Sizewell C — the last of its kind

The new finance deal for Sizewell C

RAB financing deal, developed from 2018, was announced in 2021 and legislated for in 2021-2022 when Kwasi Kwarteng was Secretary of State for Business, Energy and Industrial Strategy (BEIS) and completed under Ed Miliband at the Department for Energy Security and Net Zero (DESNZ) in July 2025. The Regulated Asset Base (RAB) finance model for nuclear power plants was sold to the public on the basis that it would provide cheaper power than using the fixed power price financial model under which the Hinkley Point C reactors¹ are being built.

It was claimed the model would bring in new sources of investment, particularly institutional investors such as pension funds. The power price reduction would be achieved if the public shared the economic risks with the investors and offered limited subsidies and guarantees. Reducing the risk borne by investors would reduce the cost of capital, a major element in the cost of power from a nuclear power plant, and hence the price of electricity. The subsidies were portrayed not so much as paying costs that would be expected to have been borne by investors, as is normal for subsidies, but as giving the investors guarantees they were not at risk from the consequences of low-probability, high-consequence events and from volatile wholesale electricity market prices.

After five years of effort by government to complete the deal, a Final Investment Decision (FID) for Sizewell C was finally taken on July 22, 2025. The contracts were finalised on November 4, 2025². The largest investor is the UK government (44.9%). The other investors are the Canadian pension fund, La Caisse (20%), Centrica (15%), EDF (12.5%) and Amber Infrastructure (7.6%). Amber Infrastructure is acting on behalf of the UK’s Nuclear Liabilities Fund, NLF, (4.6%), arguably public funds, and International Public Partnerships Limited 3.0%. So only 23% of the investment will come from institutional investors, 27.5% from energy companies and about half (if we include the NLF) from public sources.

An analysis of the Sizewell C deal shows that balance of risks is one-sided with the risks falling almost entirely on taxpayers and consumers, with minimal penalties and generous incentives offered to investors. The subsidies offered are far more extensive than those acknowledged by government and represent large amounts of public money being given to the private investors for no public return. The price of power from Sizewell C is unknown and will vary unpredictably from year to year, but there can be little confidence the RAB model will produce a lower power price than Hinkley Point C even if the cost of the subsidies is not factored in. The incentives required to bring in private investors are so expensive and risky to consumers that the model should not be repeated, and, like the Hinkley Point C deal, it ought to be a one-off, not a door-opener for new nuclear investments.

“The balance of risks is one-sided with the risks falling almost entirely on taxpayers and consumers, with minimal penalties and generous incentives offered to investors.”

The Risk/Reward balance

The plan to use the Hinkley Point C finance model for Sizewell C was abandoned by EDF in 2018. This was because it was not willing to accept the financial risks it had signed up to for Hinkley by agreeing a fixed power purchase price with all construction cost and time risks falling on itself. Costs have escalated dramatically at Hinkley since the deal was signed in 2016, by up to 90% but cannot be passed through to the power purchase price: and this commitment led to EDF writing off €12.9bn of its investment in Hinkley Point C in 2023³.

The investors in the Sizewell project frequently talk about the project being ‘derisked’⁴ by which they mean not that the risk has been reduced, but that it falls on others. The idea that changing the financial model and bringing in investors with no experience of nuclear power projects would reduce the risks is hardly credible. The key risks that had to be taken account of in the Sizewell deal are the risks of cost and time escalation in the construction period — the risks that are proving so costly for EDF at Hinkley. There are three key points that trigger rewards and penalties: the target completion cost and date; the Lower Regulatory Threshold (LRT); and the Higher Regulatory Threshold (HRT) (see Table 5).

The target completion cost is £38bn (in 2024 money) but, unaccountably, the interested parties are not willing to reveal the target completion date citing commercial confidentiality, merely saying mid-late 2030s. Let us assume for illustrative purposes that the date is 2037. The LRT is £40.5bn (2024) with completion at the end of 2039, an overrun of 6.5%. Centrica describes this as a ‘moderate’ outcome and government as a ‘central’ assumption. The HRT is £47.7bn (2024), an overrun of 18% compared to the LRT ‘central estimate’ and completion in November 2043, about six years late. Centrica describes this outcome as ‘severe’ and ‘very unlikely’. The government characterises HRT as ‘remote’. We follow the government and use the LRT as the central estimate for our analysis.

