How red tape holds back nuclear power in Britain

And how to fix it

The impressive UK Day One think tank recently surveyed 43 leading British experts on economic growth (plus myself). One, perhaps surprising, finding from the research was that the experts surveyed listed ‘nuclear power’ as the number two priority area that the government ought to invest its time and political capital in.

Why did experts on growth put ‘increasing nuclear energy production’ so high? One reason might be the UK’s high industrial energy costs. British businesses pay more for power than anyone else in the world. And even as our household appliances become more efficient, for the first time in half a century Britain is set to need a major increase in the amount of electricity it produces for two reasons. 

First, hitting net zero will require us to electrify nearly everything from how we heat our homes to how we get around town. 

Second, while AI has the potential to transform our economy, it is a power-hungry technology. In Ireland, over a fifth of electricity demand comes from data centres already. Britain is pushing hard on renewables, but data centres create new challenges. Data centres typically are limited in their ability to ‘flex’ their demand in response to high prices – you want your data centre running 24/7.  As a result, the need for backup flexible generation increases. Either that’s gas (expensive and dirty) or it’s something else. At the moment, the best candidates we have for something else are either expensive, dirty, or land-hungry. On top of that all, there’s the need to build the grid out even further.

If Britain is to avoid squandering its natural advantages in AI, nuclear will need to be part of the equation. That’s not just my view – it is increasingly the view of the AI industry.

• Microsoft recently signed a twenty year deal to reopen Three Mile Island in 2028 to power its data centre with a power purchase agreement.

• Google recently announced a deal with Kairos Energy to build Small Modular Reactors (SMRs) for the same purpose.

• Amazon too has signed a number of deals with SMR developers and existing nuclear facilities, including investing $500m in X-Energy’s Series C financing round.

• Open AI’s Sam Altman has gone one step further and has his own SMR startup, Oklo.

What’s holding back nuclear power in Britain?

Harnessing the transformational technology of the 21st century will mean going hell-for-leather in building new nuclear power stations. There’s just one problem: Britain is the most expensive place to build a nuclear power station in the world. Hinkley Point C, currently under construction, will cost 6 times more to build per megawatt than its South Korean equivalent. France and Finland, which use the same reactor design as us, build for about 50% less than us per megawatt.

Why is nuclear so expensive to build in the UK? There’s not a single cause, but our planning system doesn’t help. Developers EDF were forced to produce a more than 30,000 page long environmental impact assessment to win planning permission for Hinkley Point C. They are, to this day, still fighting with the Environment Agency and nearby councils over whether or not to install a ‘fish disco’ to save the equivalent of a trawler’s worth of fish each year. 

Although Hinkley Point C uses the same EPR design as in Finland and France, the UK’s nuclear regulator insisted on 7,000 changes such as the requirement of having both an analog and digital control centre. France’s Flamanville only has the latter. Hinkley Point C will use 25% more concrete and 35% more steel than other EPRs as a result.

All of this makes the kind of fleet-based approach that works in South Korea (and worked in France and Britain historically) next-to-impossible. You do not get the benefits of learning-by-doing or supply chain investment, if there’s no certainty around what you are building and when you will be building it, if at all.

Beyond planning, there is the problem that our regulators treat nuclear differently to any other technology when it comes to safety.

The underlying principle governing the regulation of radiation in Britain is ‘As Low as Reasonably Practicable’ or ALARP. In effect, it means that whenever a productivity gain in nuclear construction is discovered, it is seen as an opportunity to make the safest way to produce electricity even safer. The Office for Nuclear Regulation’s starting point for whether or not nuclear safety measures are ‘grossly disproportionate’ is if costs outweigh benefits by a factor of ten. 

Not only does this make nuclear uncompetitive, it also makes the public less safe because instead of using safe nuclear, we burn dirty gas (or until very recently, coal) where similarly costly regulations aren’t being imposed.

What would it take then to make nuclear cheaper in Britain then? Well, a lot. Not only do we need to make our planning system faster, less onerous, and more certain, we also need to radically adapt our approach to regulating nuclear power and forensically review thousands of regulatory choices where the wrong trade-off between safety and cost was made.

All of this can and should be done, but it won’t exactly be easy. There are some things, however, that hold nuclear back and could be fixed with ease. 

Picking the Lowest Hanging Fruit

Let’s take the issue of ‘Regulatory Justification’. The ultimate example of red tape for red tape’s sake. 

As of 2004, any practice that may expose the public to ionising radiation must be ‘justified’. That is to say, the benefits must outweigh the costs of radiation exposure. This, in itself, isn’t a silly idea. We shouldn’t expose people to radiation without a good reason. It is probably a good idea to use x-ray scanners in prisons to prevent visitors smuggling in contraband (like weapons). It probably isn’t a good idea to put an x-ray scanner in every shop entrance to prevent shoplifting.

When the regulations were put in place, all existing practices that exposed the public to ionising radiation were grandfathered in. We didn’t temporarily ban the NHS from using X-rays or switch off our nuclear power stations until they submitted the relevant paperwork.

