1. Introduction
If we strip away the rhetoric of a “golden age for nuclear power” and “ripping up rules to fire up nuclear power”, Britain’s nuclear programme looks very tired, with just three projects that have a reasonable chance of completion. Two of these (Hinkley Point C and Sizewell C) had their origins nearly 20 years ago while the third (Wylfa1) was launched under Boris Johnson’s government. Increasingly, what these three projects have most in common is not the reactors, but who is paying for them. The government’s National Wealth Fund (NWF) – set up to catalyse private investment across the economy – has already had to commit up to £36.6bn of the finance for Sizewell C, on top of a loan to design Wylfa’s reactors. If the Advanced Nuclear Framework succeeds in bringing forward the Cottam, Hartlepool and Oldbury projects too, the NWF could find itself committing a further £60bn to nuclear – in total 3.5 times what it has budgeted for the other 24 sectors it exists to support.
A fund created to unlock private investment across the economy risks becoming a vehicle for channelling public money into the most expensive, slowest and least certain to be delivered of all the forms of power generation available. An analysis of costs and benefits suggests that, far from rushing to launch new nuclear projects as the nuclear industry would advocate, the British people would be better served by abandoning these projects. The money and resources released – including by freeing the National Wealth Fund to do the job it was set up to do – can be redeployed to options that will reduce electricity bills, save taxpayers money and improve our prospects of meeting our climate change goals quickly and reliably.
2. Our nuclear capacity
The UK currently has only five operating nuclear power plants: the four remaining Advanced Gas-cooled Reactor stations (4.7GW) which are likely to be retired by 2031/32,2 and Sizewell B (1.2GW), which is expected to operate until 2055. Hinkley Point C (3.26GW) is under construction and will not be complete before 2032.3 Sizewell C (3.26GW) is three years away from an above ground construction start and, if all goes to plan, might be on-line by the start of 2040.4 The Wylfa Small Modular Reactor (SMR) station (1.5GW) is three years from a firm order being placed so is also unlikely to be generating much before 2040.5
Beyond the Hinkley and Sizewell projects, there is no evidence that further large reactor projects will be pursued (the prime site for large reactors is Wylfa and this has been allocated to an SMR project). There are also no specific government-led proposals for projects to follow on from Wylfa. It appears the government is relying on private sector projects to provide additional nuclear capacity and in February 2026 it published its Advanced Nuclear Framework, which is meant to provide government support for private sector-led projects.6 Under this, several forms of support were offered including:
- Finance via the government’s National Wealth Fund;
- Support via Great British Energy – Nuclear in finding sites;
- Help negotiating power purchase agreements such as contracts for difference that would remove any exposure of the plant to the competitive electricity wholesale market;
- Assessment of the projects by Great British Energy – Nuclear.
Two private sector SMR projects – Cottam and Hartlepool, both of about 1GW capacity – had been announced in September 2025 in association with President Trump’s visit to the UK.7 In June 2026, Holtec announced its application to the Advanced Nuclear Framework for the Cottam project.8 By the end of August 2026, X-Energy, the proposed reactor vendor for the Hartlepool site, had not made an application for entry into the Advanced Nuclear Framework.
Beyond Hinkley and Sizewell, there is no evidence that further large reactor projects will be pursued. There are also no specific government-led proposals for projects to follow on from Wylfa.
In July 2026 GE Vernova, via its Polish Partner SGE (Synthos Green Energy), announced an application to the Advanced Nuclear Framework to build up to 14 BWRX-300 SMRs (4.2GW).9 The application involves a proposal to site the first six reactors at the Oldbury site located on the River Severn, currently owned by Great British Energy – Nuclear (GBE-N). This is a problematic site and the Chair of GBE-N, Simon Bowen, told the ESNZ Committee: “If you look at Oldbury, particularly, the site that is close to you, we are aware that the only way to develop the Oldbury site for nuclear development will be with a significant investment in flood defences.”10 SGE’s target to complete the first reactor by 2034 is unrealistic given that it is several years behind the Rolls-Royce Wylfa project, where the first reactor is not expected to be complete before the ‘mid-30s’.11 It lags behind Wylfa in terms of siting, finance, and regulatory approval by the Office for Nuclear Regulation. GBE-N has stated it has not been contacted by SGE about the Oldbury site.12 There is no evidence yet as to whether the Advanced Nuclear Framework will be an effective route for promoting nuclear projects.
So, if there are no further delays with Hinkley Point C, Sizewell C and Wylfa, it is likely that it will be around 2040 before nuclear capacity exceeds today’s levels and even then, will only have got back to 2019 capacity levels. The current government’s target was to essentially remove all fossil-fuel generation by 2030 so new nuclear capacity can play no role in achieving that. If the AGRs are retired in 2031/32 probably on safety grounds (due to graphite moderator issues13) the British electricity system will have to survive with only about 4.5GW of nuclear capacity and no fossil fuel generation for around a decade. So, if there is a need for new base-load capacity, nuclear cannot provide it before 2040, and even then nuclear capacity will be much less than base-load (defined as the lowest level of demand in the year).
The government’s claim that “Unlike traditional large-scale nuclear plants, these projects [SMRs] usually need smaller sites, allowing more flexibility in where they can be built” is belied by the fact that the Wylfa project (1.5GW) and the potential Oldbury project (1.8GW) are substantially larger than any completed nuclear power plant in the UK and the Hartlepool project is about the same size as the AGR stations and Sizewell B. The reality is that sites that are suitable for nuclear projects in terms of physical factors and public acceptance are, and always will be, hard to find and there will be an overwhelming incentive to fit as many SMRs on to an accepted site as possible.