If Sizewell C was completed at a cost of £47.7bn it would make it the most successful EPR project todate.

The government and Centrica’s characterisations of HRT are not credible. The two complete or nearly complete EPR projects in Europe⁵ are 13-14 years late and 3-4 times overbudget. Even the two reactors completed in China (Taishan), often portrayed as success stories we should emulate⁶ were eight years late with costs reportedly about 60% overbudget (see Table 1). By January 2024, the date of the most recent cost estimate for Hinkley Point C, the project was up to seven years late and up to 90% overbudget. It is barely credible there will not be further delays and cost increases at Hinkley in the seven years from 2025 to the expected completion date of 2032⁷. The publicity supporting the Sizewell deal continually cites the impact of ‘learning’ from the Hinkley project as the factor that will ensure its success. But it appears successive EPR projects in Europe have gone worse in terms of time and cost, so if there is learning, it has not improved performance. If the Sizewell project came in at no more than the HRT, it would make it — by a very large margin — the most successful EPR project to date (see Table 2). So, the HRT is a remote outcome, not in the sense it is highly unlikely but in the sense that it is highly unlikely it will not be exceeded.

For the Hinkley deal there were two date markers. The plant was forecast to be complete by November 2025 but if it was not complete by November 2029 (4 years late), the power purchase contract would be reduced by a year for every year of delay beyond 2029 — EDF’s January 2024 cost update⁸ showed this was certain to happen. If the plant was not complete by the ‘long-stop’ date of November 2033 (eight years late), the government had the option to cancel the power purchase contract completely. If the contract is cancelled and not replaced, EDF would be forced to sell the output of Hinkley Point C to the market at a price highly likely to be well below the £92.5/MWh (2015) contract price. In 2025, the market price was about £83.5/MWh compared to the contract price in 2025 money of £133/MWh. In 2021, EDF, citing force majeure, persuaded the government to put the long-stop date back to November 2036. Given the delays of six years announced since then, this date no longer appears the remote risk it seemed to be in 2021 when completion was expected in 2026.

The balance of risk is reversed for Sizewell C. There is no equivalent to the 2029 date at which the contract length will be reduced if the plant is not complete. If plant construction goes beyond the HRT of 2043, there is no option to cancel the power purchase contract and if the government does choose to abandon the project, it would have to compensate investors despite it being the investors that failed to control cost and schedule.

The incentives and penalties for the operating phase appear more evenly set but are not clearly specified. If the targets at which incentives are triggered are set at a low level, it would be highly likely they would be exceeded with the corresponding benefits accruing to investors at the expense of consumers.

Table 1: Comparisons of European Pressurised Reactor completion delays and cost overruns

Table 2: Comparative figures for Sizewell C

Subsidies

The government only acknowledged two subsidy elements in the entry in its Subsidies Database for the Sizewell C deal⁹ the value of the Government Support Package (GSP) and the cost of the ‘Difference’ payments¹⁰ (see Table 3). The value of the subsidy was estimated at £54.589bn payable between now and July 22, 2100. The first allocation was made on the day the contracts were finalised, November 4, 2025¹¹. This was for £50.284bn, also payable up to July 22, 2100 but with the subsidy now split into five elements:

1. Regulated Asset Base contract between the Low Carbon Contracts Company (LCCC) and Sizewell C.
2. A government commitment to provide an equity investment to the project.
3. Together with the National Wealth Fund, government has committed to provide debt financing to the project.
4. A Government Support Package (GSP) that comprises a number of products which will insure the project against high impact low probability events for which the project cannot gain insurance on the commercial market (specifically, a Supplemental Compensation Agreement, Discontinuation and Compensation Agreement, and Government Liquidity Facility Agreement, on which further detail can be found below).
5. An offtake agreement for specific equity investors in the Sizewell C project.

Numbers 1 and 4 appear to be the two elements already identified. It is not clear why the second element was not identified in the July 22 entry as the government stake was already known then. The third element appears to be an acknowledgement that the loans from the National Wealth Fund do constitute a subsidy. The fifth element appears to relate to an agreement made with Centrica to take additional volumes of the power¹² above what it will be required to take as an electricity retailer.¹³ Unless this apparently additional sale to Centrica is sold at a lower price than the price for the other retailers, the subsidy will be the same whether all the power is sold to electricity retailers or a small additional element of it is added to Centrica’s purchases. The uses of ‘investors’ suggests it is not only Centrica who have an offtake agreement but, logically, of the other investors, only EDF would be able to use this power.