But, there was a problem with the grandfathering. Rather than accepting the broad principle that some exposure to ionising radiation is justified when it provides reliable low-carbon electricity, Britain’s regulators decided that all new nuclear technologies needed to themselves be justified. That’s despite the fact that new (or not so new now) nuclear technologies produce way more power and expose the public to even less (than the already miniscule amounts of) radiation that existing power stations do.

Even though the regulations are intended to deal with broad “practices”, the rules have been interpreted to mean that every new nuclear reactor design counts as a new practice requiring regulatory justification. 

What does this mean in practice? Every single new nuclear reactor design has to apply to have their ‘new practice’ approved as a justified use of ionising radiation.

Applying isn’t cheap. Industry estimates suggest that putting together the application alone can cost millions of pounds. The real cost, however, isn’t in cash, but time. Not just the time it takes to put together an application and provide the requisite evidence, but also the time it takes to receive a decision. Once an application (which itself takes many months to complete) has been submitted, it can take over two years to receive a decision.

It also creates a capacity crunch at our regulators. For example, Andrew Bennett of VC firm Form Ventures makes a compelling case for better funding our regulators because at current resourcing they can’t respond effectively to new technologies. Yet at the moment, regulators from the Office for Nuclear Regulation to the Food Standards Agency waste their time on unnecessary paperwork. Eliminating duplicative tasks like regulatory justification, allows us to expand capacity at our regulators without adding to the budget black hole. 

And let’s be clear it is duplicative. It is certainly not the only point in the regulatory process that considers the costs and benefits of using a specific nuclear technology. This is also covered by the Generic Design Assessment, environmental permitting, the planning process, and the Nuclear Site Licensing process. Many of the latter can’t be obtained until a regulatory justification application has been successful.

It didn’t have to be this way.

What’s frustrating is that this completely redundant regulatory barrier almost wasn’t. In 2008, the Nuclear Industry Association applied for regulatory justification for the “generation of electricity from nuclear energy using oxide fuel of low enrichment in fissile content in light water cooled, water moderated thermal reactors using evolutionary designs.” This would, in effect, have granted regulatory justification to most modern nuclear reactor designs, including all the entrants into the Great British Nuclear SMR competition.

Back when he was Energy Secretary for the first time, Ed Miliband had the opportunity to approve this. Unfortunately, he didn’t. Instead, it was instead decided that they resubmit the application on a reactor-design specific basis. Some did (EDF’s EPR and Westinghouse’s AP-1000), others dropped out.

It isn’t clear why this choice was made. I suspect there may have been fears that the original application was too broad and open to judicial challenge. Or it may have been a classic case of nuclear power being treated with a disproportionate degree of scrutiny.

The NIA’s initial application wasn’t some aberration either. Most applications for regulatory justification address classes of use e.g. “Use of ionising radiation (X-ray) imaging to support scientific age assessments of age-disputed persons”. New models of X-Rays used in hospitals do not require individual approval – rather the entire use class is approved.

One way to solve this problem and eliminate this pointless exercise in red tape creation would be to use existing powers to approve all new nuclear power station designs as an “existing class or type of practice.”

For example, the first application for regulatory justification was for a safety toaster with a built-in smoke detector. It was ultimately determined to be approved as an existing use class ‘smoke detectors’.

Strangely, despite having his old job back Ed Miliband is powerless to rewrite this wrong. Due to the principle of ‘functional separation’ (the idea that departments with an interest in a technology aren’t the ones who decide whether it's justified), the Energy Secretary no longer makes the call on whether nuclear power is a justified use of ionising radiation. Instead, the power rests in another department, DEFRA.

Environment Secretary Steve Reed could eliminate a significant bureaucratic hurdle for nuclear development in Britain with the stroke of a pen. Under the Regulation 12 of Justification of Practices Involving Ionising Radiation Regulations 2004, he can establish that “the generation of electricity from nuclear energy using oxide fuel of low enrichment in fissile content in light water cooled, water moderated thermal reactors” as an existing practice. This would cover a wide range of reactor designs, including all medium-term proposals in Britain, but it wouldn’t solve the problem forever if a new generation of nuclear power stations with a radically different design comes forward.

There is legal precedent, no requirement to hold a public consultation, and no requirement for any application from the nuclear industry. All that’s needed is a short targeted four-week consultation with a number of public bodies. This could be fixed by Christmas.

Why this matters

There is no doubt that Britain’s state believes the risks from generating energy from nuclear power are massively outweighed by the benefits. Not only have they granted permission for four EPRs (two each at Hinkley Point C and Sizewell C), they’ve designated Nuclear power as being a ‘critical national priority’, and are running a competition to fund the development of SMRs.

Yet only one of the SMR technologies in the competition (Rolls-Royce SMR) has submitted an application for regulatory justification. On the one hand, we are subsidising new SMR designs in the hope of eventually committing to a fleet. On the other, we are forcing their technology vendors to spend millions of pounds and thousands of man hours to navigate a regulatory process that is duplicated elsewhere to establish a fact that most branches of the state have already accepted. 

Britain can succeed on AI and attract billions of pounds of investment into data centres, if we can bring down nuclear costs. This won’t be easy by any means, but we can make a start by removing pointless hurdles to new reactor designs like ‘regulatory justification’.