Why Britain is building the world’s most expensive nuclear plant
For over 20 years, Britain effectively gave up on building new nuclear power stations. But that’s changed now Hinkley Point C in Somerset is under construction. When completed it will provide around 7 per cent of the UK’s electricity.
Hinkley Point C is set to be the most expensive nuclear power station ever built. In fact, it is more than four times more expensive on a pound-for-megawatt basis than the average nuclear power plant built in South Korea. Even Flamanville 3, a French plant that uses the same reactor (EPR-1750) and built by the same company (EDF), is set to cost at least 25 per cent less.
Why has Hinkley Point C been so expensive and how do we make new nuclear power in Britain cheaper? Britain Remade, the campaign group I work for, travelled to Somerset to visit Hinkley Point C a few months back to try to find a satisfactory answer to those two questions. (We also really wanted to see Big Carl, the world’s largest crane.)
Hinkley Point C is, by some distance, Europe’s largest construction project. We arrived at the site through what I am reliably told is Europe’s second largest bus station. As we entered, we walked past a big rock with ‘safety is our overriding priority’ engraved on it. Nuclear, of course, is the safest form of power generation there is.
In fact, it’s probably too safe. Nuclear expert Jack Devanney argues safety comes at a cost. There are some nuclear safety measures that cost between one million and several billion pounds per life saved. By pushing up the cost of building and running nuclear power stations, the regulations mean that we end up using much more dangerous forms of energy generation like gas or coal, which are not only bad for the climate but also kill millions with air pollution.
Most popular
Britain is living beyond its means
Still, on a building site, safety is a sensible priority and some of the safety measures employed at Hinkley Point C could (and should) be rolled out much more widely. Take the constant signs telling you not to look at your phone while walking – something for TfL to look into perhaps.
Practice makes perfect. The more times you do something the cheaper it is to do. Nuclear is no exception to this rule. For example, EDF told us that welding on the second reactor at Hinkley Point C is being done four times faster than on the first one. In fact, the second reactor is set to be around 30 per cent cheaper than the first.
It shouldn’t really surprise us that Hinkley Point C has been so expensive. It is, after all, the first nuclear power station Britain has built in almost three decades. South Korea, who have kept costs low, have done so by building fleets. There’s really no shortcut to an experienced skilled workforce. Some mistakes need to be made so we can learn from them.
Hinkley Point C is on paper the 4th EPR-1750 plant being built. In theory, we should be able to learn lessons from EPR-1750 projects from around the world. In reality, it’s more complicated.
The UK and France (and for that matter, Finland) have different approaches to nuclear regulation. On the continent and across the pond, regulators tend to set out clear rules on the design of nuclear reactors. By contrast, the UK takes an outcome-based approach to regulation. In essence, we don’t care how you make your reactor safe as long as it is safe. Our approach is more flexible, which is good for innovation and why small modular reactor developers see the UK as a great destination, but it creates problems too.
EDF were forced to prove that their reactor design, already approved elsewhere, could meet the UK’s goals on safety. The UK’s Office for Nuclear Regulation, it turns out, often takes a different view to its French counterpart on whether a design is up to standard.
In total, meeting the requirements of the Office for Nuclear Regulation led to a staggering 7,000 design modifications. The result is Hinkley Point C will use 25 per cent more concrete and 35 per cent more steel than it would otherwise.
Hinkley Point C is so different from its counterparts in Flamanville and Olkiluoto that it shouldn’t be thought of as just another EPR-1750, but rather the first ‘UK EPR’. A brand new design.
The consequence of that is many of the lessons EDF have learned the hard way on Flamanville 3 can’t always be applied to Hinkley Point C. We get to make entirely new mistakes of our own.
Nuclear power is the densest form of energy there is. In fact, you could bury all the nuclear fuel used in US power stations over the last 60 years in a ten-metre-deep football field. To produce as much electricity with solar as Hinkley Point C would use a plot of land almost 50 times bigger than Hyde Park.
