Commonwealth Fusion Systems Taps NVIDIA and Siemens for AI-Driven Digital Twin of SPARC Fusion Machine

New updates have been reported about Commonwealth Fusion Systems.

Commonwealth Fusion Systems (CFS) has entered into new collaborations with NVIDIA and Siemens to create an AI-enabled digital twin of its SPARC fusion machine, a move aimed at compressing development timelines and accelerating the path to commercial fusion power. Using Siemens’ Xcelerator industrial software portfolio—including NX for advanced engineering design and Teamcenter for product lifecycle management—CFS will aggregate and structure large volumes of design, manufacturing, and operational data into a unified environment that can feed high-fidelity simulations and AI models. This digital twin, which will be visually showcased at CES 2026, is intended to let CFS rapidly test design and operating hypotheses, compare real-world experimental data from SPARC to simulations, and iterate much faster than traditional physical prototyping would allow.

The project will also leverage NVIDIA Omniverse libraries and OpenUSD to integrate classical physics models with AI-powered simulation, giving CFS a virtual replica of SPARC capable of running thousands of scenarios in parallel. According to CEO Bob Mumgaard, this infrastructure is expected to convert years of manual experimentation into weeks of virtual optimization, potentially shortening the company’s timeline to grid-scale fusion power and transforming how future fusion machines are designed and operated. Beyond SPARC, CFS is already applying Siemens’ digital tools to enhance manufacturing efficiency at its magnet factory in Devens, Massachusetts, building a data-rich operational base that can further improve accuracy and speed of simulations. With nearly $3 billion raised since its 2018 founding, CFS is positioning these partnerships as core enablers of its strategy to be first to market with net-energy, carbon-free fusion and to scale the technology into a commercially viable power platform.