The International Atomic Energy Agency raised its global nuclear power forecast for the sixth consecutive year on September 14, projecting capacity could reach 1,284 gigawatts electric by 2060 in its high case, roughly 3.4 times the 377.1 GW(e) operating at the end of 2025. For the first time, the annual outlook models small modular reactors as their own category, and the agency's low case still nearly doubles the fleet.
Key facts
- Global nuclear capacity stood at 377.1 GW(e) across 413 reactors at the end of 2025; the high case reaches 1,045 GW(e) by 2050, up from 992 GW(e) projected last year.
- The high case projects 1,284 GW(e) by 2060, approximately 3.4 times the 2025 level; the low case reaches 696 GW(e), nearly double.
- Small modular reactors account for 28% of new capacity added by 2060 in the high case, up from 24% in last year's edition; the low-case SMR share jumped from 5% to 23%.
- In North America, about 60% of new nuclear capacity could come from SMRs; in South-East Asia and in Latin America and the Caribbean the figure is 40% in both scenarios.
- Nuclear generation rose 1.1% to 2,689.1 terawatt-hours in 2025, but nuclear's share of global electricity fell from 8.7% to 8.4% because total generation grew 2.7%.
- Two out of three operating reactors are more than 30 years old, and 45% have run for more than 40 years.
What happened
IAEA Director General Rafael Mariano Grossi presented the projections at the 70th IAEA General Conference in Vienna, contained in the 46th edition of the agency's annual Energy, Electricity and Nuclear Power Estimates report, known as RDS-1. This year's edition extends the projection horizon from 2050 to 2060 for the first time.
The report is the sixth year running that the agency has revised its outlook upward. The high case now sees capacity hitting 1,045 GW(e) by 2050, up from the 992 GW(e) forecast a year ago, before climbing to 1,284 GW(e) by 2060. Even the low case, which assumes current trends continue with few policy changes, reaches 696 GW(e) by 2060 and 641 GW(e) by 2050, up from 561 GW(e) projected for 2050 last year.
The biggest volume of additions is modeled in Central and Eastern Asia, where capacity could grow from 114 GW(e) in 2025 to 407 GW(e) by 2060 in the high case, with nuclear generation increasing almost fivefold.
The SMR modeling is the headline structural change. The IAEA defines an SMR as up to 300 MW(e), and projects 284 GW(e) from SMRs by 2060 in the high case. At an average unit size of 300 MW(e) that equates to about 946 SMRs; at 250 MW(e), roughly 1,136 units. Jessica Callen-Kovtunova, an IAEA energy planner, told a media briefing that under the high case 70% of currently operating capacity remains in operation in 2060, while under the low case almost two-thirds of today's units would be retired.
Why it matters
This is the strongest official international validation yet of the nuclear-renaissance narrative, and it lands at the agency's flagship annual gathering. When the IAEA's own conservative low case nearly doubles the fleet, the policy debate shifts from whether nuclear grows to how fast and who captures the supply chain.
The SMR numbers give the industry a quantified demand signal it has lacked. Roughly 60% of new North American capacity coming from SMRs in the agency's modeling aligns with what hyperscalers are already signaling: flexible, factory-built units sited near data centers rather than gigawatt-scale baseload. NNN's Nature Outlook analysis of SMRs and AI power demand and our 2026 developer landscape track exactly the vendors this forecast now legitimizes.
The revision cadence matters too. Six consecutive annual increases means the errors have all run in one direction. The IAEA attributes the latest step-up to energy security concerns intensified by successive energy crises since 2022 and to renewed engagement by international financial institutions, including the World Bank Group and the Asian Development Bank, a shift the industry has lobbied for over a decade.
There is one uncomfortable counterpoint inside the report: nuclear generation grew 1.1% in 2025 but lost share of the global electricity mix because everything else grew faster. The forecast is a capacity promise, not yet a generation result.
Background
The RDS-1 projections draw on three inputs: the IAEA's Power Reactor Information System, the conclusions of its annual consultancy meeting on nuclear capacity projections, and the joint IAEA/OECD-NEA Uranium Resources, Production and Demand (Red Book) assessment of fuel availability.
The fleet it projects from is old. Two-thirds of operating reactors have passed their 30th anniversary, and 45% have passed 40 years. Offsetting retirements is already part of the arithmetic: in 2025, seven reactors totaling 2.8 GW(e) were retired, while three new reactors totaling 3.0 GW(e) connected to the grid and construction began on 12 more expected to add 14.3 GW(e). The agency calls lifetime extension of viable existing reactors one of the most cost-effective low-emission generation sources available, a point reinforced by last week's joint NEA/EPRI analysis finding life-extended nuclear the cheapest of 23 technologies assessed on levelized cost.
Asia is doing the heavy lifting. China approved eight reactors in a single batch earlier this year, and the high case has Central and Eastern Asian capacity nearly quadrupling by 2060. The industry's own trade body is more bullish still: the World Nuclear Association's World Nuclear Outlook Report, published last week, found capacity could reach 1,457 GW(e) by 2050 if all national targets are met and the existing fleet keeps running.
What's next
The immediate test of the high case is whether policy converts to concrete. Watch for SMR final investment decisions in North America over the next 18 months, licensing harmonization progress from the IAEA's SMR Regulators' Forum, and whether the World Bank Group and Asian Development Bank convert their renewed engagement into financed first-of-a-kind projects. The next RDS-1 edition, due in September 2027, will show whether the seventh straight upward revision is on the table.
Questions
- How much nuclear capacity could the world have by 2060?
- The IAEA's high case projects 1,284 GWe by 2060, about 3.4 times the 377.1 GWe operating at the end of 2025. The low case reaches 696 GWe, nearly double. Global capacity was 413 reactors at the end of 2025.
- Why did the IAEA raise its nuclear forecast again?
- It is the sixth straight annual upward revision. The IAEA cites energy security concerns after the energy crises since 2022, rising electricity demand including from AI data centers, and renewed engagement from international financial institutions such as the World Bank Group.
- What share of new nuclear capacity will come from SMRs?
- In the high case, 28% of the roughly 1,017 GWe of new capacity added by 2060 comes from small modular reactors. In North America, about 60% of new capacity could be SMRs; in South-East Asia and Latin America and the Caribbean, 40%.
- Is the IAEA's high-case projection realistic?
- The IAEA describes the high case as plausible and technically feasible, noting capacity could exceed it. Actual deployment depends on investment, workforce development, regulatory harmonization, and offsetting retirements from an ageing fleet.
Sources
About Nuclear News Network
Nuclear News Network (NNN) is an independent publication covering the global nuclear energy sector — reactor construction, SMRs, fuel supply, policy, operations and fusion. NNN publishes a daily brief, same-day analysis of major developments, and reference guides used across the industry. Articles are produced by the NNN Newsroom, an editorial automation system with human oversight, under the publication's editorial standards. Ruben Seoane is the founder and main editor of NNN.