poor investments once location, congestion and financing are recognised. (Delardas 2025; Hirth 2013)

Against reliability and security, VALCOE still has an advantage on generic adequacy representation because capacity and flexibility value are explicit components of the metric. Helios Nexum addresses reliability more through delivered economics and operational resilience than through a separate adequacy market construct. (IEA WEO 2018; Batteries and Secure Energy Transitions)

Against transparency and fair competition, Helios Nexum has a clear edge. The causal chain from assumptions to outcomes is visible to the user. A planner or lender can inspect how changes in residual load, thresholds, capture prices or battery sizing alter the result. That is closer to operational transparency than a published but non-fully-reproducible model average. (Delardas 2025)

Against affordability, the answer depends on whose affordability is meant. VALCOE is useful for long-run system planning because it identifies which resources are broadly cost-effective under evolving system needs. Helios Nexum is more useful for affordability in the sense that failed or mispriced projects eventually show up as higher financing costs, delayed build-out or poor allocation of capital. On that margin, better project realism can itself support lower system cost over time.

8. A real-conditions example: when the frameworks diverge

Consider a utility-scale solar + four-hour BESS project in Great Britain in the late 2020s. At national level, a VALCOE-style assessment may still find solar + storage highly attractive because the battery improves energy timing, adds flexibility value and supports adequacy relative to stand-alone solar. That broad conclusion is plausible and in line with the IEA’s paired-storage competitiveness logic. (Batteries and Secure Energy Transitions)

Now move from the system average to the actual project. Suppose the site faces recurring midday residual-load compression during high renewable output periods, limited local headroom, imperfect route-to-market arrangements and only partial ability to stack ancillary revenues. In Helios Nexum, those features do not remain background context. They can directly change curtailment incidence, captured spreads, delivered MWh, realised average price, LCOE on delivered energy, CFADS and debt-service resilience.

Under those conditions, both frameworks can be “right” while producing different answers. VALCOE can correctly say that solar + storage is an attractive technology class in Great Britain. Helios Nexum can simultaneously show that this particular project, at this location and under this revenue stack, is only marginally bankable. The difference is not contradiction; it is analytical resolution.

For real project decisions, the Helios answer is the more decision-relevant one. Projects fail or succeed through delivered energy, captured price and financing, not through average technology attractiveness. For broad power-system planning, however, VALCOE remains the cleaner first-pass lens.

9. Where Helios Nexum should still remain modest

A balanced conclusion requires acknowledging where Helios Nexum should not over-claim. It is not a universal replacement for system-level competitiveness metrics across all technologies. It is purpose-built for solar + BESS and for the translation of system conditions into project economics. That focus is a strength, but it narrows comparability relative to VALCOE.

It is also still a proxy-rich framework. Even though its logic is transparent and country-calibrated, it does not solve full nodal power flow, reserve co-optimisation or every market-rule nuance. Its power lies in disciplined realism, not in pretending to be a complete power-system simulator.

That said, none of these caveats reduce its contribution. They simply define it correctly: Helios Nexum is most powerful when used for the class of decisions it was built to improve.

10. Final conclusion: where Helios Nexum stands

By the end of the comparison, Helios Nexum should be understood as sitting downstream of VALCOE rather than outside it. VALCOE solves the problem of recognising that technology value depends on system context. Helios Nexum solves the next problem: how that context propagates through a real solar + BESS project and changes the financial verdict.

That is why Helios Nexum fits naturally within the post-LCOE literature while also moving beyond VALCOE on several margins. It is more temporally explicit, more transparent at asset level, more sensitive to local grid fragility, more explicit about curtailment activation, and far stronger in converting system stress into finance metrics that matter for developers, lenders and investors.

The fairest verdict is therefore not that Helios Nexum invalidates VALCOE. It is that Helios Nexum adds a missing industrial layer. It connects electricity-system fundamentals to investable reality. In a power system where renewable saturation, congestion, curtailment and value erosion increasingly determine outcomes, that is not a cosmetic refinement. It is a material analytical contribution.

Selected references

Delardas, O. (2025). Transforming marginal-cost pricing: A blockchain approach to fundamental energy system dynamics. Utilities Policy.

International Energy Agency (2018). World Energy Outlook 2018.

International Energy Agency and OECD NEA (2020). Projected Costs of Generating Electricity 2020.

International Energy Agency (2024). Batteries and Secure Energy Transitions.

International Energy Agency (2025). Global Energy and Climate Model: Techno-economic inputs.

Ueckerdt, F. et al. (2013). System LCOE: What are the costs of variable renewables? Energy.

Hirth, L. (2013). The market value of variable renewables. Energy Economics.

Joskow, P. (2011). Comparing the costs of intermittent and dispatchable electricity generating technologies. AER Papers and Proceedings.

Juelch, V. (2016). Comparison of electricity storage options using levelized cost of storage. Applied Energy.

