US fusion firm Commonwealth Fusion Systems (CFS) has announced partnerships with Nvidia and Siemens to develop a digital twin of its SPARC fusion reactor.

The companies claim they will apply artificial intelligence (AI) and industrial data tools to accelerate the path toward commercial fusion energy.

The announcement was made at CES, where CFS CEO and co-founder Bob Mumgaard appeared on stage alongside executives from Nvidia and Siemens. The SPARC tokamak reactor is a compact fusion device, expected to serve as a precursor to the company’s first commercial reactor. It is located at the company’s headquarters in Massachusetts.

According to CFS, the digital twin system will integrate engineering, manufacturing, and operational data from Siemens’ Xcelerator portfolio, including NX Designcenter and Teamcenter product lifecycle management (PLM) software, already utilized by CFS to design, catalog, and manage SPARC’s complex assemblies. The data will then be fed into CFS’ modeling and simulation workflows, enabling the application of AI-powered tools across the lifecycle of the machine.

In addition, CFS will utilize Nvidia Omniverse libraries and OpenUSD to combine the data with classical physics simulations and AI-enabled models, creating a virtual replica of the SPARC fusion reactor that runs alongside the physical machine. According to the company, the digital twin will allow engineers to run simulations, test hypotheses, and rapidly compare experimental results from SPARC with predicted performance.

“This technology allows us to compress years of manual experimentation into weeks of understanding how these machines work,” Mumgaard said. “Through this collaboration, we’re demonstrating how AI and integrated digital engineering can accelerate progress from design to grid power.”

Unlike traditional simulations, which are typically used only during the design phase, Mumgaard said the digital twin will remain active throughout SPARC’s operation, continuously consuming data from the machine and comparing it with high-fidelity simulations. The company said the approach aims to improve insight into machine behavior, reduce commissioning time, and accelerate iteration toward future commercial fusion power plants.

“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,” said Rev Lebaredian, vice president of Omniverse and simulation technology at Nvidia.

In a media call before the announcement, CFS claimed that the collaboration will complement its existing AI partnerships, including work with Google DeepMind focused on plasma physics and control. The Nvidia-Siemens deal is instead aimed at the broader industrial system, including manufacturing, plant operations, and full-machine integration.

In addition to supporting the development of its SPARC reactor, CFS will also use Siemens’ digital tools to improve manufacturing efficiency and operations at its high-temperature superconducting magnet factory in Devens, Massachusetts. The factory recently completed and installed the first of SPARC’s 18 toroidal field magnets.

“By connecting Siemens Xcelerator with Nvidia AI visualization libraries, we’re demonstrating that end-to-end digital workflows aren’t just efficient, they’re transformative,” said Del Costy, president and managing director, Americas, Siemens Digital Industries Software.

CFS is targeting first plasma from SPARC in 2027, with its first commercial fusion power plant, ARC, expected to deliver electricity to the grid in the early 2030s. The ARC plant has a planned capacity of 400MW, and is tied to a long-term Power Purchase Agreement (PPA) with Google, which last year committed to offtake 200MW from the plant upon its completion.

The ARC project is being developed at the James River Industrial Park in Chesterfield County. Should it reach operational status on schedule, it is expected to be the first fusion project of its kind to deliver power to an electricity grid.

The company, which was founded in 2018, has raised almost $3 billion in funding. Its latest B2 funding round saw it raise $863 million, with backing from Nvidia’s venture capital arm, NVentures.

Unlike fission, which splits heavy atoms, nuclear fusion generates energy by fusing light nuclei. If commercialized, it could provide a vast, nearly limitless energy source. Yet major hurdles remain, namely, high costs, heat management, and plasma confinement, which have pushed most estimates for commercialization beyond the 2040s.

Despite concerns, nuclear fusion has been gaining momentum within the data center industry, with more and more hyperscalers looking to invest in the technology. Google has also backed fusion developer TAE Technologies, which recently raised more than $150m in its latest funding round.