Can AI help solve electricity problem it is helping create?
Artificial intelligence (AI) and its insatiable appetite for computing power is behind one of the main reasons for an unprecedented global surge in demand for electricity: data centres. Ironically, AI is also touted as part of the solution to the very problem it is helping to cause.
Nuclear power generation is also seeing a surge of interest in its potential ability to meet that demand, but first the ecosystem around nuclear power plants needs the tools to develop new capacity at the speed and scale that is needed.
The numbers are astronomical. Forbes magazine, citing UN research, writes that global electricity demand is projected to grow by more than 10,000 terawatt-hours by 2035, equivalent to doubling the entire current consumption of every advanced economy on earth. Data centres are driving more than a fifth of that growth in wealthy countries.
The world as a whole is belatedly waking up to the importance of nuclear power in the overall energy mix. After more than two decades of stagnation, the capacity of the global nuclear fleet is expected to increase by one-third by 2035.
That’s the projection of the latest World Energy Outlook from the International Energy Agency (IEA) based on existing energy policies and the expectation of “a record high in nuclear power output” in 2025.
"Technology advances… are improving the outlook for nuclear power,” says the report. “As demand surges and the need for reliable, low-emission baseload electricity increases, nuclear is increasingly seen as a critical part of a secure, affordable and diverse electricity mix."
It’s a welcome reversal in fortune for the nuclear industry, which had “suffered some decline and stagnation” from the late 1970s to early 2000s, although its share of world electricity generation remained at 16-17 per cent.
Now, with the very public support for nuclear by the current US administration, the French government’s new multiyear energy programme (PPE) putting renewed focus on nuclear, and other countries looking into it, the nuclear industry appears to be gaining momentum. This point was not lost on the international visitors, delegates and exhibitors who attended the 2025 edition of the World Nuclear Exhibition (WNE) in record numbers and contributed to the buzz in the exhibition hall and conference rooms.
WNE 2023_© Collectif Favart
Enter AI: along with other advanced digital technologies, it is poised to revolutionise nuclear operations, safety, and efficiency, from real-time plant optimisation to predictive maintenance and next-generation reactor design. Given the current state of nuclear newbuild activity, most AI applications are concentrated on documentation, licensing, and compliance rather than operational control functions, which are slow to change in such a highly regulated industry.
Intersection of nuclear energy and AI
That does not impede development. What happens next at the intersection of nuclear energy and AI was the focus of a key panel discussion at WNE 2025.
Panellist Anne-Charlotte Dagorn, vice-president for North America of WNE platinum sponsor Assystem, said it was “hard to imagine” what the technological capabilities of AI will be by 2040, considering the computational power of AI has been doubling every six months for the past decade. “So from a pure technological point of view, the sky is the limit.”
AI was already producing tangible results in three key areas, she said: automation, decision-making, and knowledge transfer. However, by 2040 “we also need to find the right balance to make sure AI reinforces but does not replace expert knowledge. Performance is important but in the nuclear industry safety is key so we need to make sure that our experts stay skilled, empowered and accountable.”
Dagorn cited three key conditions to allow the industry to leverage AI at scale: collaboration, cybersecurity, and the right culture of governance and trust. On top of that, she added, “training is a must.”
Speeding up approval process
The industry has taken considerable interest in the moves by the current US administration to speed up the approval process for nuclear newbuilds, particularly small modular reactors (SMRs). The Department of Energy (DoE) said the result has been “an unprecedented transformation in the nation’s approach to delivering more nuclear power than ever before.”
Industry’s response has made front-page news. Microsoft and Nvidia formed an ‘AI for nuclear’ partnership intended to streamline the permitting, design, and operations of nuclear power plant facilities. The partners believe their AI tools can fast-track ambitious nuclear technology builds by breaking the “infrastructure bottleneck” of red tape, redundancies, and engineering and construction delays without compromising safety and quality.
At WNE 2025, platinum sponsor Westinghouse introduced its “nuclear-specific” HiVE AI system, alongside its generative AI model, Bertha. “Maintaining safety, profitability and output calls for new solutions,” the company said, adding that it was “revolutionizing our solutions with cutting-edge AI technology.”
The US’s Nuclear Regulatory Commission (NRC) has proposed a ‘data bridge’ pathway that would allow commercial applicants to reference safety data from reactors already authorized. This would enable a ‘build once, certify once, scale everywhere’ strategy – exactly the kind of repeatable-template process that AI document generation tools are built for.
The SMR designer X-Energy, for instance, projects the NRC review process for its Xe-100 reactor and fuel could eventually shrink to as little as six months once it reaches ‘next-of-a-kind’ status.
Some concerns have been expressed about whether the emphasis on speed will come at the cost of diligence. The AI Now Institute, whose focus is on social impacts of AI, argues in a 2026 report that compressing review timelines regardless of whether a safety record has been established for new designs undermines nuclear regulation and raises “serious safety and oversight concerns”. It warns that a public perception of the risks of fast-track initiatives could turn against the nuclear sector at a time when it is gaining traction for its role in global decarbonisation efforts.
Impact on future workforce
The American Nuclear Society offered a sobering assessment of AI’s impact on the future workforce in a panel discussion at its 2026 annual conference.
Although AI tools could be useful in helping bring new entrants to the industry “up to speed” quickly by bridging the gap in training left by the retirement of veterans who traditionally filled a mentoring role, an unintended consequence could be the undermining of the very skills that human operatives need.
This was the warning of Dr Katlyn Turner, an incoming professor in the Department of Information Science at the University of Colorado–Boulder. She told the audience that the key skill students must develop before entering the workforce is critical thinking. This, she said, is not a skill that she believes AI helps develop.
Panel moderator Dr Kathryn Huff agreed. Huff, new chair of the Nuclear Engineering and Engineering Physics at the University of Wisconsin–Madision, noted that unobserved assessments are worthless if the observer cannot distinguish whether they were completed by students or by AI. “Taken together, their thoughts pointed to a general air of concern over students developing the requisite skills necessary to be equipped for the nuclear industry,” concluded the Nuclear Newswire report.
The AI in nuclear debate continues at WNE 2027, which runs 7–9 December 2027 at Paris Villepinte Parc des Exposition. Exhibitor registration for this key industry event is now open.
At the WNE 2023 conference, Emilie Sauchay, HR Manager of Ineo Nucléaire, discussed their strategic workforce planning for 2022 and 2023. She noted the recruitment of skilled workers as a significant factor but emphasised the necessity of making training programs available for professionals outside of the industry.
“For me, the most important one is to be able to train people in our professions, to be able to attract employees who are not currently in the sector but who would like to join this dynamic sector,” Emilie Sauchay stated. She also added that to support the growth of their workforce, Ineo Nucléaire developed training academies and a network of training centres.