Britain is the most expensive place in the world to build new nuclear power stations

Britain Remade has reviewed every nuclear project built since 2000: Britain is the most expensive place to build per kilowatt (kW) of capacity.

Hinkley Point C is estimated to cost £46 billion, or £14,100 per kW: when finished, it will be the most expensive nuclear power station ever built.

British-built plants cost far more per kW than peers: our per-kW costs are about six times South Korea’s, and France and Finland deliver the same EPR design for less per kW (27% and 53% respectively).

Britain has gone backwards on cost: Sizewell B in 1995 cost £6,200 per kW, less than half Sizewell C’s budgeted cost.

Britain cannot end its reliance on gas with renewables alone

Britain needs to end reliance on expensive, unstable, high-carbon gas: it is costly, exposes us to global shocks, and drives climate change.

Renewables may have seen large price-falls over the last 15 years but at high penetrations costs linked to managing intermittency are high: Britain has added 40 GW since 2010 and 120 GW is forecast by 2030, yet balancing, curtailment, backup, and overbuild still add cost and leave gaps that require firm power.

The most important technology of the 21st Century is AI: it will generate a large increase in baseload energy demand that wind and solar are ill-suited to meeting without substantial overbuilding.

Britain needs cheaper nuclear

Britain Remade have modelling tested different nuclear build costs and renewable price scenarios to assess long-term impacts on household energy bills: at Hinkley Point C’s current costs, another large nuclear plant would raise bills by about £6 per household, even if renewables costs rose 30% above government forecasts.

At French or Korean build costs, new large nuclear plants would cut bills: this would save £4–6 billion over 25 years for a single plant, and more with multiple builds.

If renewable costs rise, nuclear can play an even larger role: government forecasts assume falling solar costs (-27% by 2040) and modest wind cost drops (-6%), but recent data show solar prices flat and wind costs rising. If renewables costs climb 30% above baseline, the most cost-efficient plan would be eight new plants at French prices (saving £7.6 billion) or fifteen at Korean prices (saving £21 billion) over 25 years, with benefits lasting decades.

Britain’s high nuclear costs are driven by multiple factors

Slow, resource-intensive consultation, planning, and permitting: Sizewell C’s planning process involved seven consultations, an 80,000-page environmental impact assessment, and multiple legal challenges before construction could begin.

Over-specification driven by political demands and gold-plated regulation: Hinkley Point C made thousands of design changes to meet environmental and safety rules, including unique features like a fish-return system and acoustic fish deterrent.

Stop-and-start approach to construction: with no new nuclear plants built in three decades, Hinkley Point C had to rebuild supply chains from scratch; when construction began, there was only one nuclear-qualified welder in the UK.

Lack of standardisation and frequent design changes: unlike its French and Finnish counterparts, Hinkley Point C includes extra safety systems such as a separate analogue control system, additional diesel backup, and an extra spent-fuel-pool cooling train.

Regulation acts as a barrier to cost-reducing best practices

Limiting gains from learning across projects: Frequent design changes and site-specific mitigations disrupt standardisation, undermining the cost savings seen when building multiple identical plants; for example, welding at Hinkley Point C’s second reactor is four times faster than at the first, yielding an estimated 30% saving.

Making it harder to innovate in reactor design and construction: Safety case requirements and planning rules slow adoption of proven improvements; Hinkley Point C needed extra planning approval to change its waste-storage method.

Deterring investment in supply chains: Stop-start construction and uncertainty discourage spending on specialist skills and equipment, reducing the benefits of fleet builds; for example, foreign welders were hired instead of training new British ones.

Creating barriers to entry for disruptive competition: Lengthy, expensive licensing and planning processes make it harder for new entrants to challenge incumbents with lower-cost or innovative designs.

Increasing the risk associated with financing fleets: The fleet model depends on early, settled decisions, but planning delays or forced design changes can make multi-plant financing unviable.

