Realta Fusion, Commonwealth Fusion Partner to Build HTS Magnets for Mirror Reactors
CFS and Realta's multi-billion-dollar HTS magnet deal bets that magnet supply, not plasma physics, will decide which fusion concepts reach the grid first.

The gap between fusion concept and fusion product now runs through a magnet factory in Devens, Massachusetts.
Realta Fusion and Commonwealth Fusion Systems announced a long-term strategic partnership on April 2 to design and manufacture the high-temperature superconducting magnets Realta needs for both its demonstration prototypes and eventual commercial plants. The deal carries a stated potential value in the multi-billion-dollar range, and Rick Needham, CFS's chief commercial officer, called it "the largest deal of this kind to date for CFS."
The physics of Realta's CoSMo fusion™ approach makes the HTS supply question acute. A magnetic mirror machine confines plasma in a linear geometry: two powerful end-plug coils create mirrors that reflect charged particles back into the central column, while a series of coils sustains confinement along the machine's length. The end plugs are the critical bottleneck. WHAM, the ARPA-E-funded experiment at the University of Wisconsin-Madison that gave Realta its origin story, used a pair of flat, thin REBCO-wound mirror coils designed and built by CFS. Those coils reached 17 tesla in the bore and cleared 20 tesla on the conductor itself, a world record for magnetically confined plasma, achieved in July 2024.
For a commercial CoSMo plant, Realta scales by lengthening the central column, not by redesigning the end plugs. That architecture has a direct cost implication: the central coils run at lower field strength than the mirror coils and are therefore cheaper per meter. The end plugs, by contrast, demand the same 17-plus-tesla performance that makes REBCO conductor and CFS's flat-coil geometry non-negotiable. Sourcing those coils from anyone other than the team that already built them for WHAM would require years of qualification work. By locking in CFS now, Realta bypasses that detour entirely.
"Commercializing magnetic mirror fusion systems requires integrating multiple cutting-edge technologies. By working with the world's leading HTS magnet manufacturer, we are significantly de-risking one of the most critical of these technologies," said Realta CEO Kieran Furlong. CFS CEO Bob Mumgaard framed the arrangement from the supply side: "This partnership allows Realta to tap into the world-class supply chain we built to support our advanced manufacturing capabilities, and that will help it to bring commercial fusion energy to the grid faster."

The deal is also the clearest signal yet that CFS intends to build a horizontal magnet business across multiple confinement concepts. In 2025, CFS licensed its HTS cable technology to Type One Energy for the Infinity Two stellarator. The Realta agreement goes further, covering not just technology transfer but the manufacture of finished magnet systems, including a talent-sharing arrangement for deployment and operations support. CFS has now positioned itself as the default HTS supplier for tokamak, stellarator, and magnetic mirror simultaneously, a consolidation that could compress timelines and reduce per-unit costs as production volumes increase across all three programs.
The collaboration itself traces to 2020, when ARPA-E funded WHAM at Wisconsin. Realta spun out of that project in 2022. CFS's WHAM magnets, which the company nicknamed Thing 1 and Thing 2, were delivered to Stoughton, Wisconsin in July 2024 as self-contained turnkey systems, complete with cryogenic refrigeration, vacuum pumping, and controls, all supplied by CFS. First plasma followed within two weeks.
For those tracking Realta's commercialization timeline, the magnet supply constraint just shifted from open question to contracted commitment. The demonstration prototype magnet design and delivery schedule will be the concrete milestone to watch over the next 12 to 24 months, with manufacturing capacity, not physics, now defining the critical path.
Know something we missed? Have a correction or additional information?
Submit a Tip