Zakeri, B. and Syri, S. (2015). Electrical energy storage systems: A comparative life cycle cost analysis. Renewable and Sustainable Energy Reviews.

The United States is entering a new electricity era, and it is doing so without the comfort of a single unified grid, a single planning logic, or a single market design. That matters more than most headlines admit.

The US electricity system is already being pulled into a new demand regime, and data centres are a serious part of the story. The latest detailed DOE/LBNL estimate shows data centres consuming about 4.4% of total US electricity in 2023, up from 58 TWh in 2014 to 176 TWh in 2023, with demand projected to reach roughly 325 to 580 TWh by 2028, or around 6.7% to 12% of total US electricity use. More recently, the US Energy Information Administration said it now expects the strongest four-year growth in US electricity demand since 2000, with data centres helping to drive that acceleration.

Beyond the official baseline, some analysts are now sketching much larger upside scenarios. Bismarck's recent note suggested that the current US data-centre construction pipeline could, if fully built, add as much as 140 GW of load against a current US peak demand of roughly 760 GW. Read carefully, that is less a forecast to take literally than a sign of how large the planning problem is becoming.

At the same time, EIA says retail electricity prices have risen faster than inflation since 2022 and are expected to continue increasing through 2026.

That combination has encouraged a familiar genre of commentary: AI demand is exploding, the grid is old, the politics are bad, and power is becoming unaffordable. There is truth in that. But it still does not get to the centre of the problem.

The deeper issue is that parts of the US power system are struggling to translate new supply, new investment, and new demand into affordable, reliable delivered power. This is not simply a story about expensive generation. It is a story about system readiness lagging behind system ambition.

The first mistake is to talk about “the US grid” as if it were one thing

It is not. The United States is split across major interconnections and a patchwork of regional transmission organisations, independent system operators, vertically integrated utilities, public utility commissions, and different retail structures. That fragmentation is not just background detail. It is central to why the current moment is becoming harder to manage.

FERC’s own transmission-planning reforms can reasonably be read as recognition that the existing institutional arrangements have not kept up with the grid’s evolving needs. Order No. 1920 and its follow-on orders require long-term regional transmission planning, minimum scenario analysis over a 20-year horizon, and more explicit state involvement in planning and cost allocation. In other words, the regulator is not acting as though today’s grid can simply drift into the next phase of demand growth on existing planning habits.

That should already shift the tone of the debate. The question is not whether one technology or one politician has “broken” the American grid. The question is whether a highly diversified but fragmented electricity architecture can coordinate quickly enough for what is now coming.

PJM is not the whole American story, but it is one of the clearest warning signs

If the article needs one empirical anchor, PJM is the obvious choice. It is the country’s largest synchronized wholesale power market and a place where several pressures now meet at once: data-centre exposure, transmission constraints, reserve needs, and visibly changing market prices.

In PJM’s 2026/27 capacity auction, the clearing price hit the FERC-approved cap of $329.17/MW-day across the RTO footprint, up from $269.92/MW-day in the 2025/26 auction, while PJM procured 134,311 MW of generation resources for the delivery year. That does not mean the system is “running out of power.” It does mean that dependable availability is being repriced more aggressively.

In plain English, a high capacity price is not the same thing as a high energy price. It is the price of securing enough committed resources in advance to meet future peak demand and reserve requirements. PJM itself describes its capacity market as a mechanism to ensure enough power is procured three years ahead, with resources paid for the promise to be available when called upon. The auction accepts lower-cost offers first and keeps going until enough capacity is assembled to meet the projected reserve requirement; all cleared sellers then receive the marginal clearing price.

That distinction matters because it immediately complicates the simplistic claim that “electricity is expensive because generation is expensive.” In PJM’s own description of its wholesale markets, the energy market made up about 59% of wholesale electricity cost in 2024, the capacity market about 7%, and ancillary services about 3%. Capacity is not the whole bill. But nor is it incidental. It is part of the cost of making a system dependable under stress.

Cheap generation and cheap bills are not the same thing

This is still the central confusion in much of the public debate.

The EIA states plainly that retail electricity prices include the cost of generating, transmitting, and delivering electricity to end users, as well as taxes and other fees. It also notes that utility spending on electricity distribution has surpassed spending on transmission and production, and that the generation-related portions of retail electricity prices can lag changes in wholesale fuel and power prices depending on contracts and tariffs.

That means a region can add cheaper generation and still deliver expensive electricity to households. It can lower the marginal cost of producing a megawatt-hour while raising the total cost of moving, balancing, reserving, recovering, and financing the system around it. It can also improve the supply stack while leaving the retail bill under pressure.

This is a point I have made in my own published work as well: average consumer rates do not simply reflect fluctuating wholesale prices, nor do they transparently capture the costs associated with maintaining and upgrading transmission and distribution networks.

That is not a rhetorical trick. It is a basic distinction between generation economics and system economics.

