- Artificial intelligence and nuclear power are becoming inseparable pillars of the global economy, IAEA Director General Rafael Mariano Grossi told delegates at the International Symposium on Artificial Intelligence and Nuclear Energy in Vienna on Wednesday.
Grossi said two forces, the rapid ascent of AI and the accelerating shift towards clean, reliable energy, are “reshaping humanity’s horizon” at a pace once described by futurists as arriving “too fast and in the wrong order”.
“The AI revolution, through its scale and speed, was always going to choose nuclear energy as a partner,” he said. “The only question was when. Today we know the answer is now.”
AI’s energy appetite drives nuclear resurgence
Grossi argued that AI’s vast and continuous electricity demand is converging with global imperatives for energy security and climate action. Training frontier AI models requires immense computing power, while everyday applications — from healthcare and transport to public administration — add further strain on electricity grids.
“AI runs on vast, uninterrupted quantities of electricity,” he said. “And there is only one energy source that can meet the combined demands of low-carbon generation, 24/7 reliability, massive power density and genuine scalability: nuclear energy.”
This interdependence, he said, amounts to a “structural alliance — Atoms for Algorithms”.
AI reshaping nuclear operations
Grossi also highlighted AI’s growing contribution to nuclear safety, design and efficiency. He cited four key areas:
- Operations: predictive maintenance, anomaly detection and thermal-performance optimisation.
- Design: accelerated reactor modelling, fuel-cycle simulation and materials development.
- Safety: accident simulation and emergency-response planning.
- Safeguards: automated analysis of surveillance footage and satellite imagery.
These tools, he noted, are already in use within IAEA laboratories and in national programmes. But he stressed that “despite its brilliance, AI still needs a human” for judgement, impartiality and political context.
Data centres spur demand for small modular reactors
With global data-centre electricity use exceeding 400 TWh annually and expected to reach nearly 1,000 TWh within a few years, Grossi said the world faces “a huge opportunity to make sure our digital future runs on clean energy”.
He pointed to small modular reactors (SMRs) as a particularly compelling option for technology companies, as their modularity allows energy supply to grow alongside expanding AI clusters.
SMRs’ compact footprint and enhanced safety systems mean they can be located close to industrial hubs, reducing grid pressure and transmission losses. However, Grossi noted that deployment must accelerate: “SMRs need to get from the development stage to the international market fast and safely.”
Regional shifts in the AI–nuclear landscape
Grossi gave a tour of global developments:
- United States: leading in both nuclear capacity and AI infrastructure, hosting nearly half of the world’s data centres.
- Canada: expanding data-centre investment alongside major nuclear refurbishments.
- Europe: with dense digital corridors in Frankfurt, Amsterdam and London, and renewed nuclear commitments in France, the UK and newcomer nations such as Poland.
- Russia and China: advancing in both nuclear technologies and AI development, with China building more new reactors than any other country.
- Japan and the Middle East: returning to or exploring nuclear energy while scaling AI capabilities, particularly in the UAE and Saudi Arabia.
- Asia: fast-growing markets such as India, Indonesia, Pakistan, Vietnam and the Philippines could become sustainable AI hubs if digital expansion is paired with clean power.
- Latin America: emerging AI and cloud markets in Brazil, Mexico, Chile and Colombia, with Argentina well-placed to combine digital growth with an established nuclear sector.
- Africa: undergoing a “digital leap” with rising internet adoption and growing interest in nuclear power from countries including Egypt, Kenya, Ghana and Zambia.
Across all regions, he said, the IAEA is supporting nations through review missions, planning tools and capacity building to ensure safe and effective nuclear development.
IAEA to act as bridge between sectors
Grossi positioned the Agency at the centre of the emerging nexus between AI and nuclear energy.
“We are facilitating the safe integration of AI into nuclear operations, and accelerating nuclear deployment,” he said. This includes cooperation with the World Bank, regional development banks and national regulators, as well as assistance through the IAEA’s Milestones Approach for new nuclear countries.
The Agency will use the discussions at this week’s symposium, attended by representatives from 252 organisations across industry, government, research and civil society, to develop a framework and action plans aimed at ensuring AI and nuclear technologies reinforce one another’s growth.
Grossi closed with a call for collaboration: “Talk to each other openly, continue the dialogue when you return home, collaborate across borders and disciplines, and come back to report on how you have transformed ideas into real progress.”
Invoking Niels Bohr’s famous line, “Prediction is very difficult, especially about the future”, he urged delegates to focus not on forecasting but on shaping what comes next.
“If we get this right,” he said, “AI will one day say, in whatever language it invents: ‘They understood the challenge and they did what was needed.’”
Author: Bryan Groenendaal
