11th November 2025 By:- M. V. Ramana
Nuclear power is now endorsed as carbon-free, abundant, clean energy, with 30 countries signing a pledge to triple nuclear power capacity. But author of Nuclear is not the solution M.V. Ramana argues strongly against this, claiming that nuclear promoters are ignoring the evidence. He argues, nuclear energy is not economically efficient. Even more importantly, the long timescales involved in expanding nuclear power potentially could make the climate crisis worse, not better.
In the last few years, there has been a resurgence in talk about expanding nuclear power. One often hears that nuclear power is needed to supply energy to data centers or to mitigate climate change. Neither of these arguments makes any sense when one looks at actual trends in the energy sector. Nuclear energy has become less, not more, important to supplying electricity around the world.
In the United States, which has the largest demand for energy for its data centers, and has the highest per capita emissions among leading countries, the number of operating reactors has declined by 10 units since 2006. This decline occurred despite the US government offering economic incentives to electric utility companies to build nuclear reactors through the Energy Policy Act of 2005. These incentives aimed to “reduce the financial risk of investing in advanced nuclear power plants by transferring risk to the public,” leading to US utility companies proposing to build more than thirty reactors.
Only two of these reactors were eventually built, but in the meantime many more older and unprofitable reactors were shut down. Two other reactors were abandoned mid-project following huge cost and time overruns, after over $9 billion was spent; electricity customers in South Carolina are still paying every month for that wasted effort.
Globally, the share of electricity from nuclear reactors has come down from 17.5% in 1996 to only 9% in 2024. In contrast, modern renewables (i.e., not including power from large hydroelectric dams) produced 17.3% of the world’s electricity, up from around 1 percent in the mid 1990s.
The key reason for these trends—the decline in nuclear power and the rapid increase in renewables—is economics. Solar photovoltaics, especially when built at large (utility) scale, has become the least costly option for new electricity capacity in recent years; in 2020, the International Energy Agency pronounced that solar is “the new king of the world’s electricity markets.” According to the Wall Street company Lazard’s 2025 estimates, electricity from a new nuclear power plant in the United States costs roughly three times the corresponding costs at solar or wind energy plants. This large gap between nuclear power and renewables has been growing larger because of the decline in solar and wind energy costs, and this trend is expected to continue over the coming decades.
The comparison between nuclear power and sources of electricity like solar and wind is complicated by the fact that the latter sources do not generate power steadily, and depend on how much wind is blowing or sun is shining. But the large cost differences between nuclear and renewables should be more than enough to allow for complementary technologies to compensate for variations in the outputs of solar and wind farms. More generally, renewables can be the basis of a reliable electricity system provided suitable and affordable options, such as energy efficiency, demand response, technological and geographic diversity, and some storage, are incorporated.
Building reactors is costly
The poor economics of nuclear energy are a result of the high cost of building nuclear reactors. A good illustration is the only nuclear power plant involving two EPR reactors being built in the United Kingdom—Hinkley Point C—whose cost was estimated at £46 billion in 2024. The final cost might be higher still: a decade ago, the plant was estimated to cost £16 billion; in other words, the cost has more than doubled even after taking inflation into account. And more delays and resulting cost escalations have just been announced. Similarly, the two AP1000 reactors built as part of the Vogtle project in the United States ended up costing over $36 billion, much more than the $14 billion estimated when construction of those reactors started. Both reactors also took much longer to build than had been initially expected. These recent cost escalations and delays are even more extreme than the historical pattern identified in an academic study that examined 180 nuclear power projects and found that 175 had exceeded their initial budgets by an average of 117% and took 64% longer than initially projected.
The second reason for this decline is the very long time it takes to build a nuclear reactor. The average nuclear reactor takes about 10 years from when concrete is poured into the ground to when it feeds electricity into the grid. But one cannot go from deciding to build a reactor one day to pouring concrete in the ground the very next day. The project’s proponents need to get the necessary environmental permits and safety evaluations, carry out public hearings (at least where they are held), and, most importantly, raise the tens of billions of dollars needed. In contrast, renewable sources, especially solar power, have been quick to scale and cheap.
Can new small modular nuclear reactor designs help?
When faced with these facts, some proponents of nuclear energy argue that alternative nuclear reactor designs will solve the problems confronting nuclear power. In the last couple of decades, most of these designs have been categorized as Small Modular (Nuclear) Reactors (SMRs). SMRs are designed to generate under 300 megawatts of electrical power, as compared to designs that generate in the 1,000–1,600 megawatt range that have dominated nuclear construction in recent decades.
