The Government is onto a false trail on the UK’s nuclear troubles

Introduction

There is a pervasive narrative that the UK is uniquely poor at building nuclear power plants in terms of costs and construction times. The dominant cause of these delays and high costs is claimed to be planning and regulatory processes which are slow and obstructive. The Regulatory Taskforce and Prime Minister Sir Keir Starmer have been particularly prominent in expressing these views. When he announced the setting up of the regulatory taskforce the Prime Minister stated: “For too long, blockers have had the upper hand in legal challenges – using our court processes to frustrate growth. We’re putting an end to this challenge culture by taking on the NIMBYs and a broken system that has slowed down our progress as a nation.”¹  This was a clear sign to the Taskforce of the conclusions he expected them to come to.

The Regulatory Taskforce stated that the UK “has become the most expensive place in the world to build nuclear projects.” And that the planning and regulatory system was “not fit for purpose”.² What has been less reported, and which has seen little comment from the government are the highly critical comments on the government’s role and that of industry:

• Underlying laws and regulations prioritize process over outcomes, leading to time-consuming delays and suboptimal decisions;
• Government departments are often slow and indecisive in their roles as policymakers and regulators, failing to provide clear direction;
• The near-monopolistic status of much of the industry provides weak financial incentives to reduce costs or challenge disproportionate regulatory decisions.

What is missing from the Government’s and the Taskforce’s conclusions is any analytical evidence to back up their conclusion that regulation and planning processes need to be deregulated.

Why have the UK nuclear projects announced since 2006 failed?

There have been six large reactor projects announced since Tony Blair stated that nuclear power was “back on the agenda with a vengeance” in May 2006.³ These were:

• The Hinkley Point C and Sizewell C projects each comprising two French-designed European Pressurised Reactors (EPRs) to be owned by consortia comprising EDF and a Chinese company, China General Nuclear (CGN);
• The Bradwell project comprising two Chinese design HPR1000s to be owned by CGN with EDF as a minority partner;
• The Wylfa and Oldbury projects each expected to comprise two Hitachi Advanced Boiling Water Reactors (ABWRs) to be owned by a consortium, Horizon, of several large European utilities;
• The Moorside project comprising three Westinghouse AP1000s to be owned by the two largest German utilities.

All were expected to be built using the financial model implemented with the Hinkley Point C project, although this was never formally stated, and all were expected to be in operation by 2030 except Bradwell B which never had a forecast completion date.

Hinkley Point C

A Final Investment Decision (FID) for Hinkley Point C was taken in October 2016 nine years after EDF announced its intention to pursue this project. In 2007, EDF claimed the first power from the plant would be in 2017.⁴ By 2013, when initial agreement on the deal was announced⁵ project completion was expected for 2023. By 2016, it was expected the second reactor would be completed before the end of 2025 at a cost of £18bn (in 2015 money).  The regulatory process introduced by the Blair government in 2007, Generic Design Assessment, was expected to take four years. This target would have been achieved with the European Pressurised Reactor (EPR) but the process was extended by a year to 2012 to allow the implications of the Fukushima disaster for Hinkley and the other UK reactors to be considered.⁶ The main cause of the delays in reaching the FID appears to be government and EDF negotiating on the deal and EDF trying to bring in other investors so its stake was less than 50%.

The latest forecast from January 2024⁷ is that the second unit would be completed as late as 2032 at a cost of up to £35bn (in 2015 money). A key element to the deal was that the power purchase price was set in real terms for the duration of the 35-year contract at £92.5/MWh (in 2012 money)⁸ with EDF responsible for paying for any delays and cost overruns. The cost overruns resulted in them having to write-off €12.9bn of their investment in Hinkley in 2023.⁹ This was a major factor in EDF refusing to go ahead with Sizewell C using the same financial model.

Sizewell C

The deal to take this project to Final Investment Decision (FID) was signed at the same time as the Hinkley deal but by 2018 EDF had made it clear it would not proceed with the project on comparable terms to those of Hinkley Point C. In part this was due to the risk of a fixed power purchase price outlined above but EDF would also have struggled to find the cash to finance its expected 80% share of the equity. EDF promoted the Regulated Asset Base (RAB) model as an alternative with EDF not expected to be a significant equity partner. It was not till 2022 that the government had passed the legislation needed to use this model. The next three years were largely spent by the government trying to persuade investors to invest but government was only able to cover about 40% of the cost from sources other than its own and EDF. The central estimate for completion is 2039, a lead-time from FID of five years more than was forecast for Hinkley Point C in 2016. The forecast cost is about 70% more in real terms than the estimated cost of Hinkley Point C was at FID. How far these increases are due to the estimates being more realistic and how far they are to the lead-time and real cost having gone up in the past decade will not be known until Sizewell C is complete. As with Hinkley, delays at Sizewell in getting to FID were not caused by regulatory and planning delays. Any that arose were not on the critical path. The delays were largely down to financing difficulties and government processes.