By taking up a smaller footprint than any other form of energy to produce a given megawatt, nuclear power should be a conservationist’s ideal source of power. This, by the way, was a key argument made by eco-modernists like Mark Lynas, Stewart Brand and the people at the Breakthrough Institute.
When you take into account that the energy we need to keep the lights on has to be produced somehow, a new nuclear power station will almost certainly have a positive impact on nature and biodiversity (even putting aside the climate benefits). Yet, our planning system doesn’t take counterfactuals like this into account.
To win planning approval, Hinkley Point C was forced to produce a mammoth 31,401 page environmental impact assessment – and Sizewell C’s was even longer.
A key issue that the 31,401 pages covered was the impact Hinkley Point C would have on marine life. To generate power, pressurised water reactors like Hinkley Point C use a fission reaction to heat water in order to generate steam to power a turbine. To keep generating power that steam needs to be cooled and that’s typically done by pumping seawater into a condenser. By the way, to bust a common myth, this seawater never actually enters the reactor.
When you pump in the water, fish occasionally get caught and, er, cooked. EDF has spent millions to avoid that as much as possible. Unlike in Flamanville, Hinkley Point C’s water intake has been re-designed to take in water at a lower velocity to give fish a better chance of swimming away. On top of that, they also have a fish recovery and return system.
Yet that’s not all, there’s also provisions in their Development Consent Order to install an acoustic fish deterrent. In effect, to save around 45 tonnes of fish (less than a small fishing vessel takes in each year) EDF must install 288 speakers to produce a jumbo-jet level racket playing sounds that fish don’t like – a recording of Baby Shark perhaps?
This is not only an expensive way to save a tiny number of fish (only 112 of the fish deterred are protected), but also potentially dangerous to human life. The water around Hinkley Point C is murky and EDF contend that the only way to install it is to use trained divers. Diving in this water is not without risk and there’s a real possibility that someone might drown in the process of installing (and subsequently maintaining) it.
EDF are trying to persuade regulators to remove this burdensome requirement and proposing to create a biodiversity enhancing 840-acre saltmarsh nearby instead. Yet, the process of getting out of the requirement has been extremely long. It started in 2019 and they’re not expecting a decision until 2025.
EDF have spent hundreds of millions of pounds on protecting fish and will need to pay potentially hundreds of millions more if they can agree on the salt marsh. If they can’t agree, then there’s a risk the powerplant will never be allowed to be switched on.
And remember, the biggest threat to biodiversity is climate change. If we can’t switch on plants like Hinkley Point C then it means relying on climate change causing fossil fuels for longer.
To install those cooling water intakes means dredging a lot of mud. And that mud has to be dumped somewhere. EDF originally wanted to use a site near Cardiff. Yet bogus concerns around radiation (the mud was next to two other nuclear power plants) led to massive delays to work starting.
What should have been a routine process was anything but. EDF gained a licence to dump the mud in Cardiff Grounds dumping site in 2013, but Welsh politicians and a petition led to a drawn-out dispute.
In some cases, politicians opposed to the dumping made claims that could politely only be described as utter nonsense. Welsh Assembly Member, Neil McEvoy, asserted without evidence that ‘no dose of non-naturally occurring radiation is safe.’
For context, the dose of radiation received by a member of the public from the dump would be just 0.002 millisieverts. That’s less than a 25th of the radioactive exposure that you get from a single transatlantic flight and roughly equivalent to eating 40 extra bananas in a year. In other words, it’s next to nothing.
In fact, it is so un-radioactive that it cannot legally be described as ‘radioactive’.
When they started work the plan was to dump it all in Cardiff Grounds, but when space opened up at a nearby dumping site in Portishead, they decided to dump the mud from the second phase of dredging there.