SMRs have been receiving large amounts of funding. According to the OECD’s Nuclear Energy Agency, there is “approximately $15.4 billion of financing towards SMRs worldwide”. That sounds like a lot, but this investment is split across 127 SMR designs. To add more perspective, consider the case of NuScale, the leading SMR vendor in the United States. In an August 2025 earnings call, the NuScale CEO declared that the company had invested “over $2 billion”—and it still does not have a single project under construction.
SMRs, at least on paper, will be cheaper than the tens of billions that the Vogtle or Hinkley Point C reactors cost. But they will also generate lower amounts of power, which means less revenue for the owner. When the cost of SMRs is weighted by their power output, they turn out to be more, not less, expensive than large reactors. Larger reactors are cheaper on a per megawatt basis because their material and work requirements do not scale linearly with power capacity.
This is amply borne out by the estimated cost for a proposed US project involving six NuScale SMR modules, which was billed as “the first NuScale Power small modular reactor plant to begin operation in the United States”. In 2015, NuScale claimed that its reactors would be built at an “overnight cost” of $3 billion. This rose to $4.2 billion in 2018, then $6.1 billion in 2020, and finally $9.3 billion in 2023. As costs escalated, customers withdrew from the project, and in the end the project was terminated. When the cost of the project is divided by how much power it was designed to produce, it turns out to be around 250% more than the initial per megawatt cost for the 2,200 MW Vogtle project. Of course, the Vogtle cost exploded when construction started, and there is every reason the UAMPS project would have also escalated in cost had it been built. But even without such an increase during construction, the NuScale SMR design is more expensive than large reactors on a per-kilowatt basis.
Historically, too, small reactors were more expensive than large ones. In the 1950s, the US Atomic Energy Commission funded the construction of several small power reactors that were declared to be “suitable both for use in rural areas and for foreign export”. But all these reactors ended up shutting down early because they were not economically competitive.
SMR proponents have a counterargument: the lost economies of scale will be compensated by savings through mass manufacture in factories and resultant learning. The rate of learning refers to how quickly the costs of building and operating nuclear power plants go down as production increases—as more plants are built, processes and costs should improve. But, for the price per kilowatt for a small reactor to be comparable to large reactors, SMRs have to be manufactured by the thousands, even under very optimistic assumptions about rates of learning. Even then, SMRs will only be economically competitive with the likes of the Vogtle nuclear plant and generate power at costs that are many times that of renewable sources of energy.
These assumptions about rates of learning might be too optimistic for the real world, where multiple theoretical assumptions made by SMR developers will not hold. For example, they assume that costs of nuclear power plants will decline as more of these are built; but, in both the United States and France, the countries with the most nuclear plants, costs rose with time. The theoretical prerequisite for such learning is that most reactor builders would choose a standardised design. But if 127 SMR designs are being developed around the world, it is highly unlikely that one or even a few designs will be chosen by different countries and private companies.
The bottom line is that electricity from SMRs will be even more expensive than power from large nuclear plants, rendering them commercially unviable. A recent authoritative estimate of the cost of electricity from SMRs comes from Australia’s Commonwealth Scientific and Industrial Research Organisation (CSIRO), which showed that SMRs are by far the most expensive way to generate power. CSIRO’s estimate for the levelized cost of energy for an SMR project starting to deliver power in 2030 was between A$328 and 619 (US$214-404); in comparison, solar PV and onshore wind were estimated at A$35 and 63 (US$23-41) and A$64 and 107 (US$42-70) respectively. When integration costs are included to account for the variability of wind and solar power, the cost of electricity is estimated to be in the range of A$ 90-131 (US$59-85), even when variable resources constitute 90% of the electricity supply.
SMRs also take a long time to build. Russia’s KLT-40S, which is based on the design of reactors used in the nuclear-powered icebreakers operated by Russia for decades, took 13 years from the start of construction to generating electricity, instead of the expected 3 years. In China, the twin High Temperature Gas Cooled Reactor units (Shidao Bay 1-1 and 1-2) took more than twice the promised fifty months.