Without knowing the split between the five elements of the subsidy identified, the discount rate, and the timing of the payments it is impossible to calculate what the impact on consumers is, when it will occur and what the undiscounted value is — i.e. what the public will actually be forced to pay. The value of the government equity investment and the value of the loans from the National Wealth Fund are hard to estimate. The value of the Government Support Package (GSP) is an estimate based on the assumption that it is possible to calculate the cost of obtaining insurance cover against the so-called low-probability high-consequence events covered despite the fact the Government Support Package is only necessary because such cover would be impossible to obtain. So, the actual cost to taxpayers of providing this insurance will be zero if none of the events occur or much more than the assumed insurance premium cost if one does.

The acknowledgement that ‘Difference’ payments will be substantial demonstrates that it is expected that consumers will be forced to buy power from Sizewell C that will cost more than alternatives in the market. The ‘Difference’ payment is the amount of money added to consumers’ bills to make up the difference between the amount earned by Sizewell C selling power to the market at market price¹⁴ and the financial return that must be paid to investors. The value estimated for the ‘Difference’ payments assumes that the market price for electricity can be forecast with a worthwhile degree of accuracy up to 75 years in the future — 15 years from now to commercial operation and a power purchase contract for 60 years. Given that there is little confidence in wholesale price forecasts for even five years forward, the estimate of the value of difference payments has no credibility. The guaranteed income will be calculated on the actual cost of construction (the value of the Regulated Asset Base), the cost of operations & maintenance and the rate of return set by the regulator, Ofgem.

Other than demonstrating that the estimated cost of these subsidies is huge, the cost estimates of £54.589bn and £50.284bn payable between now and July 22, 2100 have no analytical value because they are discounted¹⁵. What if anything the unallocated £4.305bn will be used for is unknown. The surreal precision of the amounts of money and the dates that have been set suggests that while the Treasury may use these numbers to fill in its national accounts, they are of no actual value.

“The acknowledgement that ‘Difference’ payments will be substantial demonstrates that it is expected that consumers will be forced to buy power from Sizewell C that will cost more than alternatives in the market.”

The main actual subsidy, not acknowledged by government, is the surcharge on electricity bills from date of Final Investment Decision (FID) to three years after Commercial Operation Date (COD). This pays for three separate subsidies:

• The return paid to investors for their equity contribution (equity is expected to pay for 35% of the construction cost) during the construction period.
• The part of the surcharge used to pay for about 60% of the investors’ equity contribution¹⁶.
• The cost of borrowing to the investors (borrowing is expected to pay for 65% of the construction cost).

The Low Carbon Contracts Company (LCCC) has said the surcharge, the Interim Levy Rate, for the period December 1, 2025 to December 31, 2025 would be £3.494/MWh and from January 1 2026 to March 31, 2026 would be £3.663/MWh.¹⁷

Investors will be paid a real rate of return of 10.8% on their equity contributions. Of this 6% can be paid to the investors’ shareholders. The other 4.8% is earmarked to pay for more than half of the investors’ equity contribution.¹⁸ So consumers will be giving investors more than half of their equity needs and the investors will be earning 6% real on this gift. The government has said it will ‘recycle any profits’ from the surcharge to electricity consumers¹⁹, but it has acknowledged it has not determined how this will be done and what constitutes ‘profits’. The government has not said whether it expects to borrow its debt contribution from the National Wealth Fund or issue gilts. It has said “the actual funding from government will be borrowed at the lower government gilt rate plus an administrative margin”²⁰. It has given no indication what the gilt rate or the administrative margin might be. However, given that borrowing only accounts for 65% of the cost and that the government’s stake is only 44.9%, ‘recycling’ some of the debt costs can only have a small impact on the consumer surcharge. If the 6% return to investors is recycled, this will also be a taxpayer subsidy investing taxpayer money for no return. The 4.8% element, termed ‘RAB Growth’ by Centrica that pays for more than half the investors’ equity contribution, might be seen as profit but it is a small element of the consumer surcharge.