Why the retail bill can keep rising even when generation gets cheaper

The EIA’s own retail-pricing explanation is useful here. Residential and commercial customers tend to pay higher prices per kilowatt-hour than industrial users because it costs more to distribute electricity to them, while industrial customers often receive power at higher voltages and more efficient delivery arrangements. The point is not merely that electricity is expensive. It is that different parts of the system recover different costs in different ways, and those costs are often embedded in the final bill long after the public conversation has moved on.

EIA’s modelling documentation makes the same point in a more technical form. Even in market-based regions, transmission and distribution are generally treated as remaining under rate-of-return regulation, with revenue requirements determined through cost recovery frameworks that include the return on the rate base, fixed operating costs, and depreciation of invested capital. In other words, a large part of the end-user bill is not simply “whatever the market price of electricity happened to be.” It is also the institutional recovery of the infrastructure built around that electricity.

That is one reason the relationship between wholesale prices and final bills is not linear, and one reason a debate that focuses only on the cost of generation will keep missing the system layer.

Data centres are not the whole story, but they are revealing the system’s bottlenecks

There is a temptation now to turn AI into the sole culprit. That is too neat.

The more serious reading is that a step change in power demand is arriving in a system that was already facing ageing infrastructure, slow transmission expansion, local bottlenecks, interconnection delays, and increasingly difficult reliability trade-offs. Data centres did not create the underlying structure of those weaknesses. They are making them harder to ignore.

That is also why the conversation is maturing. It has to. Once a country begins to believe that the next generation of industrial and digital capability depends on much higher electricity throughput, the grid ceases to be a background utility and becomes an active economic constraint. The demand story is no longer speculative. DOE and EIA are both now signalling that it is material.

That does not mean every load forecast will fully materialise. It does mean the hedge is already in. The market, policymakers, and infrastructure planners are behaving as though the next phase of demand growth is real enough to matter now.

Transmission is not a side issue any more

One of the clearest signs that the debate is moving beyond slogans is the growing prominence of transmission reform. FERC’s transmission-planning rule requires long-term regional transmission planning using multiple scenarios, greater transparency over costs and benefits, and cost-allocation methods intended to be at least roughly commensurate with estimated benefits. That points to a regulator that recognises the old planning cadence is under strain relative to the scale and pace of the coming grid transition.

That is crucial because transmission is the hinge between supply ambition and delivered outcomes. If generation expands faster than transmission, congestion rises. If local network reinforcement lags behind load growth, delivered costs rise. If the system adds variable resources without enough flexibility and transfer capability, balancing stress rises. This is not theory. It is the practical meaning of a grid being asked to do new things with insufficient structural adaptation.

On locational marginal pricing: useful, but not sufficient

The strongest defence of locational marginal pricing is that it prices electricity where it actually is, under the physical constraints of the network, rather than pretending all megawatt-hours are equally deliverable. CAISO’s tariff formulation makes this explicit: its LMP is composed of a system marginal energy component plus a marginal congestion component and a marginal losses component. PJM similarly presents LMP as a way of reflecting the cost of making and delivering electricity in real time across the transmission system.

That is the defender’s case, and it is not trivial. LMP does embed local constraints, losses, and congestion in a way that coarser pricing systems often do not.

But that still does not make LMP a full answer to the wider system problem. In my own paper, I argue that nodal systems can generate high volatility and congestion rents, and that price separation between nodes does not always serve as a reliable or intuitive long-run investment guide in the way many assume. I also note that scarcity-driven adders and extreme conditions can distort the signal in ways that complicate welfare interpretation.

The point is not that LMP is irrational. The point is that precise local pricing is not the same thing as system coherence. A market can price constraints accurately and still fail to build wires quickly enough, accredit capacity convincingly enough, or align local and regional planning with the speed of demand change.

Capacity and balancing are where the argument gets more serious

In wholesale-power debates, there is often too much focus on energy and not enough on the cost of being ready, or the cost of correcting reality when forecasts fail. Yet balancing markets exist because actual supply and demand do not line up perfectly with contracted schedules, and capacity mechanisms exist because marginal energy prices alone do not always provide stable enough incentives for future adequacy. PJM’s own market explainer is very clear on this distinction: the energy market serves near-term electricity needs; the capacity market prepares for future demand; ancillary services help maintain short-term balance and reliability.

That is also consistent with the argument in my published paper. There, I explain that capacity markets emerged precisely because marginal energy pricing, while useful for operational efficiency, does not always ensure full long-run recovery for resources needed under stress, while balancing mechanisms exist to deal with the real-time mismatch between contracted and actual outcomes.

Once that is understood, the phrase “cheap power” becomes much less straightforward. A technology may be cheap to generate with. But if it increases the system’s demand for balancing, flexibility, reserves, local reinforcement, or cost recovery, then the end-user outcome will depend on how those layers are designed and paid for.