3. Wylfa
In June 2025, the government announced the winner of its competition to select a Small Modular Reactor (SMR) design for deployment in the UK.14 The design was the 470MW Rolls-Royce PWR SMR for which three reactors are expected to be built at the Wylfa site in North Wales (1.5GW). In June 2025, the government announced that it had allocated £2.6bn to the Wylfa project up to the end of fiscal year 2029/3015. A Final Investment Decision (FID) is not expected until 2029 with the process handled by the government-owned Great British Energy – Nuclear (GBE-N) entity. A significant number of hurdles must be overcome before the FID can be taken. These include completion of the Generic Design Assessment by the Office for Nuclear Regulation (ONR). This was expected in July 2026, but unofficial reports suggest the decision will not be before December 2027. The delay is said to be due to the need for design changes. The requirements the nuclear regulator imposes will be an important determinant of the reactor cost. The Wylfa site is widely portrayed as being the best of the existing reactor sites for new development and it did undergo some investigation when it was proposed as a site for two large reactors a decade ago. Nevertheless, the site investigations must be rigorous given the site is unlikely to be cleared after reactor operation until perhaps 150 years from now. However, site work done so far will be minimal, much of it funded by GBE-N so remediation costs would be small.
None of the important financial details about the Wylfa project have been announced. These include ownership, finance method, source of finance and power sales arrangements. However, experience with Hinkley Point C and Sizewell C gives strong indications of what would be needed to finalise the project. They include a substantial government equity stake, possibly as high as 100%; a power sales agreement that ensures all investor costs are passed on to consumers/taxpayers; and cheap finance provided from the National Wealth Fund.
Experience with Sizewell C shows that private investors are not willing to be exposed to any significant part of this risk and if a deal is to be done the risk will fall on the public, as taxpayers and electricity consumers.
Private investors will be wary of this project given that, in addition to the risks all nuclear projects suffer, the design is completely untested and Rolls-Royce has no experience as a vendor of civil nuclear power plants. These risks exist and will not be changed by different methods of funding and finance. All that can be changed is who bears these risks. Experience with Sizewell C shows that private investors are not willing to be exposed to any significant part of this risk and if a deal is to be done the risk will fall on the public, as taxpayers and electricity consumers. A deal will be easiest to reach if government owns all or practically all the equity so that the risk will fall on taxpayer and electricity consumers. This will avoid the lengthy delays that occurred getting the Sizewell C project to an FID from having to negotiate, not very successfully, with potential investors.
Rolls-Royce has not made predictions of the construction cost since the design was little more than a conceptual one. In 2021, it claimed the cost would fall to £1.8bn for the ‘nth of a kind’ (NOAK) with the first expected to be generating by 2030. The NOAK costs would apply when the supply chain is well established, and learning has made the construction process well established and repeatable. In practice, how many reactor orders it will take to get to ‘n’ is not well-established and most reactor designs are superseded by new or modified designs well before they reach any credible ‘n’.16 So, these forecasts have no value in forecasting the actual cost of a Rolls-Royce SMR.
By June 2026, GBE–N had let contracts worth nearly £900m for the Wylfa project so the compensation payable as a result of abandoning these contracts would be significantly less than £1bn.17 The government’s National Wealth Fund agreed to lend it £599m to develop the design in April 2026 demonstrating that the design is still far from complete.18
4. Sizewell C
The Sizewell C plant is nearly a decade behind Hinkley Point C and the costs incurred so far are correspondingly smaller. The Final Investment Decision (FID) was taken in July 2025 with the government’s central estimate for completion at the end of 2039 at a cost of £40.5bn (2024 money).19 In the June 2025 Spending Review, ahead of the FID, then-Chancellor Rachel Reeves allocated £14.2bn of public money to be spent on the project by the end of fiscal year 2029/30.20 The government claims the “(p)roject will be built for around 20% less than virtual replica Hinkley Point C”, claiming the project “builds on lessons learnt from the construction of Hinkley Point C.” It is misleading to compare in this way the estimated cost at Final Investment Decision (FID) of a project that has not yet had a chance to go wrong – Sizewell C – with one that has been under construction for eight years – Hinkley Point C – for which five delays and cost increases have been announced. The appropriate comparison is with the forecast costs at the point of FID. For Hinkley Point C it was £24bn (2024 money) compared to £40.5bn (2024) for Sizewell C, a real increase of nearly 70%. The time from Final Investment Decision (FID) to completion for Hinkley Point C was forecast to be nine years whereas for Sizewell C it is forecast to be 14 years. So, the learning from Hinkley Point C has led to large real estimated cost increases and a much longer expected lead-time.
The government was only able to complete the deal by offering huge public subsidies and, even then, had to take nearly half the equity. In the government’s Subsidies Database, the government estimates the value of these subsidies as £54.6bn.
The government was only able to complete the deal by offering huge public subsidies and, even then, had to take nearly half the equity. In the government’s Subsidies Database, the government estimates the value of these subsidies as £54.6bn.21 This is a ‘discounted’ figure so the actual amount paid by taxpayers and electricity consumers will be substantially more.22 During the construction period, consumers will pay a surcharge on their electricity bills. This will pay all the investors’ finance charges (approximately £25bn) and will also provide a ‘gift’ to investors of 60% of the equity the investors will be required to put into the project23 worth about £8bn. Finance will mainly be provided by the government’s National Wealth Fund with a budget of up to £36.6bn, at an advantageous rate initially set at a real rate of only 4.5%.
4.1. What will the price of power from Sizewell C be?
The key element in determining the cost-benefit balance for abandoning Sizewell C is the price consumers must pay for its output. One of the few good features of the Hinkley Point C deal is that consumers know what price they will be charged for power, albeit a very high one, regardless of what costs are. As EDF states in its announcements of cost increases: “Under the terms of the Contract for Difference, there is no impact for UK consumers [of cost increases].”24 The contract was for 35 years whereas for Sizewell C it is for 60 years, so consumers will be locked into the Sizewell deal for 25 years longer. For Sizewell, the price of power will be whatever it takes to generate the revenue the investors are contractually entitled to – consumers have had a blank cheque signed on their behalf. Every project using the EPR technology chosen for Sizewell C has far exceeded the cost and time estimates made when the investment decision was taken. So, the power price forecasts for Sizewell C based on estimates made at the time of FID must be regarded as highly likely to be a gross underestimate of the outturn power price. If construction or operating costs are higher than forecast, they will be passed on to consumers. Sizewell C Ltd has made some highly optimistic predictions about the price of power from Sizewell C claiming it would be £88-102/MWh (2026 money)25 – well below Hinkley. But this comes from the project’s promoter, rests on its best-case assumptions, including no cost overrun at all, and carries no consequences for the company if it proves wrong. Given the record of every EPR built so far, it should be treated as a floor rather than a forecast.
