Nuclear Could Double by 2050. A New Global Report Says the Role Will Be Smaller Than the Hype.

The nuclear renaissance has been one of the loudest stories in energy this decade. Tech giants are signing reactor deals to feed AI data centers, governments are announcing SMR programs, and uranium markets have soared on the optimism. Now a new global assessment from the Energy Transitions Commission, published 6 October, offers the most sober accounting yet of what nuclear power can and cannot realistically do. Its title says it all: The Role of Nuclear in the Energy Transition: Complementary but Limited.

The headline finding is balanced, almost deliberately so. Global nuclear capacity could roughly double by 2050, from about 380 gigawatts today to around 750 gigawatts, supplying roughly a tenth of the world's electricity. But that outcome is conditional, expensive, and, in most countries, economically second-best to solar and wind.

What the report actually concludes

The ETC's central scenario has nuclear holding a 10 percent share of a much larger electricity system. Total electricity demand is expected to double or triple by 2050, so nuclear grows two to three times in absolute terms, to between 550 and 1,000 gigawatts, without gaining ground on renewables.

Getting there is the hard part. Average new nuclear build would need to rise from 5 gigawatts a year, the average of the past two decades, to 15-25 gigawatts a year, every year, through 2050. Annual investment would need to quadruple from about $50 billion to $200 billion. The report notes that build rate was last achieved in the 1970s and 1980s in the United States and Europe, driven partly by an oil crisis.

And for the other 90 percent of decarbonized electricity, the answer is the same one it has been for a decade: solar and wind. In most developed countries, the report finds, new solar and onshore wind are typically 20-50 percent cheaper than new nuclear, while offshore wind costs are comparable or lower. Falling battery costs are increasingly letting solar and wind deliver cost-competitive power around the clock.

The nuclear debate has been too binary for too long. This report settles it with evidence: build where it is cheap, do not where it is not.

The SMR reality check

Small modular reactors have absorbed an enormous share of nuclear optimism. They are the technology tech companies keep pointing at when asked how they will power AI data centers. The ETC's verdict is blunt: SMRs remain one to two decades away from wide deployment.

The numbers explain why. Early US cost estimates for SMR electricity run $140-270 per megawatt-hour, well above the $40-190 per megawatt-hour range for large reactors. US SMR capital cost estimates have inflated from $3,000-10,000 per kilowatt to $12,000-15,000 per kilowatt in just a few years, hit by the same overruns as large reactors. Modular designs could cut costs 20-30 percent, the report says, but only with manufacturing scale and standardization. With more than 70 competing SMR designs globally, that consolidation will take time.

Pilot projects are underway in China and Canada, but the ETC estimates widespread deployment will not occur until the late 2030s and 2040s, and only if costs decline. For businesses needing reliable clean power now, the report points to proven, cheaper alternatives: grid-scale renewables plus storage, and power purchase agreements with existing nuclear plants. Tech giants are already pursuing exactly those deals.

Where nuclear makes sense, and where it does not

The report's most useful contribution may be its insistence that there is no universal case for nuclear. A country's optimal nuclear share depends on two things: whether it can build reactors cheaply, and whether its grid actually benefits from them.

On cost, the geography gap is staggering. Similar reactor designs cost 4-9 times more to build in the US and Europe than in China and South Korea. New nuclear plants in China, South Korea, and India deliver electricity under $50 per megawatt-hour; in the US and Europe, comparable plants cost $150 per megawatt-hour or more. The drivers are unglamorous: standardized fleet-build programs, shorter construction timelines, more effective regulation, and cheaper capital. Chinese state-backed projects raise money at around 1-3 percent; privately financed Western projects pay 8-10 percent.

On grid value, the answers differ by country. The UK, with expensive offshore wind and high balancing costs, could economically support up to 20 percent nuclear generation. In India, the optimal share is only 10-20 percent, because abundant solar paired with batteries is simply cheaper. And for countries with strong renewable resources and no established nuclear industry, most of Sub-Saharan Africa, Latin America, and Australia, the report finds no economic case for nuclear at all: the same money buys faster, cheaper decarbonization from renewables. Batteries paired with solar are making that math even clearer.

The cheapest nuclear megawatt is the one already built

Amid the new-build debate, the report highlights an underappreciated lever: keeping existing reactors running. Safely extending plant lifetimes from 60 to 80 years costs $500-1,100 per kilowatt, versus $2,000-18,000 per kilowatt for new builds, and yields electricity below $40 per megawatt-hour.

Twelve US plants have been approved for 80-year operation since 2019. Extending existing global capacity to 80 years could retain 230 gigawatts that would otherwise retire by 2050, reducing new-build requirements by a third. In other words, the most cost-effective nuclear program in the world may be maintenance.

The supply chain elephant

The report flags a risk that gets far less attention than safety or waste: supply chain concentration. Russia controlled about 40 percent of global uranium enrichment capacity in 2025 and still supplied around 20 percent of EU enriched uranium in 2024, even after EU efforts to reduce Russian imports. Reactor pressure vessels, turbines, and control systems are manufactured in just six countries: France, South Korea, Japan, the US, China, and Russia.

That concentration is manageable while build rates stay modest, the report says, but diversifying the supply base enough to support a true scale-up could take a decade or more. Nuclear's greatest vulnerability, per the ETC, is not its waste or its safety record. It is who controls the fuel and the factories. Energy planners betting on storage face fewer such chokepoints.

Perspective

The report does not ask anyone to pick a side in the nuclear culture war. It asks for arithmetic: where nuclear is cheap to build and genuinely useful to the grid, build it. Where it is not, do not divert scarce capital and institutional capacity from the renewables that win decisively on cost, speed, and security. On average, the ETC lands on 10-20 percent of a country's future generation coming from nuclear plants: a meaningful role, but a supporting one.

The nuclear renaissance may yet happen. But per this report, it will be smaller, slower, and more regional than the hype suggests, carried less by breakthrough reactor designs than by extending the life of reactors we already have and building new ones where the economics actually work.