Anti-nuclear campaigners, including the keyboardist from the Super Furry Animals (a man known for Fuzzy Logic), used it as an opportunity to take EDF to court on the grounds they needed to do another Environmental Impact Assessment. Dealing with their legal challenge took six months – it was eventually dismissed at the High Court in March 2022.
But the delay to the project went beyond those six months. Construction is reliant on weather windows. In some cases a few months delay can push a project back by a year. That’s what happened here. EDF reckon that pushing it back by a year made the overall project around £150 million more expensive. All of that expense over some mud that will expose a member of the public to less radiation than they’d get from eating 20g of Brazil nuts.
But there’s more. EDF were also forced to pipe naturally draining water far out to sea because the water was high in Zinc. Not only was this expensive, costing tens of millions, but as the water was naturally draining it provided no environmental benefit. Oh, and they had to pay rent to the very agencies that told them to do this to run the pipe across their land.
It is now 11 years since Hinkley Point C’s planning application was accepted. In that time, a lot has been learnt about the best way to store spent fuel (otherwise known as nuclear waste). EDF initially planned to build a ‘wet store’ because it’s easier to inspect, but they’ve now decided to switch to a ‘dry store’ because experience shows they’re simpler to run. In terms of safety, there’s no difference and the Environment Agency has granted them the relevant permit to make the change. However, switching systems will mean the building the waste is actually stored in will need to be about 79m longer.
This, it turns out, meets the threshold to be considered a ‘material change’ and as a result EDF will have to apply for special permission from the planning inspectorate. As a result, it means more delay and more legal fees for a relatively minor change.
This is part of a wider problem. By making it hard to make changes, Britain builds new infrastructure slower and less efficiently than it should. Most major infrastructure builders when asked could list examples of beneficial changes they didn’t make due to the cost of going through planning again, according to research from the National Infrastructure Planning Association. Even modest and clearly beneficial changes like the Lower Thames Crossing using one tunnel boring machine instead of two mean that lawyers and environmental experts are drafted in to prove the change is ‘non-material’.
Building Britain’s first nuclear power station in almost three decades was never going to be easy, but boy have we made it harder than it needs to be.
The sheer number of changes required by our nuclear regulator not only meant using more concrete and more steel, they also limited our ability to learn from Finland and France’s experience. Delays caused by a drawn-out process of permitting and planning haven’t helped either. But, even within a single project, productivity is increasing as engineers learn from experience.
So, what needs to be done then to cut costs? I can think of a few things.
Replication and learning are key to improving productivity, so reforming our nuclear regulator should be a priority. Where designs have been approved elsewhere by respected regulators in places like France or the US, we should defer to their judgement and not insist on as many changes. Just as our Medicines and Healthcare products Regulatory Agency now grants automatic approval to drugs approved elsewhere by respected regulators, the Office for Nuclear Regulation should automatically approve reactor designs certified as safe elsewhere. It would mean that if South Korea’s KEPCO wanted to build their tried-and-tested low-cost reactor here then they wouldn’t have to make thousands of changes and build an almost bespoke design just for us.
But wider changes to planning policy are needed. Environmental Impact Assessments and Habitats Regulations should be reformed too so developers don’t need to engage in extremely costly mitigation measures and produce tens of thousands of pages of documentation to build a new nuclear power station. It also needs to be made much easier to make changes as a project develops.
And there's a deeper question that needs to be asked about the way we regulate nuclear energy. Have we gone too far in the pursuit of safety? Do we pay more to save a hypothetical life in nuclear energy than we do in any other areas of life, such as on the road or when we regulate other forms of energy?
On all of this there’s some hope. The day we visited Hinkley Point C, the government published its civil nuclear roadmap. It includes plans to allow small modular reactors to be built ‘anywhere’ that meets certain criteria (not just designated sites), a commitment to more international regulatory cooperation, and a new smarter regulation challenge.
Could it go further? By all means, but it’s a good start.
This article first appeared in Sam Dumitriu’s Notes on Growth Substack.