Unlike Hinkley Point C, consumers will be required to pay all the finance costs incurred in the construction period as they arise and they will also pay more than half of the investors’ equity contributions. The National Wealth Fund is providing most of the finance at subsidised rate representing a taxpayer subsidy. These additional costs must be factored into any projection of the cost of power from Sizewell C. So, all things being equal, the price to consumers of power from Sizewell C ought to be significantly less than for Hinkley Point C which had none of these subsidies. However, all things are not equal: the forecast construction cost of Sizewell is 70% higher in real terms than the cost of Hinkley on which the power purchase price of £89.5/MWh (2012 prices) was based.
If the outcome cost and time did reach the HRT, it would mean that the construction cost, which would be passed on to consumers as part of the Regulated Asset Base (RAB), would be double the forecast level for Hinkley Point C, the level on which its power price was calculated.
The Sizewell deal includes a Higher Regulatory Threshold (HRT) level with construction costs £47.7bn and completion in 2043 compared to the central assumption, the Lower Regulatory Threshold (LRT), which assumes a construction cost of £40.5bn and completion at the end of 2039. Sizewell C Ltd’s baseline estimate is completion in July 2039 at a cost of £38.2bn. Unlike the LRT and HRT, this estimate has no contractual force and there will be no financial consequences when it is proved wrong. While the HRT is described by government as having a ‘remote’ probability, all experience with nuclear projects, especially those involving EPR technology, suggests that an outcome of costs less than 20% over budget and a three-year delay would be an unprecedentedly good result. If the outcome cost and time did reach the HRT, it would mean that the construction cost, which would be passed on to consumers as part of the Regulated Asset Base (RAB), would be double the forecast level for Hinkley Point C, the level on which its power price was calculated. At that point, the contractual requirement on investors to continue to invest would end with no penalty to them – they keep the full return on what they have already invested – and government would either have to take over funding or abandon the project. By contrast, if Hinkley Point C is not completed by 203626, the government has the option to tear up the power purchase agreement leaving EDF to try to negotiate a new deal or abandon the project and lose the money it invested in the project.
4.1.1. The NAO report
The NAO report on the Sizewell C project publishes a large amount of data on the project including the forecast range of costs of Sizewell over its entire life in real terms.27 Its lower figures correspond to the LRT values and the higher to the HRT. Its data show the flow of cash during the construction period including the ‘net consumer payments’ figure of £24.7-34.5bn. We assume this includes the consumer contribution to investors’ equity costs. NAO’s figures show the cash flow during operation and the total consumer contribution in that period is £111.8-124.0bn making a total consumer contribution of £136.5-158.5bn.
The NAO analysis reports that the price of power is expected to be £133–155/MWh (2024/25 prices), or £152–178/MWh without the ‘recycling’ of the government’s profit as the 44.9% shareholder. Recycling passes that profit to consumers’ bills instead of to the Exchequer. It is a large subsidy from taxpayers to electricity consumers, and it means the lower price understates the full cost of Sizewell C’s power to the public. It is also a very bad precedent to invest so much taxpayers’ money for no return. The NAO notes: “Recycling is a policy decision, which the government could stop or modify at a later point, changing the cost to consumers.”28
As the NAO explains, this is a levelised figure: the discounted value of consumer payments, including those during construction, divided by the discounted value of output, using the Treasury’s standard discount rate. Because consumers pay years before the power arrives, it is higher than the undiscounted average, about £91–106/MWh on the NAO’s own totals – but only if costs stay within the HRT. Given the record of every EPR to date, we regard that as unrealistic.
The NAO identifies some of the assumptions that lie behind the power price forecast, including:
- Sizewell C’s cost will not exceed the HRT;
- Sizewell C will have a load factor (power produced divided by maximum possible power) of 90%, well above the average for large reactors;
- Electricity demand will increase by 84-108% between 2024 and 2050. By comparison, electricity demand fell by about 25% in the period 2005-2025.
Even with these questionable assumptions, the project is not expected to break even until 2064 and there is ample scope for the price to be significantly higher than the upper end of the range but negligible scope for it to be lower than the lower point. The NAO acknowledges that “although Sizewell C should cost less than Hinkley Point C to build, consumers may still pay more for Sizewell C’s electricity”. It could have added that unless the Sizewell project progresses significantly better than Hinkley Point C, the difference could be large. NAO also says, “To match HPC’s fixed in real terms £129 per MWh in 2025 prices SZC would need to come in below the project baseline.”29 The baseline is Sizewell C Ltd’s own risk-adjusted forecast. So, merely to match Hinkley Point C’s price, Sizewell C would have to outperform its own company’s target – an outcome the record of every EPR built so far makes highly unlikely.
Given the improbability of the Sizewell C project being completed on time and to budget even under the LRT assumptions, the price of power from Sizewell C will inevitably exceed the price of power from Hinkley Point C and probably by a large amount. For our estimates of the cost of power from Sizewell C and the savings from abandoning it, we take as our lower estimate the price of power from Hinkley Point C, £129/MWh, recognising that achieving such a low cost is highly improbable. For our upper estimate, we assume a price 20% higher, £155/MWh, to match the difference between the LRT and HRT. We also include two cases with the recycling option withdrawn, which we assume, based on the NAO’s figures, increases the power price by 15% to £148-178/MWh.
The situation if costs exceed the HRT is complicated. The NAO report states30: “If the Company requires additional funding to cover these costs, it must apply any available cash and outstanding shareholder commitments and seek private funding first, before applying for taxpayer funding under the government support package. The Secretary of State can also terminate the project instead of providing taxpayer support.”
