Nuclear power did not stall because the underlying science proved inadequate. It stalled because the technology arrived carrying the image of the atomic bomb, suffered several spectacular accidents, became trapped in increasingly complicated regulatory and planning systems, and proved exceptionally difficult to finance in liberalised electricity markets.
France demonstrated what nuclear power could achieve when a country standardised its reactors and treated electricity as strategic national infrastructure. China and Russia have preserved much of that state-directed model. Britain retained considerable scientific and engineering expertise but lost continuity, supply chains and the institutional ability to build nuclear plants quickly and economically.
Small modular reactors may provide a route back. But we should describe them honestly: they are a promising industrial proposition, not yet a commercially proven Western energy industry.
The nuclear age began with a contradiction.
Human beings unlocked the enormous energy contained within the atomic nucleus, but the world's first encounter with that achievement was not an abundance of inexpensive electricity. It was Hiroshima and Nagasaki.
That beginning has shaped public attitudes ever since. Nuclear power became associated not simply with complicated machinery but with radiation, contamination, mutation and annihilation. No coal-fired or gas-fired power station has ever had to overcome an equivalent psychological inheritance.
The great historical question is therefore irresistible. If nuclear fission had first entered public consciousness as a source of heat and electricity rather than as a weapon, would the world have built far more nuclear power stations — and burned far less coal, oil and gas?
Perhaps we would now be in a much better position regarding climate change and global warming. Perhaps we would have consumed far less fossil fuel to generate our energy. We cannot prove that counterfactual, but it raises an important question: why has a technology capable of producing large quantities of dependable, low-carbon electricity remained so difficult for Western democracies to build?
The Bomb Came First
The Manhattan Project did not discover nuclear fission, but it converted the science into two workable weapons.
Los Alamos developed two fundamentally different bomb designs. Little Boy, dropped on Hiroshima, used highly enriched uranium in a relatively simple gun-type assembly. Fat Man, dropped on Nagasaki, used plutonium surrounded by carefully timed conventional explosives that compressed it into a critical mass. This was the technically more demanding implosion design.
The Trinity test in New Mexico tested the plutonium implosion design before it was used over Nagasaki. A proposed plutonium gun-type weapon had been abandoned because contamination by plutonium-240 made premature detonation too likely. This forced the Los Alamos scientists to solve the much harder implosion problem.[1]
This history matters because military and civil nuclear technology have remained psychologically entangled. A nuclear power reactor cannot explode like an atomic bomb, but that distinction has never been as emotionally powerful as the images of mushroom clouds.
The bomb does not, however, explain everything. Substantial civil nuclear programmes were subsequently built in the United States, France, Britain, Canada, the Soviet Union and Japan. Nuclear construction slowed much later.
The fuller explanation includes serious accidents, constantly changing regulations, repeated alterations to reactor designs, weak project management, cheap natural gas, high financing costs, the loss of experienced supply chains and political decisions that can be reversed several times during the extraordinarily long life of a nuclear asset.
The bomb created nuclear power's image problem. It did not, by itself, cause the West's construction problem.
What Nuclear Power Actually Offers
A nuclear power station is almost the mirror image of a gas-fired plant.
Most nuclear expenditure occurs before the first unit of electricity is sold. The plant can then operate for many decades using a remarkably small physical quantity of fuel. A gas-fired plant is comparatively cheap and quick to build, but it exposes consumers to continuing fuel costs and movements in international gas prices.
Nuclear power offers dependable electricity in almost every weather condition, very low operational carbon emissions, high energy density, comparatively modest land requirements and much less exposure to fossil-fuel price shocks. Nuclear reactors can also provide industrial heat, district heating and energy for hydrogen production.
The disadvantages are equally real. Nuclear plants require enormous up-front investment. Planning and construction can take many years. Delays accumulate financing costs before any electricity can be sold. Radioactive waste requires competent institutions over exceptionally long periods, while decommissioning must be considered from the beginning rather than left to future taxpayers.
