Thea Energy completes Amphion quench test and raises $100M

The circular, planar geometry of the Amphion coil is standardised across every shaping coil planned for Eos, enabling assembly-line production rather than the bespoke fabrication historically required for stellarator magnet systems

(Image courtesy of Thea Energy)

Thea Energy‘s first full-size, full-current, full-field planar shaping coil, the Amphion, reached a magnetic field above 6 T at 20 K earlier this year, meeting stated performance requirements for a power plant-relevant, steady-state fusion system. Subsequent quench survivability campaigns, which test a superconducting magnet’s ability to safely manage a sudden loss of superconducting state, have since concluded. Brian Berzin, co-founder and chief executive officer, said the results give the team confidence to proceed. “We have now completed the quench survivability campaigns and we have the successful results to be confident in scaling manufacturing of these coils for Eos,” he said.

The Amphion result follows more than 50 design iterations carried out entirely in-house at Thea Energy’s Kearny, New Jersey headquarters over the past 18 months. Because every shaping coil in the Eos array shares a single circular, planar geometry, the manufacturing path treats each unit as a repeatable part rather than a bespoke component. Additional test campaigns have also shown the software control systems can dynamically tune out hardware variability, assembly errors, and wear during operation. Thea Energy says the completed test programme verified both magnet performance and manufacturing capability, and that the company is already scaling its manufacturing infrastructure for Eos and subsequent Helios power plants.

Software carries the precision load

The deeper technical claim underpinning Thea Energy’s architecture is that the system shifts precision requirements from hardware to software. Earlier work on the Canis 3×3 planar coil array, published in IEEE Transactions on Applied Superconductivity, demonstrated closed-loop field control to within 1% of predicted values at 20 K. To stress-test that control capability, the team deliberately mismanufactured magnets and ran the software compensation system against the resulting variability. “Our software control systems were capable of analysing various types of off-nominal hardware and then successfully tuned out that variability via software level controls,” he said.

Berzin explains the system has 300 independent variables to tune magnetic field performance. He claims the architecture tolerates an order of magnitude more hardware variance than conventional three-dimensional modular coil stellarators, with allowable manufacturing tolerances running to multi-centimetres against sub-millimetre precision requirements for 3D coil designs. The approach also relies on basic manufacturing processes and traditional workforces rather than specialist fabrication, with Berzin explaining this supports easier systems integration, lower maintenance downtime, and reduced operational costs compared to prior modular coil stellarators. “We can have significant manufacturing economies of scale with core components made on an assembly line similar to car manufacturing, rather than a boutique laboratory,” he said. Those figures and comparisons represent the company’s stated position and have not been independently verified. The peer-reviewed Canis results confirm that closed-loop field control at 1% precision is achievable in a planar array at prototype scale.

DOE milestone and the commercial pipeline

In January 2026, the US Department of Energy certified Thea Energy’s preconceptual design for Helios, making it the first company to complete the final major milestone in Phase 1 of the DOE’s Milestone-Based Fusion Development Programme. Thea Energy says the review involved a multi-hundred-page design briefing report, in-person sessions at Kearny, and a panel of independent experts drawn from US national laboratories. “The DOE certification of our Helios power plant design review substantiates the credibility of our approach to commercial fusion power,” Berzin said. The company says it is in discussions with more than a dozen offtakers, hyperscalers, utility partners, and suppliers, a result the company says has directly accelerated those conversations.

The $100 million Series B was led by Thomas Tull’s US Innovative Technology Fund, with General Innovation Capital Partners and Linse Capital each making their first investment in fusion. The round also drew participation from a broader syndicate of new investors alongside existing backers including Lowercarbon Capital, Hitachi Ventures, and Prelude Ventures. The capital will fund expansion of Thea Energy’s magnet manufacturing infrastructure, including a second facility in Northern New Jersey due to open before the end of 2026, and support Eos siting and construction.

What Eos is being built to demonstrate

Site selection for Eos is expected before the end of 2026. Berzin says the physical and infrastructure requirements are less technically constrained than might be expected for a fusion device, and are not unique compared to other proposed fusion systems or other types of energy projects. Key criteria include high-capacity grid access, community support, and proximity to established supply chains. The facility footprint runs to 20 to 30 acres using traditional concrete structures as bioshielding. The neutronics engineering is well understood and does not significantly complicate site selection, he adds. Eos is targeted to be operational by 2030.

Eos is designed to de-risk Helios, Thea Energy’s planned first commercial fusion power plant, by demonstrating that the company can build fusion systems faster and for less capital than prior approaches. Eos serves as the prototype for the first commercial power plant and for scaling what the company describes as an abundant source of clean, safe fusion energy. Performance targets include steady-state operation with pulse lengths of up to 24 hours, with Thea Energy saying Eos will achieve world-leading fusion performance while also de-risking the systems integration, construction, and execution required to build fusion power plants at scale. For context, Wendelstein 7-X set a world record for the triple product in long-duration stellarator operation with a 43-second high-performance plasma discharge in May 2025. Thea Energy’s planar coil architecture approaches plasma confinement differently to W7-X and the comparison is not direct, but the gap between current demonstrated stellarator performance and Eos targets gives some measure of what the construction and operational programme needs to deliver. The company is progressing toward starting construction of the first Helios power plant before the end of the decade.

Stay ahead in the fusion revolution explore more breakthroughs from leading innovators in clean energy technology.