Editor’s note: This is the third story in a series of Research Money stories that will examine the benefits and risks of Canada’s nuclear power expansion. The first story was published on March 18. The second story was published on March 25, 2026. Our stories will include the perspectives of both nuclear power proponents as well as those who think nuclear is not the option Canada needs.

Nuclear power expansion underway in Ontario and planned in some other provinces is not a cost-efficient or timely way of generating electricity or reducing greenhouse gas emissions, says a Canadian energy policy expert and nuclear energy critic.

Building more nuclear reactors will lead to project cost overruns, higher electricity costs for consumers, more nuclear waste for which there’s no proven solution, and a greater risk of nuclear weapons proliferation, said M. V. Ramana (photo at right), professor and Simons Chair in Disarmament, Global and Human Security in the School of Public Policy and Global Affairs at the University of British Columbia.

Ramana, who has a PhD in physics, is the author of the book Nuclear is Not the Solution: The Folly of Atomic Power in the Age of Climate Change.

The amount of nuclear power globally has been declining during the last three decades, Ramana said during a webinar presented by environmental advocacy organization Northwatch.

“The reason why nuclear power has been declining is because it’s not economically competitive [with other energy sources], and in turn that’s because nuclear reactors cost too much to build,” he said.

The maximum amount of nuclear power globally totalled about 17.5 percent of the world’s electricity in 1996, which had declined by 2024 to about nine percent of electricity worldwide, he said, citing the World Nuclear Industry Status Report.

At the same time, renewable energy sources like solar and wind power have steadily become cheaper. Renewables have grown from supplying about one percent of the world’s electricity in the mid-1990s to about 17 percent in 2024, according to data from the Energy Institute.

Building a nuclear reactor is very complex, requiring large amounts of concrete and steel and highly trained workers, and must be designed and built to not allow any escape of radioactive materials into the environment, Ramana said.

This is why nuclear reactors that have recently been built in the U.S., France and the U.K. typically costs tens of billions of dollars, with the latest grid-connected reactors in the U.S., the two Vogtle reactors in Georgia, costing about $37 billion, he said.

Also, historically the cost of building a new nuclear reactor has always ended up being higher than the original estimate, and completion of the new plant has always taken longer than initially forecast, he noted.

In the 1980s, several new nuclear plants in Washington were cancelled after estimated construction costs increased from $4.1 billion to over $24 billion.

Two new reactors in South Carolina were cancelled after costs rose from $9.8 billion to $25 billion.

The two new Vogtle reactors in in Georgia ended up costing about $37 billion, up from $14 billion when their construction started in the 2010s.

The cost of the U.K.’s first new nuclear plant in a generation at the Hinkley Point C site has risen from £18 billion when it was given the green light in 2016 to approximately £35 billion.

French utility EDF said the first reactor at the Hinkley Point C site in Somerset will begin operations in 2030 – almost 13 years after construction work began – after a series of delays to the project.

In Canada, building the G7’s first four small modular nuclear reactors at the existing Darlington nuclear power station site is forecast to cost $20.9 billion, with the initial reactor forecast to cost about $7.7 billion.

“So there’s no way you can ever imagine that there is a nuclear reactor that is actually cheap,” Ramana said.

Cost and time required to build new nuclear power plants have increased

Commercial nuclear power plants whose construction began in the late 1960s cost $1,000 per kilowatt-electric (1,000 watts of electrical power generated) or less (in 2010 dollars); plants started just 10 years later cost nine times that much, according to an article by Brian Potter (photo at right), published by the Institute for Progress.

A 1982 analysis by the U.S. Energy Information Administration of 75 U.S. nuclear plants found that cost estimates steadily increased as construction went on, with final construction costs two to four times as high as the initial estimated cost.

Plants were constantly late and over-budget, and became an increasingly unattractive financial investment. Starting in the late 1970s, plants began to be canceled in large numbers, which accelerated following the accident at Three Mile Island.

Recent attempts at nuclear plant construction have at best ended with massive budget overruns (in the case of Vogtle Units 3 and 4), Potter said. At worst, they’ve ended in failure after billions were spent (in the case of VC Summer Units 2 and 3 in south Carolina).

