US-based start-up Commonwealth Fusion Systems (CFS) is to collaborate with AI infrastructure company NVIDIA and Germany-based global technology company Siemens to develop a digital twin of its SPARC fusion machine. The digital twin will apply artificial intelligence (AI) and data and project management tools to accelerate commercial fusion.
Since it was spun-out from the Massachusetts Institute of Technology (MIT) in 2018, CFS has raised almost $3bn in capital, which it plans to use to complete its SPARC demonstration tokamak under construction in Devens, Massachusetts. It will also progress development work on its planned 400 MWe ARC fusion power plant in Chesterfield County, Virginia, with grid connection expected in the early 2030s.
In April, CFS said it had begun assembling its SPARC tokamak, which had been under development for years. CFS began by manufacturing SPARC’s superconducting magnets based on large-scale prototypes. CFS has already installed ancillary equipment around SPARC. This includes the systems to power and cool the tokamak’s super-strong magnets, the diagnostic sensors to monitor the fusion process, and the heating system to turn SPARC’s hydrogen fuel into a plasma for the fusion process.
The digital twin will leverage troves of data from the Siemens Xcelerator portfolio of industrial software, including its Designcenter NX for advanced product engineering and Teamcenter product lifecycle management (PLM) tools, which CFS uses to create, catalogue, and process machine designs and assemblies. These designs and assemblies can then be used in CFS’s modelling and simulation workflows, including the layering of AI-enabled tools.
CFS will use NVIDIA Omniverse libraries and OpenUSD to integrate data with classical and AI-powered physics models to create the digital twin of SPARC. This virtual replica of SPARC will provide CFS with a user-friendly way to run simulations, test hypotheses, and quickly compare the experimental results from the machine to the simulations. This ability to rapidly analyse data and iterate will speed CFS’ efforts to make fusion energy a commercial reality.
“CFS will be able to compress years of manual experimentation into weeks of virtual optimisation using the digital infrastructure developed by NVIDIA and Siemens,” said CFS Founder & CEO Bob Mumgaard. “Through this collaboration, we’re demonstrating how AI and integrated digital engineering can accelerate progress from design to grid power. This will allow us to transform how we build and operate fusion machines in the race to commercial fusion.”
Rev Lebaredian, Vice President, Omniverse and Simulation Technology at NVIDIA, noted: “Delivering commercial fusion demands that we simulate and solve incredibly complex physics problems,” said. “By using Siemens NX software and NVIDIA Omniverse libraries to create a high-fidelity digital twin of SPARC, CFS will be able to accelerate its engineering and shorten the timeline to clean power.”
CFS is also using Siemens’ digital tools to improve the efficiency of its manufacturing processes and operations at the company’s magnet factory in Devens. “By connecting Siemens Xcelerator with NVIDIA AI visualisation libraries, we’re demonstrating that end-to-end digital workflows aren’t just efficient, they’re transformative,” said Del Costy, President & Managing Director, Americas, Siemens Digital Industries Software. “Fusion is complex, but data doesn’t lie. When you aggregate real manufacturing intelligence, apply AI, and run thousands of scenarios, you remove guesswork and accelerate innovation. This is the future of industrial engineering.”
During the 2026 Consumer Electronics Show (CES), which opened on 6 January 6 Las Vegas, CFS said it had installed the first of 18 magnets in its SPARC fusion reactor. Key components of SPARC’s magnets have been completed, and the company expects to install all 18 by the end of the summer, according to Bob Mumgaard.
When installed, the D-shaped magnets will sit upright on a 24-foot wide, 75-tonne stainless steel circle known as a cryostat, which was set in place in March 2025. The magnets weigh about 24 tonnes each and can generate a 20 tesla magnetic field, about 13 times stronger than a typical MRI machine – strong enough to lift an aircraft carrier,” Mumgaard said.
To achieve that strength, the magnets will be cooled to -253 degrees C so they can safely conduct over 30,000 amps of current. Plasma will be burning at more than 100m degrees C. The digital twin will be used to work out as many kinks as possible before SPARC is turned on. CFS aims to run experiments or tweak parameters in the digital twin before applying them to Sparc itself. “It will run alongside so we can learn from the machine even faster,” Mumgaard said.