That is not an anti-renewables point. It is what integration means.

PJM, ERCOT and California are not the same problem wearing different clothes

It is important not to overgeneralise from PJM.

PJM uses a forward capacity market and highly developed locational pricing. ERCOT is different. EIA notes that ERCOT is unique among balancing authorities in not having a centralized capacity market; many other ISO/RTO regions use capacity markets as a separate mechanism to encourage investment and ensure adequate reserves. In that 2019 example, EIA also noted that ERCOT’s wholesale prices rose sharply during summer tightness, reaching the then-$9,000/MWh cap for several hours during stress conditions.

More recently, NERC’s 2025 Summer Reliability Assessment said ERCOT expected sufficient operating reserves for the August peak hour under normal conditions, but still faced a higher risk of emergency conditions in evening hours when solar generation ramps down and loads remain elevated; it also pointed to South Texas transmission constraints as a continuing system limit.

California is different again. CAISO’s market architecture is deeply nodal, and NERC’s 2025 assessment did not place WECC-California in the same elevated-risk category as some other areas for summer adequacy, noting that reserve margins were not anticipated to fall below the reference margin in the upcoming summer assessment. Yet California’s system still reflects the same broader structural theme: local constraints, congestion, balancing needs, wildfire-related transmission risk, and the value of a wider day-ahead market footprint for efficiency and renewable integration.

So the right conclusion is not that PJM explains America. It is that PJM shows one especially visible form of the broader American challenge.

The transition is hitting system reality

The United States is not just adding cleaner generation into a stable demand environment. It is moving into a more electricity-intensive economy altogether. Data centres, artificial intelligence, electrification, storage, transmission reform, and changing reserve needs are not separate stories. They are parts of one restructuring.

That is why the debate needs to mature.

It is no longer enough to say one technology is cheap, another is reliable, another is clean, another is domestic, another is politically attractive. The real question is whether the system as a whole can absorb rapid demand growth, preserve reliability, maintain investability, and deliver electricity at costs households and firms can tolerate.

That question will not be answered by pretending the United States is one uniform market. It will not be answered by treating wholesale prices as if they were retail bills. And it will not be answered by assuming that precise pricing at the node automatically implies coherent planning at the system level.

What is emerging is harsher and more interesting than that.

We are watching the next generation of power demand collide with the physical, financial, and institutional realities of the grid.

Conclusion

Electricity may well become more expensive in many parts of the United States over the next several years. But the deeper story is not just expensive power.

It is the widening gap between generation growth and system readiness.

PJM is one warning sign. Data-centre demand is another. Transmission reform is another. Rising retail prices are another. Taken together, they suggest that the American power problem is not simply a shortage of generation, nor simply a failure of one policy camp over another.

It is a system under pressure to do more, faster, with institutions and infrastructure that still struggle to align.

Cheap generation is not the same thing as cheap delivered power.

And until that distinction becomes central to the public conversation, the United States will keep misreading a system problem as if it were only a fuel story, a technology story, or a political story.

Most solar plus storage models still behave as if the hard part is estimating energy.

I do not think that is the hard part anymore.

The hard part is translating energy into bankable, resilient, investor-relevant cash flow once you admit that modern power systems are increasingly shaped by congestion, curtailment, negative-price risk, battery degradation, and financing discipline. That shift is becoming more visible as photovoltaic (PV) penetration rises and flexibility becomes more commercially important.

That is the problem I have been trying to solve with my 8,760-hour Solar + Battery Energy Storage System (BESS) model.

This is not a spreadsheet built around one annual yield number and a few pricing assumptions. It is built around a different premise: the commercial asset is not nameplate capacity. The commercial asset is the net delivered and monetisable energy profile that survives through system conditions, storage behaviour, and capital structure constraints.

The real asset is not nameplate capacity

A solar plant with a co-located battery does not earn money because its installed capacity looks impressive in a model summary.

It earns money if generation is actually absorbable by the grid, if the battery can shift value without degrading too quickly, and if the resulting cash flow is strong enough to support debt and equity returns under downside assumptions.

That sounds obvious, but a lot of models still stop too early. They model generation. They do not model the system that generation is landing into.

I wanted to build something that starts one layer deeper.

Why I moved to an 8,760 architecture

The core of the model is hourly.

Not because hourly detail is automatically superior, but because a lot of the economics that matter in solar plus BESS projects disappear when everything is averaged into a single annual number.

At hourly resolution, the relevant questions change. You can see when solar is actually available, when the system is more likely to absorb it, when it is at risk of curtailment, when storage genuinely helps, and how much delivered energy is really left after system friction. That is the difference between modelling output and modelling behaviour.

And behaviour is where value now lives.

The key assumptions are organised around how value is created and lost

What matters in advanced solar plus BESS modelling is not the number of tabs. It is whether the assumptions are organised around the actual pathways through which value is created, lost, protected, or financed.

In my model, the key assumptions sit in five layers.