There are also security and proliferation concerns and the possibility of a very rare — but socially, politically and financially devastating — accident.
The International Energy Agency says nuclear projects are difficult to finance because of their scale, capital intensity, technical complexity and long construction periods. Government involvement, predictable future revenues and a low cost of capital are frequently indispensable.[2]
France: The Strongest Western Case
France did not build its nuclear system as a collection of unrelated individual projects. It created a national industrial programme constructed around standardised reactor families, a state-backed utility, long-term planning and a domestic supply chain.
The results remain striking.
France generated 539 terawatt hours of electricity in 2024. Nuclear power supplied 361.7 terawatt hours — approximately 67% of the total. Nuclear and renewable generation together made about 95% of French electricity low-carbon. Fossil-fuel generation fell to only 20 terawatt hours.
France also exported a record net 89 terawatt hours of electricity, with an estimated export value of €5 billion. The carbon intensity of French electricity was just 21.7 grams of carbon dioxide equivalent per kilowatt hour.[3]
France demonstrates the economic, environmental and strategic benefits of nuclear power. But it also provides a warning.
Corrosion discoveries, maintenance backlogs and other outages contributed to exceptionally poor nuclear availability in 2022, when France became a net electricity importer for the first time since 1980. An excessively homogeneous fleet can create common problems across several reactors, while postponing essential maintenance eventually catches up with even the most successful programme.
The lesson is not that France's nuclear strategy was mistaken. The lesson is that a nuclear fleet requires continual investment, technical competence and institutional attention. It cannot be constructed and then taken for granted.
Germany's Great Energy Gamble
It is tempting to say that Germany closed its nuclear plants in favour of Russian gas. That is directionally true but too simplistic.
Germany accelerated its nuclear phase-out following the Fukushima accident while simultaneously expanding renewable generation. Coal remained important, and the country was already heavily dependent upon inexpensive Russian gas for industry, heating and some flexible electricity generation.
Russia's invasion of Ukraine then exposed the geopolitical weakness of that dependence.
Germany voluntarily removed dependable, low-carbon generating assets while remaining exposed to coal and imported gas. It discovered, at considerable cost, that energy policy cannot be separated from industrial policy, foreign policy or national security.
The International Energy Agency identifies the simultaneous loss of nuclear power, planned coal closures and the end of Russian gas supplies as central challenges for Germany's energy transition.[4]
Germany's experience should be studied carefully in Britain. Closing existing generating capacity is easy. Replacing its dependable output at an acceptable price is considerably harder.
The Great SMR Illusion
There is now almost constant discussion about small modular reactors. Governments make announcements, companies produce attractive illustrations and investors talk confidently about factory production.
The language often gives the impression that fleets of these reactors already exist.
They do not.
As of 2025–26, the International Atomic Energy Agency identified only two commercial small modular reactor power-station designs in operation.
Russia's Akademik Lomonosov is a floating nuclear plant containing two reactors of approximately 35 megawatts each. It has operated commercially at Pevek in the Russian Arctic since 2020.
China's HTR-PM is a 200-megawatt high-temperature, gas-cooled demonstration plant. It entered commercial operation in December 2023.[5]
Despite all the political announcements, investment presentations and confident predictions, there is still no commercially operating, grid-generating small modular reactor in the Western world.
Canada has moved furthest towards changing that. Ontario Power Generation has begun construction of a GE Hitachi BWRX-300 reactor at Darlington. It represents real progress, but it is not yet producing electricity.[6]
The key promise of an SMR is not simply that it is smaller. It is that major components can be manufactured repeatedly in controlled factory conditions rather than every reactor being constructed as a unique national megaproject.
The theory is straightforward. Complete and license a stable design, manufacture standardised modules, transport them to prepared sites, build the same reactor repeatedly and use the accumulated experience to reduce the time and cost of every subsequent unit.
This is exactly what much of the Western nuclear industry has failed to do with large reactors.
