European nuclear could be cheap

A UK taskforce shows how to cut cost without touching safety

The primary argument of the more sophisticated opponents of nuclear power is its cost. They recognize that nuclear power is safe, and that it is clean. But it is much more expensive than solar, taking billions of public subsidies to finish a plant. We should let the market decide, they say, and nuclear power is not competitive with wind and solar energy.

But nuclear power is more competitive than it seems. The truth is that nuclear costs, like so many things in Europe, are a result of policy choices. The last plants built in Europe took nearly two decades to complete. In South Korea, it takes an average of five years. Across the continent, nuclear power is stuck in a doom loop. Safety rules make projects slow and expensive. Because they are slow and expensive, few get built. This breaks the positive learning curves seen in earlier programs. France's nuclear expansion in the 1970s and 80s, for example, showed that as more reactors of a standard design were built, costs steadily decreased—by as much as 23% for a given series. With so few projects today, the workforce and supply chains lose the skills to build them well.

Two of Europe’s most recent projects are Finland’s Olkiluoto 3 and France’s Flamanville 3 reactors. Both were treated as one-off experiments. Olkiluoto 3 was finished about 14 years late in 2023 and cost nearly four times its original budget, rising from €3 billion to around €11 billion. Flamanville 3 connected to the grid in December 2024, 12 years behind schedule; its cost had quadrupled from €3.3 billion to over €13 billion in 2015 euros according to EDF (France Cour des Comptes reports 24 bn in current euros including financing). On a consistent 2010-$ per-kWe basis, Olkiluoto-3 cost about $6,600/kWe and Flamanville-3 about $9,800/kWe ex-financing (≈$14,100/kWe including financing) — two to four times the standardized French builds of the 1970s–80s.¹

Making nuclear truly competitive is critical, because a fair market comparison must also account for reliability. On that front, wind and solar cannot be the whole solution to decarbonization. As we have argued before, a dark, calm week in winter—a ‘Dunkelflaute’— can stall renewable output across much of the continent. Batteries cannot fill this gap; Germany’s planned storage for 2030, for example, would cover less than one percent of the energy needed for such an event. Nor can interconnections do it, given the high correlation between weather patterns across the continent.

The UK case

To understand why, we can look at Britain, where a government task force has made proposals that could lead to faster, cheaper plants. The problems it describes are common across Europe. The primary culprit behind expensive nuclear is the “As Low As Reasonably Practicable” (ALARP) principle, which EURATOM, the EU’s nuclear body, adopted as a standard in 1980. This approach requires extra measures unless their cost is grossly disproportionate to the safety gain. While this principle comes from a seemingly laudable goal of maximizing safety, the taskforce finds it drives delay and expense without commensurate safety gains. A profitable operator is seen as leaving cash on the table that could have been used to mitigate radiation, no matter how tiny the benefit.

The results are absurd. One UK plant builder had to redesign a reactor — already used safely across Japan—to cut radiation emissions by 0.0001 millisieverts (mSv). This is the dose you get from eating a single banana. The UK's public dose limit from nuclear plants is 1 mSv/year. The average person already receives about 2.7 mSv/year from natural sources and medical scans.

Right now, none of the (safe!) designs used abroad can be used to avoid requirements in the UK. Even when a component or design has operated safely for years abroad, UK regulators commonly require new testing and a fresh decision, which “effectively leads to double regulation.”

By contrast, countries that build nuclear energy efficiently, like South Korea, rely on a fleet approach, treating the tenth reactor like the tenth plane off a production line.

While Europe struggles with two-decade projects, South Korea builds identical plants on predictable schedules, finishing them in under five years. Their reactors run more reliably than those in the US, France, or Japan — delivering an average 96.5% of full capacity. When South Korea built four of its reactors in the UAE, the project was delivered roughly on time (2020-2024) and on budget, in sharp contrast to the European experience.

The cost of delay

The problem with European nuclear regulators is that they measure the cost of not regulating, but not the opportunity cost of regulating. A delay on a major nuclear plant means years of burning more fossil fuels, higher electricity prices, and greater import dependency. These are real, quantifiable costs to society and the economy, yet they are largely absent from the regulatory equation. This extends to the workforce, where we're likely to see a "cliff-edge drop in capability due to retirement" as long project timelines deter a new generation of engineers.

As Britain’s report puts it, “indirect costs are often not adequately considered by regulators in their analysis of what measures are proportionate to reduce the risk".

Olkiluoto 3 now produces 14% of Finland's electricity. Had it opened on schedule in 2009, Finland would have avoided 14 years of extra carbon emissions and dependence on foreign fossil fuels. Regulators should have to quantify those foregone benefits when deciding if another micro-reduction in dose is “reasonable”.

The path forward is to fix this regulatory dysfunction. Today, regulators face no penalty for delays that have wrecked European electricity markets. Governments must change these incentives. This means ending the absurd practice of re-evaluating reactor designs that have run safely for decades in allied nations like the US or South Korea. It means committing to a fleet approach, building the tenth reactor like the tenth car off an assembly line. These changes would defang nuclear’s smartest critics: clean baseload power does not need to take decades and costs tens of billions. Europe could have abundant, cheap, clean electricity if its regulators let it.

1. To obtain the estimates, convert euro project costs to 2010 euros using euro-area inflation, then convert to 2010 dollars at the 2010 average FX (≈$1.327/€), and divide by net capacity. For OL3, taking ≈€11bn (current €) and deflating to 2010 € gives ≈€8.0bn; per-kWe ≈€5,000 → $6,600/kWe (2010-$). For Flamanville-3, using €13.2bn in 2015 € ex-financing and deflating to 2010 € gives ≈€12.1bn; per-kWe ≈€7,450 → $9,800/kWe (2010-$). If instead one uses the Cour des comptes total cost (≈€23.7bn in 2024 €), deflating to 2010 € yields ≈€17.3bn; per-kWe ≈€10,600 → $14,100/kWe (2010-$). These figures are directly comparable to the historical French fleet costs in the Figure, which average ≈$1,600–$1,900/kWe in 2010 dollars.