Combined, these two trends imply that nuclear energy will not help to solve climate change. For nuclear energy to play a significant role in mitigating climate change, its share of the electrical energy produced around the world has to increase, as fossil fuels are replaced by uranium. And the shift has to occur rapidly. Nuclear energy is simply not up to this challenge.
What about AI and data centers?
If nuclear power cannot be scaled up quickly enough to meet climate targets, then the chances of reactors being built to meet the increased power demands from data centers and artificial intelligence are even lower. Their demand is now or in the next few years, not in two decades. There is even the likely possibility that what we are witnessing is an “AI bubble” because these companies do not have a viable business model to pay for the immense expenses of building and operating the immense infrastructures needed.
The likely motive for these cloud companies to talk about nuclear energy is to greenwash themselves, a cheap way to reduce bad publicity about their environmental footprints. The amounts these companies seem to be investing in nuclear power are but a tiny fraction of their bloated revenues. At the same time, these investments are completely inadequate when compared to the funding needed to build even a single nuclear reactor, small or large.
Risks of nuclear power
In addition to cost and time, there are also good reasons why nuclear reactors are not a desirable way to produce electricity: the unavoidable risk of severe accidents, the inextricable connection to nuclear weapons proliferation, and the inevitable production of hazardous radioactive waste.
All nuclear power plants rely on the physical phenomenon of nuclear fission. This process will inevitably produce substances that are radioactive and are hazardous to human health and the environment. Despite decades of well-funded research, there is no demonstrated method of managing these materials safely due to a combination of social and technical problems. Even if these wastes are stored in geological repositories, the proposed management method that is currently most widely accepted, it is impossible to be confident that the radioactive materials stored in these will not leak out into the water and earth over the hundreds of thousands of years during which they will remain hazardous
Normally, these radioactive materials are kept contained inside reactors or in infrastructure designed to store them under water or with air cooling. But there are abnormal circumstances. All nuclear plants, including SMRs, can undergo accidents where large amounts of fission products are expelled from these reactors and result in widespread radioactive contamination. This possibility was on full display in 2011 when three reactors at Japan’s Fukushima Daiichi nuclear plant melted down. The smallest of these, Fukushima Daiichi-1, could generate just 460 megawatts, only slightly larger than the maximum output of 300 megawatts that characterizes an SMR.
All else being equal, making reactors smaller reduces the risk and impact of accidents. Smaller reactors have a lower inventory of radioactive material and less energy available for release during an accident. But even a very small reactor (say, one that generates under 10 megawatts of electricity) can undergo accidents that result in significant radiation doses to members of the public.
In the case of a small modular reactor, commercial proposals often envision building multiple reactors at a single site. The aim is to lower costs by taking advantage of common infrastructure elements. With multiple reactors, however, the combined radioactive inventories might be comparable to that of a large reactor. Multiple reactors at a site also increase the risk that an accident at one unit might either induce accidents at other reactors or make it harder to take preventive actions at others. This is especially the case if the underlying reason for the accident is a common one that affects all of the reactors, such as an earthquake. In the case of the accidents at Japan’s Fukushima Daiichi plant, explosions at one reactor damaged the spent fuel pool in a co-located reactor. Radiation leaks from one unit made it difficult for emergency workers to approach the other units.
Expanding nuclear power also makes the problem of nuclear weapon proliferation worse. Plutonium is created in all nuclear power plants that use uranium fuel. Practically any mixture of plutonium isotopes could be used for making weapons. However, the plutonium is produced alongside intensely radioactive fission products, and this makes it hard to utilize this material for any purpose. If any state carries out “reprocessing” of the spent fuel, then the plutonium can be separated from these fission products and used to make nuclear weapons.
Conclusion
Those who advocate for nuclear energy are ignoring the evidence for the decline in importance of nuclear energy and its inability to compete economically with renewable sources of energy. New reactor designs will not rescue nuclear power from this fate.
The climate crisis is urgent. The world has neither the financial resources nor the luxury of time to expand nuclear power. From the perspective of minimizing cost and time, expanding nuclear energy only makes the climate problem worse. First, the money invested in nuclear energy would save far more carbon dioxide if it were invested in furthering the switch to renewables. There is thus an economic opportunity cost to investing in nuclear energy. And the long timescales involved in expanding nuclear power mean that the reduction in emissions from alternative investments would not only be greater, but also quicker.
https://iai.tv/articles/the-hype-behind-nuclear-energy-doesnt-match-the-reality-auid-3415?_auid=2020