Nuclear fusion, if it can be made to work at scale, could provide an important new low carbon energy source, so it is little surprise that the big tech companies are keen to invest in and support its development, as they look for ways to supply their power-hungry data centres.
Fusion combines atoms to produce energy, in contrast to the fission used in today’s nuclear plants which splits atoms – and it could produce many times as much energy as fission from the same amount of fuel without generating long-lived nuclear waste. But the technology has yet to be perfected, is proving complicated to get right and is costing billions of dollars to develop.
Nevertheless, there is no shortage of effort being applied to making it work. A November 2024 survey from the Clean Air Task Force (CATF), an environmental organisation, showed that the number of private fusion companies had increased from 10 in 2017 to more than 40 firms, with some $7bn invested in them since 2020.
Google is among the tech companies supporting the drive to make fusion commercially viable, recently entering into a partnership with Bill Gates-backed Commonwealth Fusion Systems (CFS) to develop, deploy and scale the technology. Google agreed to take 200 megawatts of electricity from CFS’ first plant in the US state of Virginia, from which CFS expects to be supplying grid electricity in the early 2030s.
The tech giant has been investing in the fusion developer since 2021, a year in which the CFS succeeded in raising a total $1.8bn from investors. Google also recently announced it was investing in another fusion developer, California-based TAE Technologies.
Two years ago, Microsoft signed an agreement to take power from Helion, a fusion developer backed by Sam Altman, by as early as 2028. The current timetable for the project remains unclear, but Helion maintains that its will be the first fusion power plant.
But tech firms aren’t just contributing funding – the AI applications they are generating are also helping to speed up progress towards mass-market fusion power, including all the unglamorous but essential tasks needed to get the ball rolling.
Earlier this year, Microsoft reported on how its agentic AI tools were helping the people running the International Thermonuclear Experimental Reactor (ITER) fusion project, located in France, to speed up administrative chores, for example through development of a chatbot to search the project’s database of over 1 million documents.
Modelling experiments
Also significant for the project’s future is an agreement between ITER and Microsoft Research to collaborate on research into modelling future experiments at ITER. The aim is to strengthen the modelling of gas plasmas and carry out analysis and processing of the vast amount of data generated by the project. Alain Becoulet, deputy director general of ITER, said the collaboration could result in “a major leap forward in ITER’s ability to optimize its operational regime”.
In May, ITER also said it had completed all the components needed to assemble the electromagnet system required to contain the plasma used in its reactor. This represents significant progress for a multi-country project that Is already several years behind its originally envisaged schedule.
Meanwhile, Amazon has highlighted the use of its AWS cloud infrastructure to support fusion technology, for example by supporting US developer Realta Fusion to run its simulations.
Numerous academic institutions around the world are also using AI to hone techniques used to develop fusion power. In one recent example, Eindhoven University of Technology (TU/e) reported on how one of its research students, Yoeri Poels, had used machine learning to create faster simulation models and developed more robust monitoring systems capable of detecting subtle changes in plasma.
The CATF report also provides illustrations of how AI and high-performance computing are already driving progress in the fusion sector.
Doubts remain over whether any of the current fusion projects under development will be able to scale up their operations at the pace they are projecting – there are plenty of scientists who believe that commercialisation could be decades, rather than years, away. But investment continues to flow into the sector and if developers’ ambitions are realised, the start of a further radical reshaping of the global energy mix, profoundly affecting the hydrocarbons industry, could be just round the corner.
(Photo: Shutterstock)