Given the improbability of the Sizewell C project being completed on time and to budget even under the LRT assumptions, the price of power from Sizewell C will inevitably exceed the price of power from Hinkley Point C and probably by a large amount.
Centrica’s analysis has a rather different emphasis31: “Above the HRT, in the event further funding is unavailable from the private market, HM Government has committed either to provide the required funding or to discontinue the project and provide compensation to investors in respect of the RAB as at the discontinuation date. Returns continue to accrue during construction in the event of delay.” This appears more realistic. The private market is unlikely to fund a failing project. Centrica states: “that its £1.3bn [investment] is capped” and, if HRT is exceeded: “If market financing not available, Government to fund with no impact on Centrica’s RAB”.32 We assume that if the HRT is exceeded, the government of the day is unlikely to see abandonment as politically feasible. By then, the project would have been under way for more than a decade, and cancelling it would mean paying private investors about £26bn – their 55% share of a RAB of around £47bn. Cancelling now would cost under £5bn. The deal gives investors little reason to control costs. The NAO found that, as construction costs rise below the HRT, investors receive more cash overall, if later and at a lower rate of return, and that even at the HRT their returns remain “in line with comparator investments”. Above it, investors need put in no more money. The Secretary of State must then choose between adding the extra costs to consumers’ bills, funding them from the Exchequer, or cancelling the project and compensating investors. Whichever is chosen, the public pays. As the NAO puts it, “it is not clear that the deal sufficiently incentivises investors to keep construction costs well below the higher regulatory threshold.” For investors it is heads they win; for the public, tails it loses.
We assume the wholesale electricity price, the price that will determine how much extra consumers will have to pay because they are forced to buy the output of Sizewell C, matches the highest successful bids of about £90/MWh to the government’s Allocation Round 7, announced in February 2026.33 We also assume Sizewell C produces 1500TWh over the 60-year contract length using the assumption of a 90% lifetime load factor.
All savings figures in this piece are in today’s prices but undiscounted. They multiply the extra cost per MWh by each plant’s lifetime output, rather than discounting future payments as official appraisals do. Discounting gives less weight to costs falling on future bill payers, and these contracts run to the end of the century, so we regard the full sum as the fairer measure of what the public would be committed to.
Table 1: Power price savings from abandoning Sizewell C
| Power price £/MWh | Saving (£bn) | |
| Lower price case with recycling | 129 | 58 |
| Lower price case without recycling | 148 | 87 |
| Higher price case with recycling | 155 | 98 |
| Higher price case without recycling | 178 | 132 |
The savings in consumer power price from abandoning Sizewell C, assuming recycling continues, range from £58bn in the highly optimistic lower price case to £98bn in the still optimistic higher case. If the taxpayer subsidy that recycling represents is withdrawn, the savings go up to £87-132bn. All the history of large reactor construction this century, especially using EPR technology, suggests there is considerable scope for the project to cost significantly more than the HRT making the potential savings even larger.
The savings in consumer power price from abandoning Sizewell C, assuming recycling continues, range from £58bn in the highly optimistic lower price case to £98bn in the still optimistic higher case. If the taxpayer subsidy that recycling represents is withdrawn, the savings go up to £87-132bn.
4.2. What would the cost of abandonment be?
The cost of remediating the Sizewell C site would be relatively small – despite extensive environmental damage already incurred, some not remediable – given that construction start (pouring of first structural concrete) is not expected until 2029.34 So, the main cost would be compensating the investors. The Sizewell C annual report shows that by the end of fiscal year 2025/26, Sizewell C had spent a total of £8.7bn. However, the government stake is 44.9% so only £4.8bn would have to be repaid to private investors.
The savings from abandonment would be the extra cost of Sizewell C’s power: on the cases in Table 1, £58–98bn with recycling and £87–132bn without, and potentially much more if the project follows the pattern of every EPR built so far. Part of this is the surcharge consumers are already paying, which would stop immediately. The government would also avoid committing up to £3.8bn of equity and £36.6bn of NWF lending, freeing that capacity for other uses. Against this, compensation to private investors would be around £5bn.
5. Hinkley Point C
5.1. What will the price of power from Hinkley Point C be?
The contracts for the construction of Hinkley Point C were signed in October 2016 when the price paid for power over the length of the 35-year contract was set at £92.5/MWh (2012 prices) indexed to inflation reducing to £89.5/MWh if Sizewell C went ahead. These prices were based on an assumed construction cost of £18bn (2015 prices or £24bn in 2024 prices35) and plant completion by the end of 2025. Construction started in December 2018. Since 2016, there have been five cost and time increases, the first even before construction started, with the latest, in the EDF 2025 annual report36 projecting completion at a cost of £35.2bn (2015 prices or £46.6bn in 2024 prices).37 The project originally involved a consortium of EDF (66.5%) and the Chinese company, CGN (33.5%). However, CGN and EDF were only contractually obliged to invest up to the point that the forecast £18bn cost had been spent and CGN ceased to contribute in 2024. EDF chose to continue to invest. At the end of 2025, CGN’s share in the project had fallen to 23.3%.38 If the plant is completed for £35bn, CGN’s stake will fall to 17%.
The price that will be paid for Hinkley Point C power for the 35 years of the CfD contract will be £89.5/MWh (2012 prices), or £129/MWh in 2025 prices. So even in the three-year period of the highest wholesale electricity prices that there has ever been or, arguably, ever will be, Hinkley Point C would have saved consumers a minimal amount. In other years, it would have been a cost burden on consumers. For the future, there should be no exposure of wholesale electricity prices to fossil fuel prices after 2030 when the remaining fossil fuel, natural gas, is expected to have been largely phased out of the electricity generation sector. The Clean Power 2030 Action Plan anticipates that by 2030, natural gas would provide no more than 5% of Britain’s power and should therefore have little if any impact on power prices.39 By then, the system will be dominated by renewables mostly paid via fixed price CfDs negotiated by the government. This suggests the wholesale market price will tend to be stable and not affected by world energy markets from then on. In February 2026 in its Allocation Round 7,40 the government contracted more than 15GW of new renewable capacity: offshore wind at about £90/MWh, onshore wind at £72/MWh and solar at £65/MWh. Nearly all of this capacity will come online by 2030.