But the SMR proposition contains a serious risk. If governments order too few reactors, developers continually alter their designs and regulators treat every location as a completely new project, an SMR becomes merely a smaller first-of-a-kind nuclear plant. It may then produce more expensive electricity because it has surrendered the economies of scale available to a large reactor without achieving the economies of mass production.
For SMRs to succeed, governments must be prepared to order a series rather than a solitary demonstration plant.
Rolls-Royce and the Western Contenders
Rolls-Royce is proposing a 470-megawatt pressurised-water reactor in Britain. Great British Energy–Nuclear selected the company as its preferred bidder in June 2025 and signed a development contract in April 2026. The Government has allocated approximately £2.6 billion to the programme.
This is a considerable commitment. But an investment decision, site approvals and further regulatory work are still required. The ambition is for the first reactors to supply the grid during the mid-2030s.[7]
Rolls-Royce has some important advantages. Britain already possesses nuclear engineering expertise through its submarine programme. The proposed reactor uses recognisable pressurised-water technology rather than depending upon an entirely new scientific concept. Its size is also large enough to make a meaningful contribution to the grid.
Nevertheless, Rolls-Royce does not yet have an operating SMR. Its success will depend upon whether Britain creates a genuine production programme with sufficient orders to justify factories, training, tooling and a permanent supply chain.
GE Hitachi's 300-megawatt BWRX-300 is probably the most advanced Western commercial project because construction has started at Darlington in Canada. It uses established boiling-water reactor technology in a simplified design.
NuScale's 77-megawatt module has received American regulatory design approval, but its original Utah project was cancelled after projected costs increased. That illustrates the gulf between achieving regulatory approval and delivering commercially affordable electricity.[8]
X-energy is developing the Xe-100, a high-temperature gas-cooled reactor intended initially for an industrial site operated by Dow. TerraPower's Natrium project combines a sodium-cooled reactor with molten-salt heat storage. The stored heat could allow the plant to vary its electricity output and complement intermittent wind and solar generation.
Other serious projects include Holtec's SMR-300, Westinghouse's AP300 and eVinci microreactor, Kairos Power's Hermes test reactor and France's NUWARD programme. South Korea is also developing its SMART and i-SMR technologies.
The Organisation for Economic Co-operation and Development's Nuclear Energy Agency tracks dozens of designs, but only a small number have reached construction. No Western developer has yet established a mature, serially manufactured commercial fleet.[9]
The next decade will reveal whether SMRs represent the rebirth of a manufacturing industry or merely another collection of expensive prototypes.
Russia and China: Building While the West Discusses
China now has approximately 60 operating nuclear power reactors and 37 under construction. Nuclear power still supplies only about 5% of its enormous electricity system, so there remains considerable room for further expansion.[10]
China's advantage is not necessarily that every individual reactor is technically superior. Its advantage is continuity: repeated construction, state-directed finance, an experienced workforce and a supply chain that can expect further orders.
Russia operates approximately 34 conventional reactors in addition to the two units aboard Akademik Lomonosov. Nuclear supplies roughly 19% of Russian electricity, while several conventional and advanced reactors remain under construction.[11]
Akademik Lomonosov supplies electricity and district heating to Pevek. Floating nuclear power is particularly attractive in the Arctic because it can replace ageing coal stations and provide energy to remote ports, communities and resource projects where transporting large quantities of fossil fuel is difficult and expensive.[12]
Russia has also developed reactors for nuclear icebreakers and is constructing additional land-based and floating SMR projects.
China and Russia remain competitive because they have preserved entire nuclear ecosystems: engineering businesses, fuel-cycle facilities, experienced construction teams, state finance, specialist training and continuing order books.
Britain and other Western countries would, however, be understandably reluctant to make critical national infrastructure dependent upon either nation. A nuclear station involves security, software, sensitive operational knowledge, fuel supplies and commercial relationships that may continue for a century.
The cheapest initial construction proposal is not necessarily the lowest strategic cost.
