$130 million in private capital. A $20 million federal award on top. No fusion plant has put a watt on the grid yet.
Thea Energy, a Princeton spinout, is betting the gap between those facts closes with magnets, not physics.
The company has raised $130 million in private capital, including a $100 million Series B in May, and holds a Department of Energy (DOE) certification for its Helios pilot plant preconceptual design.
The thesis: move stellarator complexity from precision mechanical fabrication into software-controlled arrays of identical planar coils.
Magnets were always the bottleneck
Magnetic-confinement reactors hold plasma above a hundred million degrees using fields generated by superconducting magnets. The magnet array is the costliest part of the machine. It is also the hardest to build.
Most fusion companies treat that as a supply-chain problem. Order the magnets. Wait. Assemble.
Thea treats it as a factory problem.
Its stellarator uses planar coils. Flat, modular superconducting sheets, each controlled by software, arranged around the reactor like pixels on a display. Conventional stellarators need three-dimensional coils machined to exacting tolerances, which is why stellarator history is littered with cost overruns. It moves the complexity out of precision mechanical fabrication and into control software.
That is the whole bet.
What the money buys
The $20 million came through ARPA-E's SCALEUP program, which funds previous awardees as they move toward commercial adoption. Announced July 27, it pays for manufacturing and test equipment for the first domestic production line of modular HTS magnets.
Read that sentence again. It buys a production line for magnets, and the magnet is where fusion's cost problem actually lives.
Magnet manufacturing line
First domestic production line for modular HTS magnets, in New Jersey. · Thea Energy, July 2026
The financing picture stacked up through the year. A $100 million Series B closed in May, led by Thomas Tull's US Innovative Technology Fund, with General Innovation Capital Partners and Linse Capital participating. The DOE certified the Helios pilot plant preconceptual design in January, the first company in the Milestone-Based Fusion Development Program to receive that distinction.
Eos stellarator design
First large-scale integrated stellarator, targeted for 2030. · Thea Energy, 2026
Milestones line up behind the money. A nine-coil prototype called Canis ran at 20 kelvin and produced fields within 1% of predictions. A full-size Eos-spec coil passed 6 tesla in May. Eos, the first integrated stellarator, is designed around roughly 300 of these coils, with a 2030 target.
Our magnets are designed to be mass manufactured and in just a couple of years, we've iterated these coils over a hundred times, finalized the design, and derisked this core technology, positioning us for this next phase, manufacturing scale up.— Brian Berzin, co-founder and CEO, Thea Energy
The counterweight
The company has been straight about the timeline. TechCrunch reports Thea plans a commercial-scale plant in the mid-2040s, in line with the rest of the industry. Eos in 2030 is an integration test, not a power plant.
Stellarators carry a long history of engineering pain. Wendelstein 7-X in Germany, the largest stellarator ever built, took decades and ran over budget before experiments began. Thea's answer is that its architecture removes the source of that pain.
The counterweight is manufacturing credibility. One prototype line is not a supply chain. No HTS magnet producer has yet delivered at the scale fusion needs. It has not disclosed the line's annual capacity, its cost targets, or a commissioning date.
Why the magnet, not the reactor, decides fusion's economics
Confirmation criteria: disclosed line capacity, unit cost per coil, and a coil yield high enough to support Eos.
Who is paying attention
Thea says it is in discussions with over a dozen power offtakers, hyperscalers, and utility partners. That list matters as much as the technical milestones. Fusion's commercial logic depends on buyers willing to sign long-term contracts for firm baseload power before the plants exist.
The buyer side is already moving. Hyperscalers committed roughly 9.8 GW of nuclear capacity through 2026 power purchase agreements, mostly restarts and large-reactor deals. That wave validated the demand for firm, carbon-free baseload. It also showed the market will commit to power that arrives years from now.
Will a production-line magnet reset fusion economics this decade?
Probability: 40%. Modular coil manufacturing attacks the right cost, but no reactor-scale integrated stellarator has run yet.
✅ Arguments for
A $130 million private war chest plus $20 million federal plus a DOE certification gives it runway most peers lack.
Offtaker discussions with hyperscalers align with the firm-baseload procurement wave of 2026.
Confirmation criteria: Eos operating on target and HTS line capacity disclosed.
❌ Arguments against
A single prototype line is not a supply chain, and the company has not published capacity or timing.
First power in the mid-2040s means more than a decade of dilution and execution risk before revenue.
Disconfirmation criteria: Eos slips, magnet yield disappoints, or offtakers move to existing reactors.
Eos site selection, expected later this year
HTS line capacity, unit cost, and commissioning date disclosures
Full-size coil test results at reactor-relevant field strength
New offtake agreements with hyperscalers or utilities
Development scenarios
🟢 Optimistic scenario (25%)
Implications: It becomes the magnet supplier to the industry, an asset business rather than a reactor developer.
🟡 Base-case scenario (55%)
Implications: Continued dilution and a timeline past 2030, while the manufacturing thesis survives.
🔴 Pessimistic scenario (20%)
Implications: It retrenches to a component supplier or runs out of runway before first power.
For an investor, it is a bet on manufacturing being the binding constraint. Whoever industrializes the magnet captures the margin in a market that hyperscalers have already decided to fund.