For natural gas plants, up to 70 percent of their electricity cost comes from the cost of fuel. With nuclear plants, on the other hand, 60 percent to 80 percent of their electricity cost comes from constructing the plant itself, Potter noted. “Decreasing the construction cost of plants would thus drive the cost of the electricity they provide down substantially.”

Partially because of continual regulatory change, the time required to build a nuclear plant in the U.S. has continuously increased. The minimum time required to build a plant increased from four years in the late 60s to eight years in the mid-1970s, with 75 percent of reactors taking 10 to 15 years to build.

Ramana cited a report by U.S.-based global financial services firm Lazard that found it costs US$180 per MW hour to generate power from a new nuclear reactor in the United States, compared with US$61 for onshore wind and US$58 for utility-scale solar power.

“The fact that nuclear reactors are so expensive means the [cost of] the electricity they produce is going to be very high,” he said.

Nuclear energy proponents maintain that new reactor designs and modular construction methods, especially for small modular reactors (SMRs), will dramatically lower construction costs and make building reactors more efficient.

But Ramana disagreed, pointing out that SMRs are an unproven technology, with no commercial-scale, operating SMR connected to the electricity grid anywhere in the world.

A final reactor design for the GE Hitachi BWRX-300 reactor to be used in the four SMRs at Darlington hasn’t yet been finalized or approved for all four units, even though Ontario Power Generation has secured a construction license.

The challenges of the Darlington SMRs being first-of-their-kind projects are bound to lead to increased costs for the reactors and for electricity consumers, along with project delays, Ramana said.

Small modular reactors cost more per unit of electricity generated than large conventional reactors

Another problem with SMRs is that an SMR will produce 300 megawatts (MW)or less power compared with a large conventional reactor that produce 1,000 MW or more.

The law of scaling in engineering economics holds that the bigger the nuclear reactor, the smaller the cost per unit of electricity generation, Ramana said.

A 900-MW reactor won’t need three times as much concrete and steel and construction workers as one 300-MW reactor, so the construction cost of the larger conventional reactor will actually be cheaper per unit of electricity generated than the smaller SMR, he said.

The larger reactor also produces three times as much electricity as the smaller SMR, so it also produces three times as much revenue for the reactor operator, he added. “It’s much more profitable for companies to build larger reactors.”

Research from CSIRO, Australia’s national science agency, shows that the highest cost of available energy sources in Australia is SMRs, which is higher than all of the other alternatives and much higher than renewables including solar photovoltaic and onshore or offshore wind power.

“Small modular nuclear reactors (SMRs) remain the highest cost option, even with new data from Canada’s Darlington project,” according to CSIRO’s GenCost 2023-24 report.

Renewables (wind/solar) cost $73 to $128 per megawatt-hour (MWh), while large nuclear is $141 to $233/MWh (by 2030), and SMRs are $230 to $382/MWh. 

Proponents of the four SMRs at Darlington maintain that the cost of each SMR build will be cheaper than the previous SMR, because of what’s learned from constructing the first SMR and the subsequent reactors.

But Ramana said such an argument flies in the face of historical data on the cost of constructing nuclear reactors. In both the United States and France, the cost of constructing new reactors was more expensive than that of earlier nuclear plants.

“So there is no historical reason to assume that the costs of small modular reactors are going to decline with time,” Ramana said.

Even allowing for optimistic learning rates, the number of SMRs that would need to be built for their cost-per-unit of electricity to be comparable with large conventional reactors would range from a few hundred SMRs to several thousands. “So the question is, ‘Who’s going to pay for all of these lost leaders, as it were?’”

Nuclear energy proponents also argue that renewable energy sources like wind and solar are intermittent, so can’t provide the 24-7 “baseload power” – the minimum amount of power required at any time –required by heavy industry and other large electricity consumers.

But Ramana argued that the term baseload power is misleading and an outdated way of looking at electricity grid management.

A nuclear reactor is designed to produce power at a steady level, so varying the output in order to meet baseload power requirements and manage renewables’ intermittency will actually drive up the cost of producing power, he said.

CSIRO’s GenCost 2023-24 report found that with various renewables making up as much as 90 percent of the electricity in the grid, “the cost of power from nuclear power plants, in particular small modular reactors is going to be much higher,” he noted.

The way to deal with renewables intermittency is to balance the variability of the renewable sources used with a diversity of technologies and geographical sourcing, and enhanced, flexible transmission lines, Ramana said.