The extra cost consumers will have to pay is about £1bn per year or £35bn over the life of the contract.
In June 2026, the electricity regulator, Ofgem, shortlisted 16 projects for Long Duration Energy Storage (LDES) with a capacity of 7.6GW and 136.9GWh, which it is ‘minded’ to fund. 136.9GWh is equivalent to almost two days of output from Hinkley Point C or Sizewell C at full power. The main technologies are pumped storage hydro (3.9GW of power and 87.8GWh of storage) and battery energy storage systems, mostly lithium-ion (3.6GW and 47.1GWh).41 This will provide the flexibility needed to cope with the variability of renewables (see Annex 1). A further round of LDES projects is expected in 2027.42
So, on the same undiscounted basis, if we assume the long-term wholesale price is £90/MWh and that Hinkley Point C produces 25,000GWh per year at £129/MWh, the extra cost consumers will have to pay is about £1bn per year or £35bn over the life of the contract.
However, the near doubling of the construction cost and the seven-year time overrun means that the £129/MWh electricity price is a gross underestimate of what an economic price based on actual costs and times would have been. Given that, in the ten-year period since contracts were signed, Hinkley’s costs have gone up on five occasions, it seems highly unlikely that in the remaining six years of construction there will not be further cost increases and delays, making the project cost even higher. In 2024 EDF wrote off €12.9bn of its investment in Hinkley Point C because of the cost escalation43. So, the Hinkley Point C deal will achieve the unenviable feat of being a heavy loss-maker for its owners, EDF, and a liability for its consumers. Given the likelihood of further cost increases, EDF would be likely to be quietly relieved if the project was abandoned, to avoid further losses on the plant, especially given that it would expect to be compensated for some of its costs in the event it was abandoned.
5.2. What would the cost of abandonment be?
The costs of abandonment must be set against the estimated £35bn that will be added to consumers’ bills over the 35-year power purchase contract. EDF and CGN were contracted only to invest £24bn (2024 prices). EDF might claim that because they will have spent more than this, they are entitled to more than their share of £24bn. However, it was their decision to continue to invest and the fact that costs have escalated is the result of their failure to complete the plant for the cost they signed up to. Why should consumers have to pay for this failure? So, compensation to the investors ought to be no more than the £24bn the owners were contracted to invest.
There have clearly been massive civil works at the site that would have to be remediated. There has been no part-built nuclear power plants abandoned and the site remediated in the past 35 years so there is no good reference point for the cost. One reference point for remediating the Hinkley Point C site if it was abandoned is the estimated cost of decommissioning Sizewell C. This is essentially a clone of the Hinkley Point C plant, and the claimed worst cost case for decommissioning the plant after 60 years of operation is £12bn (2025 money).44 However, this will be after a lifetime of operation by which time, significant parts of the plant will have become highly radioactive. This will make the task of decommissioning reactors at the end of their life, including disposal of the large amount of radioactive and other toxic waste produced and the cost of workers having to work in hazardous radioactive areas, far more demanding and expensive than demolishing an unirradiated plant. The cost of decommissioning a ‘clean’ plant will therefore be a small fraction of £12bn. So, the cost of compensating investors and remediating the site seems sure to fall well below the savings from not requiring consumers to buy Hinkley Point C’s expensive power.
Table 2: A comparison of Hinkley Point C, Sizewell C and Wylfa assumptions and status
| Hinkley Point C | Sizewell C | Wylfa | |
| Assumptions | |||
| FID to construction start (years) | 2 | 4 | ? |
| Construction time (years) | 7 | 10 | ? |
| Cost estimate £(2024)/reactor at FID | 12 | 20.25 | ? |
| Finance charges during construction | Investors pay | Consumers pay | ? |
| Power price £(2024-25)/MWh | 129 | 133-155* | ? |
| Finance provider | Investors choice | National Wealth Fund | National Wealth Fund |
| Status | |||
| Expenditure committed/spent (£bn) | 2445 | 8.746 | 0.947 |
| Construction status | About 50% built | Construction start 2029 | Investment decision 2029 |
6. Bygones are bygones
British government ministers have repeatedly failed to abandon projects that were clearly going badly wrong preferring to pass the problem to their successor. In the 1970s, a former Treasury economist characterised the AGR and Concorde projects as “two of the three worst civil investment decisions in the history of mankind.”48 In terms of their key characteristics such as construction time and reliability, the five AGR stations ordered in the mid- to late-60s were clearly the worst set of reactors built worldwide in the history of nuclear power. Incomprehensibly, the government ordered two more AGR stations a decade later and the best that can be said about these is that they did not turn out as badly as the first five. In 1979, the Sizewell B project was announced but was not completed until 1995 at a cost of more than £3bn, all paid for by an electricity consumer subsidy that should have gone to pay towards the decommissioning of the UK’s nuclear power plants. A year after its completion, to privatise it, the plant was essentially given away, so the decommissioning provisions were lost. Yet only six years later the privatised company went bankrupt and was saved at huge further public expense. Hinkley Point C is the latest nuclear white elephant, now not expected to be complete until 15 years after its owner, EDF, promised Christmas turkeys would be cooked in 2017 using power generated from it. Experience with the HS2 high-speed rail project suggests the British government has still not understood the lesson that the amount of money spent so far should not influence whether to continue. That money has been spent and cannot be recovered. The decision should be based on the remaining costs and benefits. Bygones are bygones: don’t throw good money after bad.
7. Opportunity costs
The direct costs of failed nuclear projects can be high but the costs that are even more important are the opportunity costs. Pursuing nuclear power will inevitably take resources away from options that would more effectively help the UK meet its climate change goals (see Annex 2). Of all the power generation technologies, nuclear power is by far the most capital intensive and the longest to build and the most likely to fail at some point between project announcement to completion. Nuclear power also consumes a large proportion of the government’s financial and human resources despite its contribution of 2.2% to meeting UK energy demand and 12% of electricity demand.