This is precisely why Britain needs to rebuild its own capability — and why Rolls-Royce represents much more than another possible source of electricity. If successful, it could support a strategically important national industry.
Are Naval Reactors Really SMRs?
Submarines and aircraft carriers have used compact nuclear reactors successfully for decades. In a physical sense, these reactors are both small and modular. But they are not SMRs in the normal commercial meaning of the term.
Naval reactors are designed specifically for propulsion. They must operate inside extremely confined spaces, tolerate shock and rapid changes in power, minimise noise and remain close to their crews. They may also use much more highly enriched uranium than commercial reactors.
They operate within closed military organisations with specialised crews, dedicated maintenance facilities, centralised procurement and an exceptionally committed customer. National-security secrecy also protects much of their technology.
It is therefore not quite correct to attribute their success to lighter military regulation. American naval sources say the standards imposed upon naval reactors can be more demanding than civilian requirements because components must withstand battle shock, crew proximity, noise restrictions and frequent changes in power.[13]
The more important difference is unified authority. The military can commit to a design, construct it repeatedly, train people around it and maintain the necessary organisation over many decades.
Naval experience proves that compact reactors can operate reliably. It does not automatically prove that the same reactors can generate commercially competitive electricity. But it does show what continuity, standardisation and a committed customer can achieve.
Nuclear-Powered Commercial Shipping
Civil nuclear propulsion has been demonstrated but has never become a substantial commercial industry.
America built the NS Savannah, Germany produced the Otto Hahn, Japan developed the Mutsu and the Soviet Union constructed the nuclear-powered cargo vessel Sevmorput. Russia has also operated a fleet of nuclear-powered icebreakers.[14]
The economics of icebreakers are unusual. They require enormous amounts of power and must remain at sea for extended periods in remote regions where conventional refuelling can be difficult. Nuclear propulsion therefore provides a clear operational advantage.
Ordinary container ships have not adopted it because reactors are expensive and require specialist crews, security arrangements and maintenance infrastructure. There are also unresolved questions involving insurance, liability and permission for nuclear-powered commercial vessels to enter ports in different countries.
However, the need to decarbonise international shipping may cause the idea to be reconsidered. A reactor could potentially power a large vessel for many years without consuming oil or producing carbon emissions during operation.
Once again, the difficulty is not simply whether the technology works. It is whether the international regulatory, insurance and port system can accommodate it.
Accidents, Perception and Reality
Three Mile Island, Chernobyl and Fukushima are frequently mentioned together, but they were very different events.
Three Mile Island suffered a partial core meltdown in 1979 following a combination of equipment failure, design weaknesses and operator error. The United States Nuclear Regulatory Commission found no detectable health effects among workers or the public. Nevertheless, the accident caused enormous public fear, damaged confidence in the industry and produced sweeping regulatory changes.[15]
Chernobyl was far more serious. A fundamentally unsafe Soviet reactor design, combined with a dangerous test and a deficient safety culture, caused explosions, a fire and a large radioactive release. Workers and emergency responders died, thyroid cancers increased among people exposed as children, and extensive territories were contaminated.[16]
Fukushima followed a massive earthquake and tsunami in 2011. The flooding disabled electrical and cooling systems, causing core damage and radioactive releases. There were no acute radiation deaths. The United Nations Scientific Committee on the Effects of Atomic Radiation says no adverse health effects among residents have been documented as directly attributable to radiation exposure. The evacuation itself, however, produced serious social, psychological and health consequences.[17]
These distinctions matter, but public attitudes towards nuclear power are not formed through statistical comparisons alone.
People assess nuclear risk according to visibility and dread. A nuclear accident is dramatic, concentrated and unforgettable. The damage caused by coal, oil and gas is dispersed through air pollution, industrial disease and climate change. It occurs continually and is therefore less likely to dominate the news.
This creates one of nuclear power's central political problems. Its rare accidents are remembered for generations, while the continuous harm caused by fossil fuels can become almost invisible.