Bradwell B

The third project agreed in 2016 with the EDF/CGN consortium, this time led by CGN (80%), was for two Chinese design reactors at Bradwell. The Office for Nuclear Regulation (ONR) successfully completed its Generic Design Assessment (GDA) on the design in 2022.¹⁰ Little other progress was made with the project and a Development Consent Order (planning permission) was not applied for. The government effectively vetoed it in 2021 because of the national security issues raised by CGN’s involvement although the project has yet to be formally abandoned.

“As with Hinkley, delays at Sizewell C in getting to Final Investment Decision were not caused by regulatory and planning delays. Any that arose were not on the critical path. The delays were largely down to financing difficulties and government processes.”

Wylfa/Oldbury

The consortium of European utilities, Horizon, which set up these two projects, sold the venture in 2013 to the Japanese reactor vendor, Hitachi. The utilities withdrew because the impact of the Fukushima disaster made new nuclear projects problematic and because the opening to competition of electricity markets meant utilities could no longer justify building economically risky and probably uneconomic facilities. Hitachi does not have the financial capability to finance nuclear power plants, and it hoped it would be able to sell on the projects when the they were further advanced but still using Hitachi’s reactors. In 2019, the Wylfa venture collapsed because of a lack of investors despite the UK government offering to take a 30% equity stake and providing all the finance. Little progress had been made with the Oldbury project which was also abandoned.

Moorside

In 2012, the NuGen venture by two German utilities was sold to Toshiba, the then owner of Westinghouse reactor vendor, because of the Fukushima disaster and the confirmation of the nuclear phase-out policy in Germany. As with Horizon, the plan was to progress the project then sell it on. However, in 2017, Westinghouse filed for Chapter 11 bankruptcy protection and the project collapsed.

Summary

The failures and delays with these projects have not come from obstructive planning or regulation. The issues appear to have been the difficulty of finding finance partly because of the scale of investment needed and partly because of the economic risk of nuclear projects. Delays and cost overruns at the one project to start construction, Hinkley Point C, appear to result from the difficulty of building the design and the skills of the contractors.

Acoustic Fish Deterrent

The measures to protect fish from being killed by being drawn into the coolant system are often cited by those claiming the problems with building reactors in the UK are due to over-zealous regulators and planners. This is wrong. The case for Hinkley Point C put forward by EDF and approved by the Planning Inspectorate included measures to protect fish from the cooling system. The number of fish that would be killed is not precisely known. EDF estimated that the removal of the AFD would result in the death of between 1.89 million and 2.9 million fish annually.¹¹ However, a report from the Welsh Government indicated that the removal of the AFD could lead to the death of at least 182 million fish each year.¹² The Environment Agency estimated that 178 tonnes of fish would be drawn into the cooling system every year.¹³ This is not an issue just for Hinkley Point C, it will arise for coastal sites including Sizewell and Wylfa.

Planning consent given in 2013 required EDF to include three measures to protect fish: an Acoustic Fish Deterrent (AFD); low-velocity side intakes; and a Fish Recovery and Return system. In 2022, EDF appealed against the need for an AFD¹⁴ claiming that installing the system would mean an unacceptable safety risk to those maintaining it.¹⁵ Their appeal was rejected in August 2022 by the UK Secretary of State for the Environment.¹⁶ EDF attempted to side-step the requirement by offering to build 800 acres of saltmarsh on the River Parrett instead, a plan that also raised environmental concerns.¹⁷ Whether this would have been a worthwhile project is not the issue, it clearly would have done nothing to prevent the killing of a huge number of fish in Severn Estuary. The saltmarsh plan was effectively abandoned in early 2025 when EDF claimed they now had an AFD that was workable, albeit at a cost of £700m.¹⁸ However, it appears the £700m covers the cost of all three systems required in 2013 with the AFD accounting for only £50m of the £700m.¹⁹

EDF has hinted that the process of deciding on its appeal could have delayed the completion of the station, but it has a strong incentive to suggest that the delays at Hinkley are not its fault. They have provided no evidence that there have been any actual delays due to the AFD issue. EDF clearly has a commercial interest in portraying the damage from not installing AFDs as low as possible while estimating the cost as high as possible so that the proportionality of the measures is called into question. A cost of £700m to save less than 3 million fish per year is a very different proposition to a cost of £50m to save 182 million fish every year.