For the past six decades, almost every British government has operated on the basis that a substantial expansion of nuclear power capacity must be pursued. Despite this, Britain’s nuclear generating capacity is less than half the level it was three decades ago with four large projects abandoned in the past decade and one, Sizewell C going ahead nearly a decade later than planned.49 Rather than come to the obvious conclusion that we should give up with nuclear power, governments continue to give it one more chance. This is grossly distorting government resources, particularly with the National Wealth Fund, as the next section sets out.
7.1. The National Wealth Fund
The UK Infrastructure Bank was set up in June 2021 and renamed the National Wealth Fund (NWF) in October 2024 with a budget of £27.8bn and an objective “to catalyse private investment in the market”. Its projects are usually for between £25m-1,000m with a significant proportion of the cost coming from private-sector banks.
Table 3: National Wealth Fund’s Budget
| Budget (£bn) | |
| Non-nuclear budget | 27.8 |
| Sizewell C | 36.6 |
| Potential new nuclear* | 60 |
| Total | 124.4 |
For the Sizewell C project, it has had to be given an additional budget of £36.6bn with minimal contribution from other banks, the only significant contribution being £5bn in loan guarantees from the French government’s BPI France Assurance Export. The NWF’s participation in the Sizewell C project is in total contradiction to its founding principles. Indeed, the NAO calculates that using public finance, mainly the NWF’s loans, rather than wholly private finance adds around £21bn to the project’s net benefits, which is “sufficient under DESNZ’s modelling to make the net present social values positive”.50 Without the NWF’s cheap money, Sizewell C would not be value for money even on the government’s own figures. The NWF’s lending involves no private sector finance. Its terms, including the interest rate, were imposed on it by the UK government whereas the NWF normally negotiates terms with an applicant. Its financing of Sizewell C, 130% of its budget for other projects, is grossly disproportionate and distortive. Indeed, only £17.8bn of the NWF’s £27.8bn is for loans and equity – the rest is guarantees – so its lending to Sizewell C is about twice what it can lend to all its other projects put together.51 It has also been required to lend up to £599m to Rolls-Royce to complete their SMR design.
If the Advanced Nuclear Framework (ANF) does prove effective in bringing in new nuclear capacity and, if the Cottam, Hartlepool and Oldbury projects do go ahead as well as Wylfa for a total of about 5.4GW of capacity, it is inconceivable the promoters would not look to the NWF for finance. The ANF states: “government will encourage Pipeline projects [such as the three identified above] to engage with NWF on financing proposals.” Based on the budget required for the 3.2GW Sizewell C project of £36.6bn, these four SMR stations would require a budget of about £60bn and would bring NWF’s total budget to about £125bn, about 75% of which would be for nuclear. In its 5-year plan, the NWF states “On top of this we are quantifying our ambitions by driving more than £100 billion in finance by the end of this strategic plan.” The £100 billion target could be consumed entirely by nuclear projects.
Its financing of Sizewell C, 130% of its budget for other projects, is grossly disproportionate and distortive. Indeed, only £17.8bn of the NWF’s £27.8bn is for loans and equity – the rest is guarantees – so its lending to Sizewell C is about twice what it can lend to all its other projects put together.
The ANF states: “Nuclear deals remain high-risk investments and investment decisions will be based on NWF’s rigorous due diligence, commercial assessments and internal governance processes.” There is no evidence that these ‘rigorous processes’ took place with the Sizewell C deal. There must be doubts whether, if the NWF’s requirements seemed to be blocking a project supported by the government, government would not find ways to overrule it.
NWF claims it is “bringing together the strengths of both the public and private sectors to unlock the UK’s future” and it has “selected 25 sectors and their value chains which are important to the UK’s long-term growth and clean energy ambitions, and where we will actively pursue opportunities for private investments and support for regionally significant projects.”52 These ambitions are hard to square with an organisation that is going to spend almost 3.5 times as much on nuclear with little or no private financial participation as the rest of its activities put together. Abandoning these projects would end the distortion of the National Wealth Fund and allow its investments to fully reflect its remit as a national wealth fund rather than a means of making UK consumers poorer.
8. The case for continuing
The main arguments put forward for continuing with these three projects are:
- Power price. Power from the nuclear plants would be cheaper than the alternative low-carbon options;
- The need for base-load power. Nuclear power provides reliable base-load power unlike most renewables which are weather dependent;
- System costs. Because of the variability of renewable output, there is a need for additional features such as storage.
The previous analysis shows that the argument on power price is not credible and, with the RAB model, consumers are facing an addition to their bills more than a decade before first power will be produced. Renewables can clearly be contracted without the need for this upfront payment.
Four of the six nuclear projects that the government forecast would be operating by 2030 failed completely.
The argument on base-load power also does not stand up. Nuclear power has its own inflexibilities and it is problematic to operate nuclear power plants at other than their full available power. The reality is that nuclear power plants do break down and are not the 24/7 source of power they are portrayed as. So, when demand is less than the supply available from renewables and nuclear, it is renewables that must be turned off. The flexibility given by fossil fuel generation is being lost as it is phased out and any combination of nuclear and renewables will need the flexibility offered by Short and Long Duration Energy Storage (see Annex 1) and other measures such as Demand Side Management that allows demand to be reduced when supply is tight.
The success rate of nuclear projects must also be considered. Four of the six nuclear projects that the government forecast would be operating by 2030 failed completely, one (Sizewell C) is only now going ahead and will be at least a decade behind the target and even Hinkley Point C will miss the target by at least two years. It was only the gross over-forecasting of electricity demand by the government and the availability of renewable projects to be rapidly deployed to compensate that prevented the most serious consequences of this failure.