Why Britain's Recent Projects Cost So Much
The recent European Pressurised Reactor programme provides an uncomfortable case study.
Olkiluoto 3 in Finland was completed approximately 14 years late. Flamanville 3 in France was also about 14 years late and cost roughly four times its initial estimate. Hinkley Point C's expected opening moved from 2025 to approximately 2030–32, while its estimated construction cost rose substantially.[18]
These projects were not delayed because nuclear fission suddenly stopped working. They suffered from immature designs, inadequate preparation, supply-chain weaknesses, quality problems and the loss of recent construction experience.
Much of nuclear power's final cost is therefore not uranium, concrete or steel. It is the cost of time.
Every planning delay and construction overrun accumulates interest before the plant can sell any electricity. The longer construction continues, the greater the investor's exposure to political change, inflation, regulation and movements in the electricity market.
This helps to explain why countries with state-backed utilities and lower financing costs may build nuclear plants more successfully than countries attempting to finance them entirely through private capital.
Decommissioning is another legitimate concern, but the figures must be treated carefully. Britain's nuclear provision of more than £100 billion includes historic Magnox stations, Sellafield facilities and liabilities inherited from the early civil and weapons-era programmes. It should not be presented as the normal decommissioning cost of a modern commercial reactor.[19]
Every new plant should make proper provision for decommissioning and waste from the electricity it sells. But the cost of cleaning up Britain's early nuclear history should not be used dishonestly to suggest that every future reactor will create an equivalent liability.
Fusion Is Different — and It Is Not Yet an Energy Industry
Fission splits heavy atomic nuclei. Fusion attempts to combine light nuclei and release energy through the process that powers the Sun.
Researchers have achieved fusion reactions. America's National Ignition Facility has also achieved experimental ignition, in which the fusion energy released from a target exceeded the laser energy delivered directly to it.
That was an important scientific achievement, but it was not a commercial power station. The total facility consumed considerably more energy than the target produced.
ITER, the major international fusion experiment under construction in France, is intended to demonstrate sustained, high-gain fusion plasma. It will not generate electricity for the grid. Its revised programme anticipates research operations during the 2030s.[20]
Later demonstration plants must still solve several formidable problems: producing or breeding sufficient tritium fuel, developing materials capable of surviving intense neutron bombardment, extracting heat continuously and converting the entire process into economically competitive electricity.
It is therefore wrong to say that fusion research has achieved no success. The scientific progress is real. But it is equally wrong to talk as if commercially viable fusion power is just around the corner.
Fusion may eventually transform human civilisation. It cannot substitute for the power stations and electricity networks we need to build during the next two decades.
Standardisation Matters More Than Size
One of the most important lessons from France, China, South Korea and the military is that countries build efficiently when they construct the same design repeatedly.
An SMR programme divided between several competing designs, with only one or two orders for each, will not create a genuine industry. Nor will it produce the promised reductions in cost.
Britain needs to decide whether its objective is to buy a few reactors or create a repeatable nuclear manufacturing system.
That means limiting design changes, preparing sites in advance, coordinating regulators, training sufficient engineers and giving the supply chain confidence that further orders will follow.
It also means accepting that the first reactor may be expensive. The economic argument for modular construction depends upon the fifth, tenth and twentieth reactors being cheaper and quicker than the first.
Financing May Matter More Than Uranium
Nuclear fuel represents a relatively small proportion of the total cost of nuclear electricity. The cost of capital can be far more important.
A plant that takes ten or fifteen years to complete can become prohibitively expensive even if its underlying construction cost is manageable. Government guarantees, regulated returns and long-term electricity contracts can reduce the final cost to consumers by reducing financing risk.
This does not mean giving the industry a blank cheque. Public support should be conditional upon finalising designs before construction, maintaining transparent budgets and demonstrating measurable improvement between successive reactors.
The public should share in the benefits of lower financing costs, not merely absorb the consequences of every overrun.
