Thea Energy's Helios Design Turns Fusion Into a Manufacturing Problem

Fusion energy has spent four decades as the technology that is always thirty years away. Thea Energy is making a narrower promise: that its first commercial plant will not be a physics problem at all.

On September 9, the New Jersey company published sixteen peer-reviewed papers on Helios, a stellarator power plant designed to deliver roughly 400 megawatts of electricity to the grid. The work ran as a special issue of Fusion Engineering and Design. It describes manufacturing and maintenance more than plasma, and that emphasis is the argument.

What the design actually claims

The specification is unusually concrete for early fusion. Helios is designed for 400 MW of net electrical output and 1.1 GW of total thermal power, wrapped around a plasma with a major radius of 8 meters. The company calls it the most compact optimized stellarator power plant architecture on record.

Between the plasma and the magnets sit 1.2 meters of shielding, enough that the coils should survive the life of the plant. A tokamak-style X-point divertor handles exhaust. Maintenance is organized by sector, so one slice of the machine opens while the rest stays sealed.

Each choice serves a single goal. It is making a fusion plant that can be financed, insured and serviced like infrastructure rather than curated like a laboratory. The stellarator, in other words, has to become a product.

Why the magnets decide everything

The stellarator has always been fusion's elegant branch. Its twisted magnetic field holds plasma in a steady state, with none of the current-driven disruptions that can damage a tokamak. The catch was hardware.

Producing that field required complex three-dimensional coils, each one a bespoke object, each one difficult to build and harder to replace. Engineers spent decades describing the stellarator as the machine that would work perfectly if only someone could manufacture it.

The company's answer moves the complexity out of the metal and into software. Its magnets are flat planar coils, mass-manufacturable and individually adjustable. The control system shapes the field in real time instead of freezing an intricate geometry at the factory.

Thea Energy Helios design papers, Fusion Engineering and Design, 2026

The company demonstrated a full-scale superconducting planar coil array before this publication. The paper set builds on that hardware result rather than replacing it, which is why the design reads as an engineering program and not a thought experiment.

The Helios power plant is technically viable, robust, and requires no scientific miracles to commercialize.— David Gates, co-founder and chief technology officer, Thea Energy

Peer review is a filter, not a warranty. Sixteen papers clearing expert review means the physics and the engineering hold together on paper. It does not mean a plant will run.

The capital and the clock

Thea has disclosed $140M across three rounds. A $20M Series A in February 2024. A $100M Series B in May 2026 led by Thomas Tull's US Innovative Technology Fund, with Hitachi Ventures, Prelude Ventures and Lowercarbon Capital also on the cap table. An extension in August 2026, joined by Brevan Howard Macro Venture and others, at an undisclosed size.

Public money carries part of the load. In July the company won a $20M ARPA-E SCALEUP award to expand magnet manufacturing lines, the same unglamorous capacity the design depends on. We covered that grant when it landed: as we wrote in July, the magnet factory, not the plasma, was the constraint the company was trying to relieve.

The rest of the field is not standing still. Commonwealth Fusion Systems, Helion Energy and Pacific Fusion are pursuing rival architectures with far deeper pockets, which sets the bar Thea has to clear on cost rather than on physics.

The near-term schedule runs through Eos, a large-scale integrated stellarator meant to demonstrate power-plant-relevant, steady-state fusion. The company says Eos will be online by 2030 and is choosing among five states for the site, with an announcement expected this year. The commercial plant follows in the 2030s.

That schedule is the risk. Fusion timelines have slipped for fifty years, and every company in the sector carries a version of this chart. Its edge is that the date rests on factory throughput and construction, quantities that can be measured, rather than on a plasma result that may or may not arrive.

What the papers do not prove

Three questions stay open.

Net energy gain. Eos must reach a triple product of 10²¹ keV·s/m³, the benchmark that combines density, temperature and confinement time. Until it does, the plant is only a design.

Cost. The papers validate feasibility. They do not yet show a cost per megawatt-hour that competes with solar, wind and grid storage, all of which keep getting cheaper.

Execution. A design certified on paper still has to be built, licensed and connected to a grid already straining under data-center demand.

Can Thea hold the 2030s date?

Signals to watch

Development scenarios

Sources