In terms of system costs, the dramatic fall in price of Battery Energy Storage Systems, down 84% in the past decade53, means that system costs have fallen dramatically. Nuclear plants, because of the size of their output, typically 3GW from one site, bring their own system costs of ensuring the demand met by the plant can be instantaneously replaced if it breaks down.
9. Conclusion
It is possible to make some estimate of the costs and benefits of completing Hinkley Point C and Sizewell C and in both cases, the evidence is decisively in favour of immediate abandonment. The case for abandoning Sizewell C is clearly the strongest, but the case for abandoning Hinkley Point C is also strong even after construction has been underway for eight years.
Little is known about the costs and commercial arrangements for the Wylfa project, particularly who will bear the economic risk, but if a deal is to be concluded, it is only likely to be possible if these risks are borne by the public. Wylfa therefore represents a major gamble with public money and there can be little confidence the power produced will be any cheaper than that from Hinkley Point C and Sizewell C and the benefits to the British people of abandoning it correspondingly large. It will also open the door for other SMR projects. Table 4 summarises the costs and savings of abandoning these projects. There is uncertainty about all the numbers but the net position is overwhelmingly in favour of abandonment.
For a new government with a substantial mission for regional growth and tackling the cost-of-living crisis, reallocating funds from nuclear energy to economic growth should be viewed as a real option.
Table 4: Costs and savings of abandoning Wylfa, Sizewell C & Hinkley Point C (£bn)
| Site remediation | Compensation | Consumer savings | Net savings | |
| Wylfa | Negligible | <1? | ? | ? |
| Sizewell C | <1? | 4.8 | 58-98 | 52-92 |
| Hinkley Point C | ~2 | 24 | 35 | ~9 |
| Total | ~3 | ~30 | 93-133 | ~61-101 |
For a new government with a substantial mission for regional growth and tackling the cost-of-living crisis, reallocating funds from nuclear energy to economic growth should be viewed as a real option. While government does not have a fixed amount of capital available, its financial resources are finite and abandoning Hinkley Point C, Sizewell C and Wylfa would reduce pressure on Treasury funds making more capital available for the sort of projects NWF was set up to promote. Cutting DESNZ’s spending on Sizewell and Wylfa – £14.2bn was allocated to Sizewell and £2.6bn to Wylfa in June 2025 to be spent in the following five years – and taking projects such as Cottam and Hartlepool off the table would also substantially reduce DESNZ’s budget, again freeing resources for more productive investments such as in renewables and energy efficiency. It would also save consumers about £93-133bn from avoiding paying above market prices for the output of Hinkley Point C and Sizewell C plus the savings from not having to buy Wylfa’s output. By comparison, the cost of abandonment – remediating the sites and compensating investors – should be less than £35bn. Most of that is compensation for Hinkley Point C, which might be paid in instalments rather than up front, matching the savings to consumers as they accrue over the 35-year contract. Abandoning Sizewell C, by contrast, saves public money from the start: the consumer surcharge ends, and the NWF loans and government equity are never committed. Redirecting funding from the UK’s expensive and uncompetitive nuclear projects into its new policy priorities is one of the few large-scale sources of funding available to it outside further borrowing. There are not that many win-win options open to the Government. This is one. It should be assessed by both the Prime Minister and the Chancellor.
Annex 1: Long Duration Energy Storage
Long Duration Energy Storage (LDES) covers storage facilities that can store energy for at least eight hours and up to weeks. In June 2026, Ofgem published a ‘Minded to Decision’ list of 16 LDES projects with 7.645GW with a capacity of 136.9GWh.54 The largest element is the three Pumped Storage Hydro (PSH) projects with a capacity of 3.9GW and a duration of 15-32 hours. The next largest element is 11 Lithium-Ion Battery Energy Storage Systems (Li-Ion BESS) ranging from 100-500MW and a duration of 8-18 hours. The other two projects were Compressed Air Energy Storage, CAES (50MW, 30 hours) and Vanadium Redox Flow Batteries, VRFB, (65MW, eight hours). A final decision on which projects to fund is expected in Autumn 2026.
The agreement will be a ‘cap and floor’ scheme designed to reduce the risk to investors that their income will not cover their costs. The ‘floor’ will guarantee developers a minimum level of income if market revenues are lower than expected and will be funded by consumers. The cap sets a level above which a share of additional revenues is returned to consumers.
These projects were chosen from 77 projects shortlisted by Ofgem with a capacity of 27GW to fill between 2.7-7.7GW of capacity. A second window for awards is expected in 2027. While PSH is a long-established technology, battery technology is moving very fast and Li-ion BESS was not going to be included initially.
Annex 2: Climate Change Targets
The UK is struggling to meet its climate change targets. None of the six nuclear projects55 (18GW) that were forecast to be online by 2030 by the Brown and Cameron governments will meet that target date. The reason the UK is not more off target is because renewable projects – which are built quickly and to time and cost, typically in four years – have been drafted in to fill the gap despite an arbitrary ban on new onshore wind power projects for a decade up to 2025. In addition, forecasts that electricity demand would increase by about 25% between 2005-25 have proved wrong with demand falling by 25%. This gross over-forecasting of demand by the government, used by them to justify large nuclear programmes, has been a feature since the mid-1960s. The expectation has always been that this demand growth would be driven by new applications requiring electricity and now the needs of data centres, battery vehicles and electric space-heating appear likely to drive electricity demand up substantially. However, in the past, new applications have always been partly offset by electricity efficiency improvements and there is no reason to think this will not be the case in the future especially if the government were to adopt a more active energy efficiency policy.
Government effectively abandoned its major energy efficiency programme more than a decade ago. If renewables and storage were pursued more vigorously and systematically and energy efficiency programmes, especially those targeted at low-income households, reintroduced, not only would our climate change goals be achieved more quickly, more reliably and at lower cost, but the price of electricity would also be reduced and fuel poverty would be reduced.
References
1. In June 2026 Great British Energy – Nuclear renamed the Wylfa site Gwyndod. However, most people will know the site from its original name and in this report, we will refer to it as Wylfa.
