US-based nuclear fusion firm Commonwealth Fusion Systems (CFS) has collaborated with Google’s DeepMind division to use artificial intelligence to fine-tune and enhance the operation of its forthcoming SPARC fusion reactor.
The partnership aims to accelerate the path to commercial fusion energy by using AI to simulate and improve control of the superhot plasma that will fuel the machine.
The plan involves using DeepMind’s specialized, open-source software known as Torax to model the plasma inside the SPARC reactor, which is located at CFS headquarters in Devens, Massachusetts.
By pairing Torax with advanced AI models, the companies will explore the most effective ways to run the machine, potentially making future commercial fusion power plants, called ARC, more efficient and economical.
“It’s a big deal. AI tools carry serious potential to accelerate learning on SPARC and then to push toward making our ARC fusion power plants more efficient and economical,” said the energy firm in a press release.
“And that acceleration could dramatically boost fusion energy’s positive impact on the world.”
Building on existing relationship
This collaboration builds on an existing relationship between the two companies. Google is an investor in CFS and has previously agreed to purchase 200 megawatts of power from the first ARC power plant, expected in the early 2030s.
The AI approach will use reinforcement learning, a technique famously used by DeepMind’s AlphaGo program. The AI will explore numerous operational settings for SPARC—such as fueling rates, radio-frequency heating, and electrical currents in its powerful electromagnets—and will be “rewarded” for finding configurations that maximize performance while maintaining safe operational limits.
This is not DeepMind’s first venture into fusion research. The AI team previously used reinforcement learning to successfully control the magnets and explore new plasma configurations in the Tokamak à Configuration Variable (TCV) research machine at the Swiss Plasma Center.
Key component of project
For SPARC, the AI could be used in advance planning to map out operations or be integrated directly into the control system during live operations. It could also help manage the intense heat exhaust in a region of the machine called the divertor by precisely controlling the plasma’s shape.
The Torax software is a key component of the project. Released in 2024, it provides a modern framework that can link multiple AI tools and computing engines, greatly speeding up complex plasma simulations that previously relied on a patchwork of different software.
CFS is developing a more compact and economical type of tokamak reactor by using innovative high-temperature superconducting magnets, which can create much stronger magnetic fields than conventional designs.
The company believes that applying advanced AI will add another layer of computing power to further accelerate the development of clean, affordable fusion energy.
“CFS has already benefited from sophisticated modeling to design SPARC,” concluded the press release.
“Now AI could add another layer of computing power — and another avenue to take fusion energy to the next level.”