1. Resource and yield assumptions

This is the starting point.

The model uses PV resource history and constructs multiple yield cases rather than relying on a single deterministic annual number. In the pro workbook, the yield-factor layer derives central and downside cases from historical PV yield patterns rather than simply assuming that Probability of Exceedance 90 (P90) is some arbitrary percentage below Probability of Exceedance 50 (P50).

That matters because weather uncertainty is not just a technical footnote. It affects revenue variance, debt sizing, and debt service resilience. Lender-oriented solar guidance has long linked downside yield cases and coverage analysis to debt capacity, because debt is serviced by cash flow, not by optimistic average production.

2. Grid absorption and curtailment assumptions

This is where the model stops behaving like a conventional solar template.

Instead of assuming that every generated megawatt-hour is equally exportable, the model uses grid-side proxy mechanisms to estimate how much solar can actually be absorbed in each hour. These proxies are not trying to replicate a full nodal market model. They are trying to capture the fact that exportability depends on broader system conditions, not just plant output.

That distinction matters more as PV penetration rises. The International Energy Agency Photovoltaic Power Systems Programme (IEA PVPS) has highlighted that congestion, oversupply, curtailment, and negative prices are becoming more material in higher-penetration markets, and that flexibility, storage, interconnection, and alternative services are increasingly important to preserve profitability.

3. Battery operating assumptions

The BESS is not treated as a generic uplift factor.

It is governed by explicit assumptions around power, duration, minimum and maximum state of charge, target state of charge, charge and discharge efficiency, auxiliary consumption, cycling behaviour, equivalent full cycles, end-of-life capacity, and degradation over time.

This matters because storage value only means something if the physical wear required to create that value is also reflected in the model. National Renewable Energy Laboratory (NREL) work on hybrid plants and degradation-aware control makes the same general point: storage dispatch and degradation cannot be separated if the objective is realistic lifetime value rather than short-term revenue optics.

4. Revenue and pricing assumptions

The model does not stop at megawatt-hours. It converts delivered energy into financial outcomes through price and capture assumptions.

That means the relevant question is not just “what is the market price?” but “what price is this asset actually likely to capture once timing, curtailment, storage shifting, and system saturation are taken seriously?”

As more solar enters a system, average market price and realised asset price increasingly diverge. That is one reason why flat annual merchant assumptions are becoming less informative in high-solar systems.

5. Financing and downside assumptions

The final layer is the capital structure.

The model separates the case used to show base operating performance, the case used for debt sizing, and the case used for covenant stress. That is fundamental. A project can look healthy in a central case and still support less debt than expected once downside resilience is tested.

The debt service coverage ratio (DSCR) remains one of the central underwriting constraints in utility-scale renewable project finance, and NREL benchmarking still uses explicit DSCR assumptions for utility PV and utility PV-plus-battery cases.

How the grid proxy mechanisms work conceptually

This is one of the most important parts of the model, because it is where the structure stops behaving like a simple yield-times-price template.

A lot of renewable models treat curtailment as a flat percentage. I do not think that is enough.

Curtailment is a system outcome. It emerges when generation arrives into a system that cannot fully absorb it, whether because of network bottlenecks, timing mismatch, oversupply, or insufficient flexibility. That is why I use a set of grid proxy mechanisms rather than a single blunt haircut.

At a high level, the model works through three linked ideas.

Residual mix

The model uses a residual supply-demand style signal to represent how crowded or constrained the system is in each hour.

Its purpose is to distinguish between hours when the grid is more likely to accept additional solar generation and hours when solar is more likely to be pushed toward spill, curtailment, or weaker pricing.

Grid fragility or congestion parameters

The model then applies shaping parameters to convert that system condition into a solar absorption profile.

Conceptually, this is doing something important: it allows the model to say that the same megawatt-hour of solar is not equally valuable in every hour. Its commercial relevance depends on the condition of the surrounding system.

Solar absorption versus solar at risk

Once those proxies are applied, generation is split into solar absorbed by the grid, solar at risk before storage, solar that can potentially be shifted by the BESS, and solar ultimately curtailed.

That sequence matters because it stops the model from treating storage and curtailment as independent. In reality, they are part of the same commercial chain.

Storage is not just a revenue uplift

I do not think it is good enough to treat a battery as an arbitrage attachment that simply improves the blended price.

A BESS is a degrading asset. It has cycling limits, charge and discharge constraints, minimum and maximum state of charge, round-trip losses, auxiliary loads, and a replacement profile that eventually becomes financial, not just technical.

That is why my BESS module explicitly carries state-of-charge constraints, charge and discharge logic, round-trip efficiency, auxiliary consumption, equivalent full cycles, end-of-life capacity thresholds, cycle fade, and maintenance and replacement capital expenditure.

If storage value is real, battery wear is real too. A serious model has to carry both. That is consistent with broader NREL work showing that degradation-aware storage scheduling can materially affect lifetime battery outcomes.