2. Hartlepool, Heysham A, Heysham B and Torness.
4. For an analysis of the Sizewell C project, see https://policybrief.org/briefs/sizewell-c-the-last-of-its-kind/
5. For an analysis of the Wylfa project, see https://policybrief.org/briefs/how-the-sun-went-down-on-uk-nuclear/
6. https://www.gov.uk/government/publications/advanced-nuclear-framework
7. https://www.gov.uk/government/news/golden-age-of-nuclear-delivers-uk-us-deal-on-energy-security
8. https://holtecinternational.com/hh-41-09/
9. https://sge.eu/sge-unveils-plans-for-4-2gw-uk-small-modular-reactor-fleet/
10. https://committees.parliament.uk/oralevidence/17523/pdf/ q 326
11. https://www.gov.uk/government/news/rolls-royce-smr-selected-to-build-small-modular-nuclear-reactors
13. https://www.theferret.scot/torness-bid-cracked-reactors-safety/
14. https://www.gov.uk/government/news/rolls-royce-smr-selected-to-build-small-modular-nuclear-reactors
15. https://www.gov.uk/government/publications/spending-review-2025-document/spending-review-2025-html. The Spending Review announced “over £2.5bn” for the SMR programme, subsequently confirmed as £2.6bn.
16. The French EPR design has been superseded by the EPR2 design after winning only eight orders.
19. https://www.gov.uk/government/news/sizewell-c-gets-green-light-with-final-investment-decision
21. https://searchforuksubsidies.beis.gov.uk/scheme/?scheme=SC11357
22. If the real discount rate is 3%, a cost of £1bn incurred in 50 years will show as only £228m in a discounted cash flow calculation.
23. The project is expected to be funded by 35% equity (investors’ own cash) and 65% debt (borrowing)
25. https://committees.parliament.uk/writtenevidence/167114/html/ and Julia Pyke, Joint Managing Director of Sizewell C, ‘This is the solution to Britain’s energy security problems’, the Telegraph, 21 August 2025. She quotes £60-70/MWh in 2012 prices.
26. This ‘longstop date’ was set at 2033 when the FID was taken but renegotiated by EDF to 2036 in 2022.
27. https://www.nao.org.uk/wp-content/uploads/2026/05/sizewell-c.pdf
28. https://www.nao.org.uk/wp-content/uploads/2026/05/sizewell-c.pdf pp 9 and 31.
29. NAO, Sizewell C, HC 33, May 2026, para 15. The ‘baseline’ cost and time estimate is the estimate made by Sizewell C Ltd and assumes a cost of £38.2bn and completion in July 2039. It excludes some costs included in the regulatory thresholds, mainly a £1.6bn payment to Hinkley Point C (NAO, HC 33, para 1.13).
30. https://www.nao.org.uk/wp-content/uploads/2026/05/sizewell-c.pdf p 47
31. https://www.centrica.com/media/tfnbsmm5/centrica-2025-sizewell-c-rns-announcement-final-220725.pdf
32. https://www.centrica.com/investors/centrica-as-an-investment/investing-for-value-at-sizewell-c/
33. https://www.gov.uk/government/publications/contracts-for-difference-cfd-allocation-round-7-results p 4
34. https://www.sizewellc.com/wp-content/uploads/2026/07/Annual-Report-2025-26.pdf p 5
35. We use 2024 prices because that is the basis used for Sizewell C cost estimates.
36. https://www.edf.fr/sites/groupe/files/2026-04/2026-04-02-urd-edf-2025-en.pdf
37. https://www.edf.fr/sites/groupe/files/2026-02/annual-results-edf-2025-presentation-2026-02-20.pdf p 14
38. https://www.edf.fr/sites/groupe/files/2026-04/2026-04-02-urd-edf-2025-en.pdf p 76
40. https://www.gov.uk/government/publications/contracts-for-difference-cfd-allocation-round-7-results p 4
44. https://questions-statements.parliament.uk/written-statements/detail/2025-09-01/hlws892
45. https://www.edf.fr/sites/groupe/files/2026-04/2026-04-02-urd-edf-2025-en.pdf p 77 and https://www.edf.fr/en/the-edf-group/dedicated-sections/journalists/all-press-releases/final-contracts-signed-for-hinkley-point-c
46. To end fiscal year 2025/26 https://www.sizewellc.com/wp-content/uploads/2026/07/Annual-Report-2025-26.pdf p 19
47. To June 2026 https://www.greatbritishenergynuclear.uk/uk-smr-programme-at-wylfa-hitting-target-for-british-contracts/
48. https://www.bsg.ox.ac.uk/sites/default/files/2023-12/BSG-WP–2023-056-When-Missions-Fail.pdf
49. The British government claimed 18GW of new nuclear capacity would be online by 2030. The Hinkley Point C and Sizewell C projects are going ahead but will miss the 2030 target, in the case of Sizewell C by at least a decade. The Wylfa Newydd and Oldbury projects were abandoned because no funders could be found, the Moorside project was abandoned when its owner, Westinghouse, applied for Chapter 11 bankruptcy protection and the Bradwell project was blocked by the UK government because of its ownership by a Chinese company.
50. NAO, Sizewell C, HC 33, paragraph 2.9.
51. https://www.gov.uk/government/publications/spending-review-2025-document/spending-review-2025-html. See Table 4.1: Public Financial Institution Capacity.
52. https://www.nationalwealthfund.org.uk/about-us/our-strategy/
53. https://www.worldnuclearreport.org/IMG/pdf/wnisr2026-v1.pdf p 429
54. https://consult.ofgem.gov.uk/energy-generation/ldes-window-1-minded-to-decision/supporting_documents/ldes-window-1-minded-to-decisions-consultationpdf, https://www.energy-uk.org.uk/publications/energy-uk-explains-long-duration-electricity-storage-and-ofgems-cap-and-floor-decisions/
55. Hinkley Point C, Sizewell C, Wylfa, Oldbury (Hitachi-GE), Moorside, Bradwell B.
