Commonwealth Fusion Systems (CFS) announced Tuesday at CES 2026 that it has installed the first magnet in its demonstration Sparc fusion reactor, which is scheduled to go online next year.
The magnet is the first of 18 that will generate a powerful magnetic field to create a donut-like shape that confines and compresses superheated plasma when the reactor is completed. If all goes well, the plasma releases more energy than is needed for heating and compression.
After decades of promises and delays, fusion power appears to be just around the corner. CFS and its competitors are in a race to deliver the first electrons to the power grid sometime in the early 2030s. If it works, fusion power could unlock nearly limitless clean energy in a package similar to traditional power plants.
The main components of Sparc’s magnets are complete, and the company plans to have all 18 in place by the end of the summer, said Bob Mumgaard, CFS co-founder and CEO. “It will continue bang, bang, bang throughout the first half of this year as we bring together this innovative technology.”
Once installed, the D-shaped magnet will sit upright on a circular stainless steel cryostat that is 24 feet wide and weighs 75 tons. This circular cryostat was installed last March. The magnets themselves weigh about 24 tons each and can generate a 20 Tesla magnetic field, about 13 times stronger than a typical MRI machine. “This is the type of magnet that could be used to lift an aircraft carrier,” Mamgaard said.
To reach this strength, the magnets are cooled to -253 °C (-423 °F) and can safely carry currents in excess of 30,000 amps. Inside the donut, plasma burns at over 100 million degrees.
To iron out as many glitches as possible before Sparc goes online, CFS announced Tuesday that it is working with Nvidia and Siemens to develop a digital twin of the reactor. Siemens provides design and manufacturing software that helps the company collect and feed data into Nvidia’s Omniverse library.
tech crunch event
san francisco
|
October 13-15, 2026
This is not CFS’ first simulation — the company already runs a number of simulations to predict the performance of various parts of the reactor — but existing efforts are providing results on their own, Mumgaard said. Regarding digital twins, he said, “These are no longer isolated simulations that are only used for design. They are parallel to the physical one from start to finish and are constantly compared to each other.”
The hope is that CFS will be able to run experiments and fine-tune parameters within the digital twin before applying it to Sparc itself. “It runs in parallel, so we can learn from the machine even faster,” he said.
Sparc was expensive to build. CFS has raised nearly $3 billion to date, including an $863 million Series B2 round in August that included investments from Nvidia, Google, and nearly 30 other investors. The company’s first commercial-scale power plant, “Ark,” will be the first. CFS estimates that this could cost billions of dollars more.
Mumgaard hopes digital twin and AI technology will help the company get fusion power onto the grid sooner. “As the machine learning tools improve and the representation becomes more accurate, we will see it get even faster, which is a good thing because there is an urgent need to get the fusion to the grid,” he said.
See TechCrunch’s full coverage of the annual CES conference here.
