The real reason Google DeepMind is working with a fusion energy startup

Energy startup Commonwealth Fusion Systems (CFS) said Thursday it is working with Google’s DeepMind division to optimize — and even improve — the operation of its prospective Sparc reactor using artificial intelligence.

The companies’ plan will simulate the plasma that would burn inside a CFS reactor using specialized DeepMind software known as Torax. They also plan to pair Torax with AI models to help CFS figure out how best to achieve fusion power.

Fusion energy promises to provide massive amounts of electricity with zero emissions from a nearly limitless fuel source: water. AI-related companies have been bullish on fusion startups as a source of electricity to power power-hungry data centers. Google seems to be eyeing them as potential customers as well.

This is not Google’s first experiment in nuclear fusion. The technology company worked with another fusion startup, TAE Technologies, to use artificial intelligence to study how plasma behaves inside TAE’s fusion machine.

There’s a reason Google keeps coming back to the problem: AI may be uniquely suited to making fusion power possible.

One of the biggest challenges facing fusion startups is keeping the plasma inside the reactor hot enough for long enough. Unlike nuclear fission reactions, which are self-sustaining, fusion reactions are difficult to sustain outside stars like the Sun. Without this kind of mass and gravity, the plasma is constantly at risk of spreading out and extinguishing itself.

In CFS reactors, powerful magnets replace gravity to help trap the plasma, but they’re not perfect. Reactor operators must develop control software that can enable the device to continuously react to changing plasma conditions.

The problem is that there are too many knobs to turn, certainly more than a human can handle. This is the type of problem at which AI excels. Experts have cited AI as one of the key technologies that has enabled remarkable progress in the industry over the past several years.

CFS is currently building its experimental Sparc reactor in a suburb of Boston. The device is roughly two-thirds complete, and when completed later in 2026, the startup expects it will be the first fusion device capable of producing more power than the factory needs to power itself.

Google said Torax can be used with reinforcement learning or evolutionary search models to find “the most efficient and robust paths to net energy generation.” The two companies are also exploring whether artificial intelligence can be used to control reactor operation.

In August, Google participated in the $863 million CFS round B2 alongside Nvidia. Earlier this year, Google also said it would purchase 200 megawatts of electricity from CFS’s first commercial power plant, Arc, which is scheduled to be built outside Richmond, Virginia. The technology company is also an investor in CFS competitor TAE Technologies.