One of the principles for licensing nuclear power plants is that harm should be As Low As Reasonably Practicable (ALARP) so that any damage is as low as possible within economic bounds. However, what ALARP does not tell you is whether the remaining damage is acceptable. So, a facility might include state-of-the-art protection measures that met the ALARP criterion, but the remaining risk might still be unacceptable.

“The failures and delays with these projects have not come from obstructive planning or regulation. The issues appear to have been the difficulty of finding finance partly because of the scale of investment needed and partly because of the economic risk of nuclear projects.”

Most expensive place to build reactors in the world

One of the often-repeated claims of those advocating deregulation of the nuclear sector is that the UK is the most expensive place in the world to build nuclear power plants. However, the analytical evidence to back this claim is missing. Making a valid international comparison of costs will inevitably not be precise because of the difficulties raised in putting costs on a comparable basis correcting for varying currency exchange rates and inflation in the different countries involved. However, making meaningful comparisons is made much more difficult by the dearth of reactor orders in the past two decades. To make a fair comparison, it is necessary to focus only on designs that meet current European and US regulatory standards – older technologies lacking currently required systems will be cheaper and easier to build. It is also necessary to focus on countries that can offer reliable, audited costs.

On these criteria, the UK has the incomplete Hinkley Point C project, which is about halfway through its construction period²⁰. The comparators in Europe are Finland with the Olkiluoto EPR completed in 2023 after 18 years of construction and France with the Flamanville EPR, in February 2026 still incomplete after more than 18 years. In the USA, the two Vogtle AP1000 reactors took 10-11 years to build compared to the four years forecast and at about three times the estimated cost. The two Summer AP1000s were abandoned after four years of construction because costs and times were out of control. Whether the record of the UK will turn out to be worse than that of the other countries seems only of academic interest given that all five projects were all clearly economic disasters.

Korea and China are often cited as countries where reactors can be built to time and cost. For China, the evidence is weak. Independently audited costs are not available. Given that the three owners of most of the plants are state-owned and are also the three reactor vendors competing to win export orders, they have a strong incentive to portray the cost of their reactors as low. Labour costs in China are far lower than Europe and the USA and this will inevitably reduce costs. If we are to draw the conclusion that we should emulate China, we would need to take account of these lower labour costs and be confident the standards of quality control, regulation and democratic accountability in China are comparable to those required in Europe and the USA. Many of the reactors completed in China in the past two decades use an old design dating from the 1970s that would not be considered elsewhere. The six reactors built using Western technology (four AP1000s and two EPRs) have the longest construction times of almost all the Chinese reactors – 10 years compared to the four years planned and not much quicker than the Vogtle project in the USA.

Claims that the UK is one of the most expensive places to build nuclear power plants are wide of the mark. Comparator projects in Finland, France and the US show similar overruns of cost and delays in construction.

Korea is also often cited as a country building reactors to time and cost.²¹ As with China, there is the issue of authenticity of costs. The utility (KEPCO) is a state-owned monopoly and is also the Korean reactor vendor. It has strong reactor export ambitions so has a clear incentive to portray low costs and times. It has completed four reactors in Korea using its most recent design, the APR1400. The first took nine years to build with the following three taking 11 years, giving no evidence for the positive effect of learning. KEPCO has acknowledged that the APR1400s it has built (four each in Korea and UAE²²) would not meet European regulatory standards. These would require a system such as a core-catcher to contain the impact of a core melt-down and a toughened reactor shell capable of withstanding impact from an aircraft.²³ Adding these would substantially increase costs.

Korea’s technology is licensed from Westinghouse who will not allow it to be built in Europe apart from in the Czech Republic so Korean technology is not an option for the UK.²⁴

UK can regain its position as global leader in nuclear technology

In its final report, the Taskforce claimed that implementing its recommendations would: “allow the nation to reclaim its position as a global leader and capitalise on the worldwide nuclear renaissance, ensuring a secure and prosperous future.”²⁵ This is a misrepresentation of the UK’s status. Along with the USA and Russia, the UK was one of the first countries to generate electricity from nuclear power plants but with a technology, Magnox, which was not pursued. It has never been a significant supplier of nuclear technology outside the UK. The record of its first reactor technology, the Magnoxes, was mediocre and that of its successor technology, the Advanced Gas-cooled Reactors, was disastrous.²⁶ Only two reactors, both Magnox design, have been exported, both ordered in 1959. Other UK nuclear technologies, such as the Thermal Oxide Reprocessing Plant (THORP) and the Mixed Oxide fuel plant have an even worse record. Given the poor economic record of nuclear technology in general, customers will be looking to reduce the risk by choosing suppliers with a substantial and successful track record in the technology. It is highly unlikely that the UK could achieve a position as a major nuclear technology supplier in the foreseeable future. The UK will need to build its nuclear skills but in decommissioning and waste disposal. It will inevitably be one of the pioneers in these areas but only because a much high proportion of its nuclear facilities than almost all other nuclear countries have already been shut down.