Smart grids utilize AI-driven forecasting to predict supply, while demand-side management, such as time-of-use pricing and electrical vehicle charging, adjusts electricity consumption to match generation.

In Canada, more interconnected electricity transmission lines between provinces would enable the transfer of electricity among provinces to help manage renewables’ intermittency, he said.

There are also technologies, such as pumped hydro and energy storage, that are becoming more advanced and growing in use and can provide electricity in response to intermittent renewable energy.

Still no proven solution for managing radioactive nuclear waste

Nuclear energy proponents also maintain that new nuclear power plants are needed to help reduce greenhouse gas emissions and combat global warming and climate change.

But Ramana said new reactors – both SMRs and large conventional reactors – won’t be built in time to reduce emissions quickly enough to have a significant impact on the urgent crisis of global warming’s current impacts. “Nuclear power is not the solution to climate change.”

History shows that the planning-to-operation times of all nuclear plants ever built have been 10 to 19 years or more.

A typical nuclear plant takes at least 10 years to be built from the time concrete starts pouring to when the plant is connected to the electricity grid. So that means the earliest any of Canada’s planned new reactors, including the four SMRs at Darlington, would be able to generate electricity would be 2036 at the earliest.

Although the first SMR at Darlington has a construction licence, it will still require a separate regulatory review by the Canadian Nuclear Safety Commission to obtain a licence to operate.

Therefore any investment in nuclear power which is going to come at the cost of investment into renewables and other sources of power is actually a negative in terms of how it contributes to [combatting] climate change,” Ramana said.

Another rationale nuclear energy proponents put forward is that new nuclear power plants are needed to power all the AI data centres that are being built or planned.

But Ramana said even the biggest U.S. hyperscalers are investing relatively small amounts in building new nuclear reactors, compared to what these reactors actually will cost.

“Big tech companies are pushing for building nuclear power to support AI and data centres, mostly as a greenwashing exercise,” he said.

The data centres themselves have problems, including enormous power consumption and freshwater use for cooling, he noted.

SMRs also will create different types of nuclear wastes than with conventional large reactors. There’s currently no proven solution for long-term management of nuclear waste, and even the deep geological repository (DGR) approved in northern Ontario wasn’t designed for new types of radioactive waste from SMRs.

Critics also contend the process for choosing the DGR site was flawed. The repository is expected to cost at least $26 billion.

Ramana pointed out that a nuclear waste isolation pilot plant repository in New Mexico, meant to store waste resulting from the U.S. nuclear weapons program, had an accident in 2014 – about 15 years after the facility started operating – that involved an explosion of a drum of nuclear waste and exposed more than 20 workers to a release of radioactive plutonium.

The accident cost the U.S. more than $1 billion and the cleanup cost about $640 million.

Also, the four SMRs at Darlington will require low-enriched uranium (typically three to five percent) as fuel. Canada doesn’t have its own enrichment facility, so will need to procure enriched uranium from countries that do, such as the U.S., Germany, the Netherlands, France and the U.K.

Some SMR designs would use reprocessed spent nuclear fuel rather than enriched uranium that conventional CANDU reactors use, which could increase the risk of nuclear proliferation, Ramana said.

Some scientists and civil society groups are calling for a ban on reprocessing spent nuclear fuel in Canada, amid signs that federal officials began reconsidering their stance on enriched uranium a few years ago.

Reprocessing spent fuel will create new radioactive waste streams – including radioactive off-gases, spent fuel salt waste, and irradiated graphite waste – that will need to be controlled and the waste treated before being stored in a deep geological repository, Ramana said.

Although reprocessing is technologically feasible, “it’s not economical and it never will be,” said Dr. Allison Macfarlane, PhD, professor and director of the School of Public Policy and Global Affairs at the University of British Columbia.

In addition, proliferation is "a serious risk and not to be taken lightly. Canada would have to be thinking about the example it’s setting for other countries if it were to go ahead with this,” Macfarlane said.

Ramana said the scientific facts and checkered history of nuclear power are well known, but the nuclear industry consistently counters with a “propaganda machine that basically keeps using the word ‘clean’ every time they say nuclear.”

The nuclear industry keeps saying “clean nuclear power, safe nuclear power,  all knowing that it is not,” he said.

R$