Do Not Forget the Reactors We Already Have
Safely extending the life of an existing nuclear station can be considerably cheaper and less risky than building a new one.
Existing plants already have grid connections, experienced employees, local acceptance, security systems and established operating organisations. Closing a serviceable nuclear plant and promising an unbuilt SMR several decades later is not an equivalent exchange.
We should distinguish between plants that have genuinely reached the end of safe operation and plants being closed because of an arbitrary political timetable.
Once a nuclear station closes, the experienced workforce disperses and the supply chain contracts. Rebuilding that capability later is exceptionally difficult.
Nuclear and Renewables Are Not Enemies
The energy debate is often presented as a choice between nuclear power and renewable energy. This is unnecessary and unhelpful.
Wind and solar can provide inexpensive, low-carbon electricity when conditions are favourable. Nuclear can provide dependable low-carbon output regardless of the weather. Storage, interconnection and flexible demand can help combine these technologies into a resilient system.
The correct question is not which single technology should supply everything. It is which combination can provide the electricity Britain needs at an acceptable total-system cost.
A system containing large quantities of wind requires dependable power, storage, interconnection or controllable demand during periods of low wind. Nuclear is one possible source of that dependable electricity.
Its value should therefore not be measured solely against the price of a unit of wind power generated at midday on a windy day. The comparison must include the cost of delivering reliable electricity throughout the year.
Waste, Fuel and Energy Security
The physical quantity of high-level nuclear waste is small compared with the waste produced by fossil fuels, but it remains hazardous for exceptionally long periods.
Geological disposal is technically achievable, but it requires competent institutions and public consent extending beyond normal political cycles. The industry damaged its credibility when it spoke as though waste was somebody else's future problem. Any renewed programme must be honest about the responsibility.
Fuel security also deserves more attention.
Uranium is sourced from several countries and can be stockpiled much more easily than natural gas. A nuclear plant can hold years rather than days of fuel. However, uranium conversion, enrichment and specialist fuel fabrication are concentrated industries, and Western countries retain some dependence upon Russian nuclear-fuel services.
Energy independence therefore requires more than constructing reactors. It requires secure fuel-cycle capacity, component manufacturing, engineering skills and long-term waste institutions.
AI Changes the Scale of the Question
Electricity demand may be about to change dramatically.
Artificial-intelligence data centres, electrified transport, heat pumps, advanced manufacturing and industrial reshoring could reverse decades of relatively flat electricity consumption.
The question is no longer simply how Britain decarbonises the electricity system it already has. It is how we build a much larger system without making energy prohibitively expensive or increasing our dependence upon other countries.
Britain talks frequently about becoming an AI superpower. But artificial intelligence ultimately depends upon physical infrastructure: data centres, semiconductors, cooling systems, transmission networks and enormous quantities of dependable electricity.
Conclusion
The tragedy of nuclear power is not that the technology stalled. It is that many of the countries that pioneered it forgot how to build it.
Hiroshima and Nagasaki burdened nuclear technology with an image no other energy source has had to carry. Three Mile Island, Chernobyl and Fukushima reinforced that fear.
But public anxiety alone does not explain the West's decline. We stopped standardising designs, dispersed responsibility among too many organisations, allowed planning and construction to last for decades and attempted to finance century-long strategic assets through short-term electricity markets.
France showed what nuclear power can achieve. China and Russia are demonstrating what programme continuity can preserve. The military has shown that compact reactors can operate reliably when there is a committed customer, a stable design and an enduring technical organisation.
SMRs may offer the West a route back — but only if "modular" becomes a manufacturing reality rather than a marketing word.
The choice is not between nuclear power and renewable energy. It is between building a plentiful, dependable, low-carbon energy system and continuing to assume that electricity will somehow appear whenever it is required.
If Britain intends to lead in artificial intelligence, advanced manufacturing and electrified transport, it will need substantially more power. Nuclear energy cannot solve that problem alone. But excluding it because of its history may leave Britain without enough electricity to create its future.