There is also the issue of whether being a technology leader in supplying new reactors is worth aiming for. If, like previous claimed global nuclear revivals, the new one peters out with few orders placed, being a world leader will be no more valuable than leading the world with hovercraft technology.

Predictable series ordering of standardised designs will reduce costs

This has been a claim made by the nuclear industry for more than 50 years. It is a strategy that has only been tested in France, which started construction of 58 reactors in the 20 years between 1971-1991. These were based on a design that evolved slowly through ‘tranches’ of 4-16 reactors with the design within each tranche fixed. It is widely assumed that the predictability, standardisation and scale of this flow of orders led to falling real costs and construction times.²⁷ The evidence is to the contrary.

“The UK will need to build its nuclear skills but in decommissioning and waste disposal. It will inevitably be one of the pioneers in these areas but only because a much high proportion of its nuclear facilities than almost all other nuclear countries have already been shut down.”

The first eight reactors took on average 70 months to build while the last eight averaged 135 months with the slowest of the first eight taking two years less than the fastest of the last eight. There have been several analyses of French construction costs using official French data and all show the real cost of French reactors increased significantly over the period of the programme. The US energy economist Charles Komanoff found the real cost had increased by 60%²⁸; the Austrian scholar Arnulf Grübler characterised the French programme as “a case of negative learning by doing;”²⁹ the US nuclear research scholar, Mark Cooper, found that “French reactors have escalated in price at almost the same rate as those in the U.S.”³⁰

Responses to the Regulatory Review

While the press response to the Regulatory Taskforce was generally positive, there has been an increasing flow of criticism from expert bodies. Ruth Wilkinson, Institution of Occupational Safety and Health, contested the claim that relaxing health and safety rules was an effective way to speed up UK infrastructure projects. She said: “Stripping back protections in the name of speed is a false economy that risks lives, reputations and resilience. The UK’s health and safety framework is the backbone of safe and sustainable growth.”³¹

The Wildlife Trusts said: “The Nuclear Regulatory Review recommendations 11, 12 and 19³² will harm nature and biodiversity. They are based on flawed evidence relating to environmental regulations and how they have been applied.”

Conclusions

If policy is to be well-founded, it should be evidence-based, not based on shallow unsupported assertions that seem designed to generate headlines. The narrative that the problems in the UK are down to obstructive regulation and planning and regulatory processes; that other countries can build reactors to time and cost; and that the UK could become a major player in an expanding world nuclear market is a fantasy. If it leads to measures to deregulate planning and regulation, the consequences could be severe.

Two days after the publication of the Taskforce’s final report, the Prime Minister stated: “I warmly welcome it and endorse its approach and accept the principle of all the recommendations it has set out. We have committed to complete implementation within two years, subject to legislative timelines on elements requiring primary legislation.”³³ Such a precipitous response does not suggest careful consideration of the implications of the radical reforms proposed by the Taskforce.

The danger is that if government effort goes into solving an issue that is not the cause of the problems, a large amount of time and public money will be wasted failing to solve problems which may be insoluble. This is time and money we cannot afford given the need to take timely and effective measures to mitigate the climate emergency. The integrity of a regulatory system about which, less than two years ago, the government was highly complimentary about could be jeopardised. The government stated in its January 2024 ‘Roadmap’: “The UK is also a world leader in nuclear safety and security – thanks to our commitment to rigorous regulation and the creation of the internationally recognised Office for Nuclear Regulation (ONR).”³⁴ Following deregulation in the nuclear sector will only divert attention from the reality that nuclear power plants are not close to being economic; that their cost continues to increase in real terms as it has done over the entire history of nuclear power unlike competing options which are generally falling in cost; and that the foreseen worldwide nuclear revival may well, like its predecessors, not materialise. So, pursuing global leadership in a technology that has been in decline for at least 40 years would be a folly.