Advanced modelling starts with separating the cases

A lot of models use one scenario selector for everything, as if the case that tells the best story should also size the debt and test the covenants.

I do not think that is how project finance works.

In my model, the displayed base operating case is separated from the debt sizing case and the covenant stress case. That means I can ask three distinct questions:

what does the project look like in a central case, how much debt can it really support under a more conservative yield case, and does the structure survive a deeper downside?

Those are not the same question.

That distinction is critical because P50 is not the same thing as bankability. Financing quality depends on downside resilience, not base-case optimism.

Debt should be earned by cash flow available for debt service (CFADS)

This is another area where many models stay too shallow.

I do not like starting with “let’s assume 70% debt” and then building the project around that. I prefer debt to emerge from the cash flow.

That is why the model sizes senior debt from cash flow available for debt service (CFADS), target DSCR, tenor, fees, reserve requirements, and leverage caps. In other words, leverage is not just chosen. It is earned by the resilience of the cash flow.

That approach aligns with standard project-finance logic: DSCR measures the project’s ability to meet debt service from operating cash flow, while coverage metrics such as loan life coverage ratio (LLCR) and project life coverage ratio (PLCR) are used to judge credit robustness.

It also helps expose something simpler templates often hide: a project can look acceptable in project internal rate of return terms and still be unconvincing for equity once the financing structure and hurdle rates are imposed.

What the template inputs actually give you

The practical value of the template is that it allows a user to move from a manageable set of technical, commercial, and financing inputs into a much richer decision framework.

At input level, the template lets you define things such as:

• project location and PV resource profile

• PV capacity and yield cases

• battery power, duration, and degradation settings

• grid proxy and curtailment sensitivity assumptions

• pricing, capture, and contract structure assumptions

• capital expenditure, operating expenditure, and tax settings

• debt sizing case, DSCR, tenor, and reserve assumptions

From those inputs, the model gives you a structured view of:

• gross versus net delivered energy

• absorbed, shifted, and curtailed solar volumes

• battery cycling, degradation, and sustaining capital needs

• revenue and cash flow under different operating cases

• debt capacity and sculpted debt service

• DSCR, LLCR, and PLCR

• project internal rate of return and net present value

• equity internal rate of return and net present value

• downside robustness across base, debt, and stress cases

That, to me, is the point of advanced solar plus BESS modelling.

Not complexity for its own sake.

A better translation from physical system behaviour into investable financial reality.

References

• IEA PVPS, Snapshot of Global PV Markets 2025.

• IEA PVPS, Snapshot 2025 overview page.

• NREL, Financial Cases & Methods | Electricity | 2024 ATB.

• NREL, Utility-Scale PV-Plus-Battery | Electricity | 2024 ATB.

• IFC / World Bank, Utility-Scale Solar Photovoltaic Power Plants.

• NREL, Modeling and Analysis of Clean Energy and Storage Systems.

• NREL, Battery Degradation Modeling in Hybrid Power Plants.

• ERIA, Financing Solar PV Projects chapter with coverage-ratio discussion.

The politics of UK energy bills always invites neat stories. Drill more and bills fall. Build renewables and bills fall. Scrap climate policy and bills fall. Redesign tariffs and bills fall.

The Oxford Smith School’s March 2026 rapid analysis steps directly into that terrain. Its central claim is straightforward: maximising North Sea oil and gas extraction is a weak strategy for materially reducing household energy bills, while a clean-energy and electrification pathway could deliver larger savings over time. It models North Sea household benefit mainly through fiscal redistribution and compares that with a renewables-led electricity system, electrified heat, and in the highest case, bill rebalancing through general taxation.

That core direction is not frivolous. In fact, the report is strongest precisely where it resists the most simplistic version of “drill more, bills fall.” But once the paper moves from that broad directional point into a single headline comparison, its structure starts doing more work than the analysis can fully support. The comparison is not just about evidence. It is also about framing, scale, policy ambition, and what gets counted as a “bill saving” in the first place.

So the most useful way to read the note is not as a final verdict on North Sea versus renewables, but as a bundle of claims that need to be separated, tested, and in some cases reconstructed.

What follows is a 5 Claims / 5 Critiques / 5 Reconciliations review: not an attempt to dismiss the report, but an effort to identify where it is strongest, where it overreaches, and how its argument could be made more rigorous.

Claim 1: More North Sea extraction will not materially reduce UK household bills

This is the paper’s anchor argument. The report says that more domestic extraction does not mean materially cheaper household energy because UK oil and gas are sold into market-based pricing systems. It uses Brent for oil and NBP for gas, but largely folds both into the broader claim that hydrocarbons are priced through markets larger than the UK alone. It also uses January 2026 prices as a deliberately conservative pre-shock baseline.

The broad direction holds. Domestic extraction is not the same thing as domestic price control.