The loans, up to £36.6bn (enough to provide 65% of the cost up to £56.3bn), will be provided by the UK government’s National Wealth Fund at a low rate of interest, 4.5% real. There is also a £5bn element backed by France’s export credit guarantee entity, BpifranceAC. Sizewell C talks about ‘£5 billion of debt raised through the BpifranceAE export credit facility as well as a £500 million Working Capital Facility, both sitting alongside the National Wealth Fund’s term loan. 13 banks have supported the £5 billion debt raise and a subset of these are providing the Working Capital Facility,’²¹ It is not clear how this £5bn fits in with the National Wealth Fund contribution.

It is not known whether commercial sources would have been willing to provide the loans and if they would, what the interest rate charged would have been and what government guarantees would have been required. Given that government capital resources are limited, there will be a substantial opportunity cost of other projects that cannot be financed if the money is lent to Sizewell C — the National Wealth Fund’s entire budget excluding Sizewell C is only £27.8bn.

It must therefore be assumed either that commercial finance would not have been available, or if it was, the cost would have been significantly/prohibitively higher. So, the National Wealth Fund loans represent a substantial public subsidy. If the government is going to set the interest rate, it could choose to impose the same rate whatever financial model was used. So, it cannot argue the cost of borrowing was reduced, just that this large cost will be borne by consumers, not investors. A central plank to the case for the RAB model, reduced finance charges, therefore, is invalid. If the interest rate increases, consumers, not the investors, will be liable to pay the additional cost through the surcharge.

“Consumers will be giving investors more than half of their equity needs and the investors will be earning 6% real on this gift.”

Table 3: Subsidies

Will the power be cheaper than for Hinkley?

The Hinkley Point C deal was a poor one for consumers²² with a wholesale power price of £92.5/MWh (2012 prices), £133.0/MWh in 2025 prices, against an average wholesale price for October 2024 to September 2025 of £83.50 (see Graph page 8).²³ Whether the array of public subsidies, guarantees and incentives offered under the Sizewell C deal will be sufficient to make the power purchase price less, on average, than the firm figure set for Hinkley Point C can only be judged, at least provisionally, when the plant is operating and its construction cost known (see Table 4). Even then, the power price will vary unpredictably from year to year. For Hinkley Point C, the power purchase contract was for 35 years but for Sizewell C, it will be 60 years, so, if as the government has acknowledged, the power price is not competitive with alternative options, consumers will be locked into the deal till around 2100 or later. Whether the plant will be able to operate for 60 years is unclear. No nuclear power plant has operated for even close to that long.

Table 4: Comparison between Hinkley Point C and Sizewell C

Is the Sizewell deal repeatable?

The Hinkley Point C model was not repeatable because investors were not willing to take the risk it involved of fixing the power purchase price in advance. Ironically, the model appears to work well for offshore windfarms because investors are prepared to accept these risks for offshore wind projects. The logical conclusion is therefore that it is nuclear technology that is the problem, not the finance model. For the Sizewell C model these risks fall on consumers. For the sake of consumer pockets, it must be hoped that when the extent of these risks and subsidies needed to get private funding for only half the cost of the plant is apparent, it will be politically impossible to repeat the deal even if investors can be found for follow-on projects.

If the deal proves not to be repeatable, the huge amount of government time and cost that has gone into completing the deal will, as with Hinkley Point C, have been a costly diversion of more than a decade from pursuing the cheaper, quicker and more reliable ways of meeting the government’s promises of net zero.

Information correct as of November 2025. (Source: Wholesale market indicators, Ofgem.)

Is Sizewell the last large reactor for the UK?

The Johnson government set a target of up to 24GW, the equivalent of eight Sizewell Cs of new nuclear capacity to come online in UK by 2050 with Sizewell C to be the first. The inclusion of ‘up to’ means the target was effectively meaningless. Nevertheless, the Starmer government has been careful not to commit to this target. There are no identified projects to follow-on from Sizewell C: there is no procedure to choose technology; no sites have been selected, and it seems likely a new finance model will be needed. Experience suggests that even if creating these processes were to start up tomorrow and go smoothly, a follow-on plant would not come online before 2045 but probably after 2050, perhaps two decades after fossil fuels had been largely phased out of the electricity generation mix. This is hardly an effective option for what is widely seen as a climate crisis.

In terms of technology, the EPR was effectively abandoned a decade or more ago by France in favour of a revised design, EPR2. But the first reactor using this design will not come online (in France) before 2038.²⁴ So, unless the UK takes the reckless step of ordering it before its costs and performance have been demonstrated, it is not an option. There is much talk of the attractions of Korean APR1400 design²⁵. There are question marks about its cost and performance but the intellectual property for the design is held by Westinghouse who will not allow it to be built in Western Europe, so it is not an option. This leaves only one viable technology choice (assuming we discount Russian and Chinese designs), the Westinghouse AP1000, albeit its record of construction time and cost are not much better than that of the EPR.

