The archive · Climate & Energy · Technical decision · 2022–2026
Realta Fusion's magnetic-mirror bet: direct electricity demo lights a bulb in 2026
Realta bets HTS magnets revive the magnetic mirror and direct conversion beats steam; after WHAM first plasma in 2024, its June 2026 demo powered lightbulbs.
Realta Fusion
What the business is
Realta Fusion is a Madison, Wisconsin spinoff of University of Wisconsin-Madison fusion research, commercializing WHAM-style magnetic-mirror machines for electricity and industrial heat, with a prototype reactor called Anvil as its next step.
Starting capital：$9M seed led by Khosla Ventures plus a $36M Series A led by Future Ventures in May 2025; raising a new round as of June 2026
How it started
WHAM (Wisconsin HTS Axisymmetric Mirror) began in 2020 as a UW-Madison, MIT and Commonwealth Fusion Systems partnership funded with $10M in DOE grants. Realta spun out of UW-Madison - co-founded by physicists Cary Forest and Jay Anderson with Kieran Furlong as CEO - to commercialize mirror fusion and took over WHAM as a public-private partner; in June 2023 the DOE selected Realta for milestone-based fusion funding. WHAM achieved first plasma on July 15, 2024 and set a record magnetic field for magnetically confined plasma within its first weeks of operation.
What happened
Realta raised a $9M seed led by Khosla Ventures, then a $36M Series A led by Future Ventures in May 2025 (with Avila VC, GSBackers, Khosla, Mayfield, SiteGround, TitletownTech and the Wisconsin Alumni Research Foundation), targeting power costs of $100 per MWh falling toward $40. In April 2026 Commonwealth Fusion Systems - the company that built WHAM's magnets - announced what it called its largest deal of this kind to date, supplying HTS magnets to Realta. On June 19, 2026 Realta powered lightbulbs using electricity harvested by a prototype converter attached to WHAM, which it called the first private direct-electricity demonstration from a fusion device; TechCrunch noted WHAM does not yet run deuterium-tritium fuel, so the converter harvested input power rather than alpha particles. Furlong estimated direct conversion at about 90% efficiency versus roughly 33% for steam turbines, with recirculated electricity potentially adding 20-30% to a commercial plant's output.
How it ended up
Still running: Realta has demonstrated first plasma, a record confinement field and a direct-conversion proof of concept, but not a fusion-grade reactor; finalizing the Anvil prototype design and raising a new round are the immediate next steps.
Background
Realta Fusion is a Madison, Wisconsin startup spun out of University of Wisconsin-Madison fusion research to commercialize the magnetic mirror, a reactor concept abandoned for decades because powerful end magnets could not stop plasma leaking out. Its bet is that a new component changed the physics: high-temperature superconducting magnets that are tens of times stronger than anything mirror researchers had before.
Realta took over WHAM, the Wisconsin HTS Axisymmetric Mirror, which began in 2020 as a UW-Madison, MIT and Commonwealth Fusion Systems partnership with $10M in DOE grants. WHAM achieved first plasma on July 15, 2024 and set a record confinement field within weeks. Realta raised a $9M seed led by Khosla Ventures and a $36M Series A led by Future Ventures in May 2025, targeting $100 per MWh falling toward $40; in April 2026, Commonwealth Fusion Systems agreed to supply it with HTS magnets in what CFS called its largest deal of that kind to date.
On June 19, 2026 Realta powered lightbulbs with a prototype converter attached to WHAM, saying it was the first private company to demonstrate harvesting electricity directly from a fusion device. TechCrunch's caveat was precise: WHAM does not yet run deuterium-tritium fuel, so the converter harvested input power, not fusion alpha particles. Realta estimates direct conversion at roughly 90% efficiency versus about 33% for steam turbines, and plans to use it to recirculate electricity for heating plasma. The company is finalizing the design of its Anvil prototype and raising a new round.
What has to be true
- Mirror machines failed because magnets leaked plasma; HTS magnets tens of times stronger removed the constraint, letting Realta revive a 60-year-old concept with new physics.
- The axisymmetric mirror scales by adding middle sections whose magnets are weaker and cheaper, attacking fusion's capital-cost problem rather than only its plasma physics.
- Direct electricity conversion targets the other half of the economics: roughly 90% estimated efficiency against about 33% for steam, plus recirculated power for plasma heating.
- Public-private structure carried it early: DOE milestone funding in June 2023, the WHAM partnership, then private capital and a CFS supply deal as milestones landed.
What can be applied
When a technology failed for decades, look for the component that changed underneath it: HTS magnets revived the mirror; direct conversion attacks the same cost problem from the thermodynamics side.
Aftermath
As of July 1, 2026, Realta Fusion is raising a new round and finalizing the design of Anvil, its first purpose-built prototype reactor, after its June 19, 2026 demonstration powered lightbulbs from a direct-conversion converter on WHAM. WHAM remains a demonstration device: it does not run deuterium-tritium fuel, so the converter harvested input power, and no fusion-grade reactor or grid connection exists. Its commercial targets are $100 per MWh initially, falling toward $40, with direct conversion adding an estimated 20-30% to plant output by recirculating electricity.
Sources
- Realta Fusion generates electricity directly from a fusion reaction, an apparent first
- Realta Fusion taps $36M in fresh funds for its fusion-in-a-bottle reactor
- Commonwealth Fusion Systems leans on magnets for near-term revenue
- First plasma marks major milestone in UW-Madison fusion energy research
spotted an error? The archive wants to know.
Your turn
You just read one. Describe what you are building, and see who is betting on the same thing.
Free account · 3 free questions · no card