But the compression matters. Oil and gas are not structured in the same way. Brent is a global crude benchmark. UK gas sits within the more regionally mediated NBP / northwest European gas system, shaped by LNG availability, Norwegian pipeline flows, storage constraints, and interconnector conditions. So the report is more persuasive as an argument against naive domestic-price determinism than as a complete account of gas-market policy leverage.

There is also a methodological asymmetry hidden inside the “conservative baseline” language. A lower Brent price cleanly weakens the North Sea fiscal case. A lower NBP price is more contingent: it reflects a particular regional configuration of storage, LNG, weather, and pipeline conditions in early 2026 that may not be representative of a 20-year extraction horizon. In other words, the conservative-baseline framing is cleaner for oil than for gas.

A stronger version of the claim would therefore split oil and gas twice over: first by market structure, and second by baseline logic. It could preserve the report’s core point while admitting that the route from domestic production to prices is flatter for oil and somewhat more regionally mediated for gas.

Claim 2: The main consumer value of North Sea production is fiscal rather than physical

Here the report is explicit. The household benefit of North Sea production is modelled mainly through the state’s fiscal take being redistributed across 28.6 million households, yielding around £82 per household in the “realistic” case and £16 in the “no windfall tax” case. This is presented as a transfer-payment style mechanism spread over the extraction lifetime.

That is a valid policy channel. There is nothing inherently second-rate about resource-rent capture and redistribution if the objective is household welfare. The deeper problem lies elsewhere.

The paper presents itself as answering a single policy question about which strategy better reduces household bills, yet the two sides of the comparison operate at very different levels of ambition and complexity. The North Sea side is a relatively narrow rent-distribution mechanism. The renewables side is a much broader system-transition pathway involving generation mix change, electrification, and eventually bill rebalancing. The asymmetry is therefore not cosmetic. It is built into the structure of the comparison.

And there is a second problem still. The report implicitly stages North Sea fiscal revenues and clean-energy transition costs as if they were competing policy pathways. But the paper itself says electrification depends on upfront investment and targeted support. There is no analytic reason why a portion of declining North Sea fiscal revenues could not help capitalise exactly those transition bottlenecks. The either/or structure is therefore doing political work that the analysis does not itself require.

A more rigorous version of this section would disaggregate the policy question into direct bill effects, fiscal-transfer effects, and whole-system transition effects. It would also include a hybrid pathway, in which declining North Sea rents help finance the transition constraints the report identifies.

Claim 3: A renewables-led electricity system lowers bills because renewables can replace gas as the marginal price setter

This is the intellectual bridge from “renewables are cheap” to “households save money.” The report argues that if renewables dominate power generation, they can displace gas as the price-setting technology. It uses AR7 CfD strike prices as a benchmark for renewable electricity and estimates savings of roughly £105 under partial electrification and £330 under full electrification. It also explicitly says it does not explicitly include intermittency costs.

This is where the paper becomes noticeably thinner.

The note leans on a stylised CfD / strike-price / low-generation-cost logic without properly carrying through the broader whole-system questions that determine whether lower generation costs become lower delivered costs. Balancing costs, congestion, profile mismatch, backup capacity, flexibility procurement, and the cannibalisation effect at higher renewable penetration are all either lightly treated or effectively externalised.

More importantly, this is not just a missing technical caveat. It is a market-design problem. Whether lower renewable generation costs flow through to retail bills depends on institutional choices: the future of marginal pricing, the possibility of zonal reforms, the treatment of balancing and locational costs, and the retail mechanisms through which wholesale changes are transmitted to consumers.

That matters because market design is not a background assumption. It is a contested political variable.

More importantly, this is not just a missing technical caveat. It is a market-design problem. Whether lower renewable generation costs flow through to retail bills depends on institutional choices: the future of marginal pricing, the possibility of zonal reforms, the treatment of balancing and locational costs, and the retail mechanisms through which wholesale changes are transmitted to consumers. That matters because market design is not a background assumption. It is a contested political variable.

As I argued in my published paper, Transforming marginal-cost pricing: A blockchain approach to fundamental energy system dynamics, proposals built around LCOE and fixed-price framings often overlook why marginal-cost pricing exists in the first place. They can flatten scarcity, misprice reliability, understate integration and balancing costs, and miss the system-value problem that emerges as variable generation scales. On that view, the Oxford note’s weakness is not just that it leaves out a technical detail. It is that it assumes lower generation costs can pass through to households without first resolving the market-design question that determines how those costs are discovered, balanced, and allocated.

A stronger version of Claim 3 would therefore treat market design as a core uncertainty variable, not a footnote. It would say that a renewables-led system may reduce bills, but the extent and transmission of those gains depend on unresolved choices about pricing architecture, balancing arrangements, network charging, and retail pass-through, not simply on low renewable generation costs.