Preventing the use of modular construction: Added safety and environmental features, like hard-wired backup control and instrumentation systems, increase design complexity and reduce the ability to use cost-saving factory-built modules.

Safety regulation is disproportionate to the risks of radiation

Nuclear is one of the safest energy technologies: Per unit of electricity, it is as safe as wind or solar and between 100 to 1,000 times safer than fossil fuels. Deaths due to nuclear accidents are extremely rare. Chernobyl is estimated to have caused only around 200 premature deaths, with none from other civil nuclear accidents such as Fukushima and Three Mile Island.

Radiation is natural and nuclear power is not the main source: Average Britons receive about 2 mSv/year from natural background radiation, rising to around 7 mSv/year in Cornwall due to high radon levels. Coal plants and rare-earth mining for renewables expose people to more radiation than living near a nuclear plant.

Risks are low in practice: The highest-exposed person near Hinkley Point receives just 0.32 mSv/year (two chest x-rays, or 320 g of Brazil nuts), and even in a worst-case core meltdown at Hinkley Point C, modern safety systems would limit public exposure to a few mSv—less than half a year’s natural background in Cornwall.

Britain regulates radiological risk under Health and Safety Law: The key principle is ‘As Low as Reasonably Practicable’, or ALARP. Below legal limits (1 mSv/year in normal operation) dutyholders are required to reduce risks until further measures would be grossly disproportionate. ALARP is not intended to require constant improvement, yet in practice requires design changes (or at least safety cases) covering extremely low levels of radiation exposure.

The UK applies a goal-based rather than prescriptive system: Nuclear site licence holders are free to choose how they meet safety goals, which allows flexibility and innovation. Yet the open-ended nature of ALARP creates uncertainty, often pushing operators to go beyond what the law requires, adding costly safety measures that provide little additional benefit.

The UK’s application of ALARP drives costly add-ons to cut already tiny risks: Independent economic analysis of the Office for Nuclear Regulation’s (ONR) practices found that cost-benefit analysis is almost never used to challenge design-change requests.

Dutyholders rarely contest recommendations because challenge is costly and slow: Preparing safety cases and gross-disproportion analyses takes months and can only be done by expensive specialist teams. Timelines are uncertain and repeated challenges risk harming a licensee’s perceived safety culture. All of this is undertaken knowing there is a substantial risk the ONR will reject the analysis.

Nuclear regulation does not use a default method to calculate gross disproportion: ratios vary by context, with the Sizewell B Inquiry using three for worker risk, two for low public risk, and ten for high-consequence risks. Quantified costs are often weighed against qualitative safety claims and wider benefits are excluded. Unlike the US’s Nuclear Regulatory Commission and the National Institute for Health and Care Excellence (NICE), the ONR does not use a standard “value of life” estimate to quantify costs.

To be ALARP, dutyholders must use Relevant Good Practice (RGP): RGP covers guidance, standards, and practices with a safe operating record, but there is no clear hierarchy. This ambiguity leads to frequent redesigns of proven foreign reactors. For the EPR, it led to thousands of UK-specific changes, including a separate analogue control system and two additional HVAC floors. More steel, concrete, and cabling were required as a result.

RGP has forced expensive modifications on modern designs with little-or-no safety benefit: to meet ONR requirements, GE Hitachi was required to modify the ABWR design by installing bulky HEPA filters on each HVAC duct even though they reduced radiation exposure by only 0.0001 mSv—equivalent to eating a single banana.

Inflexible planning and permitting processes hold back green nuclear power

Nuclear is low-carbon across its life cycle: Producing zero in operation and, even counting mining and construction, on average fewer carbon emissions than wind and solar. Nuclear, wind, and solar all emit about one-twentieth of the carbon emissions of gas per watt generated. Coal emits sixty times more carbon per watt generated than nuclear.

Nuclear’s fuel density keeps waste tiny: A uranium pellet can power a home for two years and a lifetime household’s electricity use in Britain would produce less than a shot glass worth of waste. All of the UK’s nuclear waste to date and all the waste for the next 100 years would fit in a building the size of Wembley.