ENDNOTES

1. https://www.gov.uk/government/news/prime-minister-clears-path-to-get-britain-building

2. https://www.gov.uk/government/news/taskforce-to-tackle-regulatory-barriers-holding-back-nuclear

3. https://www.theguardian.com/environment/2006/may/16/energy.business

4. https://www.theguardian.com/news/2021/nov/02/questions-remain-over-the-uks-nuclear-power-plans

5. https://www.gov.uk/government/news/initial-agreement-reached-on-new-nuclear-power-station-at-hinkley

6. https://www.onr.org.uk/generic-design-assessment/assessment-of-reactors/uk-european-pressurised-reactor-uk-epr

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

8. This will fall to £89.5/MWh if the Sizewell C deal goes ahead.

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

10. https://www.onr.org.uk/generic-design-assessment/assessment-of-reactors/uk-hpr-1000

11. https://bristolavonriverstrust.org/hpc-impact-to-fish/

12. https://bristolavonriverstrust.org/hpc-impact-to-fish/

13. https://www.bbc.co.uk/news/uk-england-somerset-66582623

14. https://assets.publishing.service.gov.uk/media/6310b3e08fa8f5579e65ef94/environmental-permit-appeal-app-epr-573-hinkley-point-c.pdf

15. https://nsip-documents.planninginspectorate.gov.uk/published-documents/EN010012-006627-DL6%20-%20Fish%20Guidance%20Systems%20Ltd.pdf

16. https://assets.publishing.service.gov.uk/media/6310b3e08fa8f5579e65ef94/environmental-permit-appeal-app-epr-573-hinkley-point-c.pdf

17. https://www.edfenergy.com/media-centre/hinkley-point-c-sets-out-plan-create-somerset-saltmarsh

18. https://www.edfenergy.com/energy/nuclear-new-build-projects/hinkley-point-c/supporting-the-environment/acoustic-fish-deterrent and https://www.edfenergy.com/media-centre/enabling-clean-energy-through-smarter-proportionate-nuclear-regulation

19. https://www.wildlifetrusts.org/sites/default/files/2026-01/WhyTheNuclearRegulatoryReviewIsFlawed_TheWildlifeTrusts.pdf

20. Its predecessor, Sizewell B, was completed 30 years ago using a pre-Chernobyl design that lacked the safety features that would now be required.

21. https://committees.parliament.uk/oralevidence/16943/pdf/ Q31

22. The construction record of APR1400s in UAE is barely any better than in Korea with an average of 9 years and no improvement over the programme. https://committees.parliament.uk/publications/46805/documents/240895/default/

[23] Nucleonics Week, “No core catcher, double containment for UAE reactors, South Koreans say”, 22 April 2010

[24] https://www.neimagazine.com/news/khnp-withdraws-from-polish-nuclear-project/?cf-view

[25] https://www.gov.uk/government/publications/nuclear-regulatory-taskforce/nuclear-regulatory-review-2025-summary#conclusion-a-path-to-renewed-uk-nuclear-leadership

[26] The 14 AGRs took on average 14 years to build with a lifetime load factor of 68% compared to 87% claimed for them when the technology was chosen. None of them was able to operate as designed typically only reaching about 90% of design capacity. Dungeness B took 22 years to build and had a life-time load factor of 45%.

[27] https://committees.parliament.uk/oralevidence/16943/pdf/ Q31

[28] https://www.komanoff.net/nuclear_power/Cost_Escalation_in_France%27s_Nuclear_Reactors.pdf

[29] https://www.sciencedirect.com/science/article/pii/S0301421510003526

[30] https://thehill.com/blogs/congress-blog/energy-a-environment/70994-frances-nuclear-miracle-is-more-fantasy-that-fact/

[31] https://www.theguardian.com/global-development/2025/dec/18/dont-weaken-health-and-safety-rules-in-the-name-of-growth

[32] Recommendation 11 calls for various changes to the Habitats Regulations, including removing the requirement for compensation to be like-for-like. Recommendation 12 calls for nuclear developers to be allowed to comply with the regulations simply by paying a fixed sum (an amount per acre), which would be used by Natural England for nature somewhere else. Recommendation 19 would remove the duty on Local Authorities to seek and further National Parks and Landscapes, returning to the old language of “have regard to”. https://www.wildlifetrusts.org/sites/default/files/2026-01/WhyTheNuclearRegulatoryReviewIsFlawed_TheWildlifeTrusts.pdf

[33] https://www.gov.uk/government/publications/prime-ministers-strategic-steer-to-the-nuclear-sector/prime-ministers-strategic-steer-to-the-nuclear-sector-following-the-2025-nuclear-regulatory-taskforces-review

[34] https://assets.publishing.service.gov.uk/media/65c0e7cac43191000d1a457d/6.8610_DESNZ_Civil_Nuclear_Roadmap_report_Final_Web.pdf