Claim 4: Electrification is the mechanism that unlocks the larger household savings

The report is clear that renewables alone are not the whole story. It notes that only about 30% of household energy use currently comes from electricity and that most households remain dual fuel. The larger savings come when households electrify heat, especially through heat pumps, and it acknowledges that this requires upfront investment and support mechanisms, particularly for lower-income households.

That mechanism is real. But the report still understates the difficulty of making it real.

The problem is not only financing and subsidy design, though those matter. It is also a distributional and physical retrofit problem. The households for whom lower bills matter most are often those in older, harder-to-insulate homes, with the weakest balance sheets and the least ability to absorb upfront retrofit costs. Yet even with perfect subsidy design, a large portion of the UK housing stock, especially older solid-wall and pre-war properties, presents physical retrofit constraints that money alone does not quickly solve. This is also a skills, planning, and supply-chain problem unfolding over a longer horizon than the report’s arithmetic suggests.

So financing is necessary, but not sufficient.

A stronger version of Claim 4 would make electrification conditional on three things: financing access, targeted subsidy design, and physical delivery capacity across retrofit execution, installer skills, planning, and supply chains. That would turn electrification from a generic mechanism into a more credible political-economy and delivery claim.

Claim 5: The largest household savings come from a clean-energy system plus bill rebalancing

The report’s highest figure, around £441 per household, depends not only on renewables and electrification, but also on moving policy costs from electricity bills into general taxation and increasing the wholesale share of the bill.

This is probably the sharpest technical issue in the paper.

Part of the apparent saving is not a reduction in underlying energy-system cost, but a reallocation of costs outside the electricity bill. That may still be sensible policy. It may even be good policy. But it is not the same thing as showing that the energy system itself is cheaper by the full quoted amount.

There is also a direct comparability problem. In the North Sea scenario, the report largely holds policy and network costs constant. In the top-end renewables case, it then changes bill architecture itself. So the framework is not like-for-like. One side is largely a fiscal transfer layered onto existing bill structure. The other becomes part system change, part tariff redesign.

A stronger version of this claim would therefore separate system-cost savings from bill-design savings. It should say clearly that part of the gain comes from cleaner electricity and electrification, while part comes from a political decision to shift policy costs into taxation.

The structural weakness running through all five claims

Running beneath the whole paper is one further issue that affects every comparison it makes.

The report repeatedly uses 28.6 million households as a single denominator and explicitly says values are distributed evenly across all households. It also says it did not distinguish between low-income and other households.

That is a major weakness because the paper’s stated concern is the burden on vulnerable and low-income households, yet the arithmetic assumes household homogeneity. This matters across the whole framework. North Sea fiscal transfers may look different if targeted. Electrification savings are much harder to realise for poorer households in inefficient housing. Tax-based bill rebalancing will have different incidence depending on who ultimately pays.

In other words, the headline averages do not merely simplify. They risk obscuring the very distributional question the paper claims to care about. A quintile-based analysis could produce materially different conclusions.

Overall synthesis

The Oxford note is strongest when read as a directional macro critique of the most simplistic North Sea billing politics. More domestic extraction is not a serious direct lever for structurally lower household bills in a market-priced system. On that, the report is broadly persuasive.

It becomes weaker when it turns that directional insight into a single quantified policy comparison.

At that point, the paper starts compressing differences that matter. It compresses the distinction between oil and gas market structure. It uses a “conservative baseline” that is methodologically cleaner for Brent than for NBP. It compares a narrow fiscal-rent pathway with a broader transition pathway. It treats market design as a technical qualifier rather than a live political variable. It understates both the distributional and physical barriers to electrification. It mixes system-cost change with bill reallocation. And it uses whole-population averages despite explicitly acknowledging household heterogeneity.

So the fairest reconciled version of the report’s argument is something like this:

North Sea expansion is unlikely to deliver materially lower household bills directly, though it may still have value through fiscal rent capture and some gas-related resilience at the margin. A renewables-led system may deliver larger long-run savings, but only through a broader and more uncertain package of electrification, retrofit delivery, market design, flexibility provision, grid buildout, and tariff reform. And those pathways are not necessarily mutually exclusive: a more serious policy analysis would ask how declining North Sea rents might help finance the very transition constraints the paper itself identifies.

That, in the end, is the real lesson of the review. The report is not useless. Nor is it a final answer. It is strongest as a warning against simplistic fossil-fuel bill politics, and weakest when it presents a complex, asymmetric policy comparison with more precision than the underlying framework can fully carry.

References / further reading

• Oxford Smith School, North Sea Rapid Analysis, March 2026. [https://www.smithschool.ox.ac.uk/sites/default/files/2026-03/North_Sea_Rapid_Analysis_March2026_OxfordSmithSchool.pdf]

• Orestis Delardas, Transforming marginal-cost pricing: A blockchain approach to fundamental energy system dynamics, Utilities Policy (2025). [https://doi.org/10.1016/j.jup.2025.101942]