It may be that the flurry of publicity about Small Modular Reactor projects is a tacit acknowledgement that large reactor designs have run their course and if there is any future for nuclear power, it is through Small Modular Reactors.

Table 5: Incentives and Penalties

Endnotes

^1. Both projects comprise two 1600MW reactors using the French European Pressurised Reactor (EPR) design.

^2. https://www.sizewellc.com/news-views/sizewell-c-reaches-financial-close-with-5-billion-export-credit-backed-debt-raise-and-strong-investment-grade-credit-rating/

^3. https://www.edf.fr/sites/groupe/files/2024-04/annual-results-edf-2023-presentation-2024-04-02.pdf

^4. https://www.centrica.com/investors/centrica-as-an-investment/investing-for-value-at-sizewell-c/

^5. Olkiluoto 3 (Finland) and Flamanville 3 (France). Flamanville 3 produced first power in December 2024 but has experienced significant problems and commercial operation will not be before 2026^.

^6. https://committees.parliament.uk/publications/46805/documents/242664/default/

^7. https://www.bloomberg.com/news/articles/2025-11-07/edf-braces-for-more-delays-at-uk-hinkley-point-nuclear-project

^8. https://www.edf.fr/en/the-edf-group/dedicated-sections/journalists/all-press-releases/hinkley-point-c-update-1

^9. https://searchforuksubsidies.beis.gov.uk/scheme/?scheme=SC11357

^10. The difference payment is the difference paid by consumers between the market price and the price necessary to generate the investor’s guaranteed income.

^11.. https://searchforuksubsidies.beis.gov.uk/scheme/?scheme=SC11357

^12. ‘Agreement in principle for initial 20-year offtake agreement for our share of Sizewell C’s production, and for Centrica to provide Sizewell C with route to market services for additional volumes.’

https://www.centrica.com/media-centre/news/2025/investment-in-sizewell-c

^13. The power is expected to be sold by the government’s Low Carbon Contracts Company to all electricity retailers in proportion to their share of the retail market.

^14. The market price used will be the average forward price sourced from wholesale market transaction data collected from brokers.

https://www.ofgem.gov.uk/news-and-insight/data/data-portal/wholesale-market-indicators

^15. Discounting reduces the accounting cost of future liabilities by assuming that, if there was a fund to pay the cost, it would grow in real terms by investing it. Over long periods, the impact is dramatic. For example, if a fund earned 3% real for 80 years, it would grow more than 10-fold.

^16. The government has said it will ‘recycle’ to consumers the payment its stake generates under this provision.

^17. https://www.lowcarboncontracts.uk/news/nuclear-regulated-asset-based-bulletin/

^18. Centrica has said that its expected £3bn contribution will come partly from itself, £1.3bn, but mostly from the consumers surcharge (it terms this ‘RAB growth’). https://www.centrica.com/investors/centrica-as-an-investment/investing-for-value-at-sizewell-c/

^19. Email from Ofgem to Policy Brief, November 7, 2025.

^20. Sizewell C (SZC) Final Investment Decision: Value for Money (VfM) assessment, July 2025, DESNZ, page 10. https://assets.publishing.service.gov.uk/media/687f951dfdc190fb6b84689d/szc-fid-vfm-assessment.pdf

^21. https://www.sizewellc.com/news-views/sizewell-c-reaches-financial-close-with-5-billion-export-credit-backed-debt-raise-and-strong-investment-grade-credit-rating/

^22. The NAO concluded ‘[The deal] for Hinkley Point C has locked consumers into a risky and expensive project with uncertain strategic and economic benefits.’ https://www.nao.org.uk/reports/hinkley-point-c/

^23. Wholesale market indicators, Ofgem, accessed November 2025.

^24. Cour des Comptes (2025) ‘La filière EPR : une dynamique nouvelle, des risques persistants’

https://www.ccomptes.fr/fr/publications/la-filiere-epr-une-dynamique-nouvelle-des-risques-persistants Quotations from the report in English are based on a machine translation of it.

^25. https://committees.parliament.uk/publications/46805/documents/242664/default/