Nuclear minimises land use and can support wildlife: To match Hinkley Point C’s annual output would require a solar farm the size of the Isle of Wight or an onshore wind farm the size of the New Forest. Far from harming wildlife, many UK nuclear sites act as quiet nature reserves because access is restricted and the land is left largely untouched. At Dungeness, for example, black redstarts and peregrine falcons nest on the station.

Environmental Impact and Habitats Regulation Assessments extend pre-application timelines: To secure consent, 31,401 pages of environmental documentation was produced for Hinkley Point C. When you factor in supporting documentation, it reached 80,000 pages total for Sizewell C. Many ecology surveys can only run at special times (for birds, at least one breeding and one winter season), and marine surveys often need multi-year time series—such as three years of fish-egg and larval (ichthyoplankton) data—to capture natural variation.

Post-Consent permitting multiplies paperwork and litigation risk: Sizewell C requires over 160 overlapping permits, and at Hinkley Point C an ultimately unsuccessful habitat-related challenge over mud disposal still delayed essential dredging by one year.

Habitat protections lead to costly, site-specific interventions and can lead to projects being refused altogether: Hinkley Point C’s proposed acoustic fish deterrent sat on top of a fish-return system and initially required risky diver maintenance, while Wylfa Newydd faced refusal in part due to uncertainty about Arctic terns returning.

Consultation obligations add time and embed caution: Sizewell C ran seven public consultations, and developers must consult numerous statutory bodies whose responses often miss the three-week response target, with Natural England the worst offender.

Judicial Review increases costs through delay: in the last twelve years there have been seven legal challenges across Hinkley Point C and Sizewell C, causing 866 days of delay at Hinkley and 827 days at Sizewell so far, with one coastal-defences case still ongoing.

Planning deters efficiency-improving design changes: Hinkley Point C’s switch from a wet to a dry spent-fuel store required new consent from the Environment Agency and a planning variation. The extra permissions add time and cost, discouraging late-stage changes unless savings are certain.

Planning uncertainty undermines fleet and SMR economics: investors are unlikely to order a fleet if projects two to five may not be approved. Long and expensive processes create fixed costs. Large sites can spread them over many megawatts. SMRs cannot. New SMR entrants are left carrying heavy pre-revenue costs.

SMRs could access alternative routes of finance from green and AI investors, but are held back by access to the grid

SMRs are smaller and different: factory-built modules are assembled on site and sited on a plot no larger than a football pitch. Typical unit sizes are 470 MW for the Rolls-Royce SMR and 20 MW for Last Energy, which suits repeatable fleet builds.

SMRs can be financed privately unlike megaprojects: large plants need state backing and have never been privately deployed, but lower per-unit costs make SMRs viable through Contracts for Difference or power purchase agreements with industrial customers.

AI data centres are creating firm demand: DSIT targets 6 GW of AI-ready capacity by 2030 and 11 GW by 2035. Tech majors are already signing nuclear-linked deals overseas, including Amazon’s private wire to the Susquehanna nuclear station, Microsoft’s offtake backing Constellation’s $1.6 billion restart at Three Mile Island Unit 1, and Meta’s 650 MW of long-term PPAs with AES.

Britain lacks grid capacity where it is needed: connection waits can exceed a decade and most new capacity is earmarked for wind and solar. A 75 MW AI data centre could face an annual electricity bill of £120 million, so PPAs and private wires are attractive workarounds even though sites still need grid access for stability and outages.

Nuclear is excluded from cheaper green finance and clean-power certificates: green gilts and NS&I Green Savings Bonds are open to decarbonisation projects but the Green Financing Framework omits civil nuclear, and Renewable Energy Guarantees of Origin (REGO) certificates—used to claim clean electricity—exclude nuclear and ignore half-hourly grid mixes, allowing “renewable” claims during wind and solar lulls.