Helion Energy

The thesis: fusion on a data-center clock

Helion Energy is the fusion startup that has bet its existence on a calendar rather than a physics prize. Where most fusion programs frame success as eventually reaching scientific breakeven, Helion has instead sold electricity it does not yet know how to make: in 2023 it signed the world's first fusion power purchase agreement, committing to deliver 50+ megawatts to Microsoft by 2028 with financial penalties if it misses. That contract turned a laboratory ambition into a commercial deadline, and everything the company does now is organized around meeting it.

The wager has grown far larger than Microsoft. The explosion of AI data-center demand has made firm, carbon-free, always-on power the scarcest input in technology, and fusion's promise of dense baseload generation with no fuel logistics maps almost perfectly onto that need. That alignment is why OpenAI is reportedly negotiating to buy fusion power from Helion at gigawatt scale, and why a company that has never sold a kilowatt-hour was valued at $15.5 billion in June 2026. Helion is effectively a bet that the AI industry's hunger for electricity will arrive at the same moment its reactors do.

Source

From starship propulsion to a fusion company

Helion was founded in 2013 by David Kirtley, John Slough, Chris Pihl, and George Votroubek, four scientists who met at MSNW, a University of Washington spinout that studied plasma physics for spacecraft propulsion. The founders' path ran through the idea of using magnetized plasma to push starships before they redirected the same physics toward generating electricity on Earth — a lineage that shaped Helion's unusually engineering-first, hardware-iteration culture rather than a purely academic one.

The company gained early momentum as a 2013 CleanTech Open finalist, a 2014 Y Combinator participant, and the recipient of a 2015 ARPA-E ALPHA contract to compress field-reversed plasmas to fusion conditions. Its trajectory changed decisively in 2021, when its sixth prototype, Trenta, reached 100 million degrees Celsius over a 16-month campaign and Sam Altman led a $500 million Series E with up to $1.7 billion in milestone-based commitments — then the largest private fusion round ever. Altman remains Helion's largest individual investor, a relationship that would later force him to step aside from the OpenAI supply negotiations.

Source

A contrarian machine: pulsed fusion with no steam

Helion's technology diverges sharply from the tokamak mainstream. It uses a field-reversed configuration (FRC) — a self-contained, smoke-ring-shaped plasma that holds its own magnetic field and needs fewer external magnets — combined with pulsed magneto-inertial fusion. Two plasma rings are accelerated toward each other at high speed and slammed together in a long tube; the collision and magnetic compression heat the fuel to fusion temperatures in a brief pulse, repeated over and over rather than sustained continuously.

The more radical choice is how Helion captures the energy. Conventional fusion designs, like conventional fission plants, use the reaction's heat to boil water and spin a steam turbine. Helion skips the turbine entirely: when the fused plasma expands, it pushes back against the confining magnetic field, inducing a current directly in the coils — the same principle as regenerative braking in an electric vehicle. The company claims this direct recovery can exceed 95% efficiency and eliminates the cooling towers and steam loops that make thermal fusion plants resemble large power stations.

That efficiency ambition is tied to an unconventional fuel. Helion's commercial target is deuterium and helium-3, a reaction that runs hotter but produces more charged particles — ideal for direct electrical conversion — and Helion plans to breed its own helium-3 as a fusion byproduct. Skeptics note that no one has ever demonstrated net-electricity fusion this way, and the aneutronic-leaning D-He-3 path is among the hardest fuels to ignite, which is precisely why the company's timeline draws both excitement and doubt.

Source

Polaris and the first private D-T fusion

Helion's seventh-generation prototype, Polaris, began operating in Everett, Washington, in late 2024 and is the machine on which the company's near-term credibility rests. In February 2026 Polaris reached 150 million degrees Celsius and became the first privately developed fusion machine to demonstrate measurable deuterium-tritium fusion — a milestone Helion uses to argue it can operate and scale across multiple fuels on its way to the commercial deuterium-helium-3 mix.

Reaching temperature is only part of the problem; commercial operation requires firing rapidly and recapturing electricity every pulse. Polaris currently pulses roughly once every ten minutes, and Helion needs to push that toward one pulse per second — a repetition-rate leap of several orders of magnitude — while proving it can convert each pulse into recoverable electricity. To iterate faster, Helion is building 'Tiny Merge,' an roughly eight-foot, one-eighth-scale FRC testbed targeted to come online by the end of summer 2026, whose faster experimental loop is meant to de-risk the final Orion design before the Microsoft deadline.

Source

Orion: building the plant before proving the physics

Rather than wait for a definitive net-energy demonstration, Helion is constructing Orion, billed as the world's first commercial fusion power plant, in Malaga, Chelan County, Washington, on land leased near the Rock Island Dam. Site work began in 2025, and by mid-2026 the assembly and office structures were complete, with construction moving to the fusion generator building itself. Constellation Energy will market Orion's output to Microsoft's data centers.

In June 2026 Helion cleared a genuinely novel hurdle: Washington State's Department of Health issued it Radioactive Materials and Radioactive Air Emissions licenses for Orion — the first regulatory licenses in the world for a fusion power plant. Crucially, they were granted under the lighter byproduct-material framework that regulates fusion separately from fission reactors, validating the emerging U.S. regulatory posture that fusion should not be governed like a conventional nuclear plant. That distinction is a quiet but significant advantage for the entire private fusion industry, and Helion is the first to bank it.

Source

Customers first, revenue later: the business model

Helion's commercial strategy inverts the usual deep-tech sequence: it signs marquee offtake agreements years before it can deliver, using them to de-risk the venture in investors' eyes and to force internal urgency. The anchor is the 2028 Microsoft PPA for 50+ MWe; beyond it sit a customer agreement with steelmaker Nucor to develop a 500 MWe plant by 2030 (Nucor also invested), and the reported OpenAI negotiations for 5 GW of power by 2030 scaling to 50 GW by 2035 — roughly an eighth of Helion's planned production. As of mid-2026 the OpenAI deal remained a term sheet, not a signature.

The capital to chase those commitments has arrived in escalating waves: a $425 million Series F in 2025 and, on June 4, 2026, a $465 million Series G led by Thrive Capital at a $15.5 billion post-money valuation, bringing total funding to roughly $1.5 billion. The proceeds are earmarked for scaling manufacturing and accelerating Polaris and Tiny Merge. Helion remains pre-revenue and pre-net-energy, so its valuation is underwritten almost entirely by the credibility of its contracts and the strategic scarcity of firm clean power for AI.

Source

What to watch

The single question that governs Helion's future is whether it can turn a machine that pulses every few minutes into one that delivers grid electricity to Microsoft by 2028 — a timeline most fusion physicists regard as extraordinarily aggressive, and one carrying contractual penalties if it slips. Watch for Tiny Merge coming online, for Polaris demonstrating not just fusion temperatures but actual electricity recapture at rising repetition rates, and for progress on the harder deuterium-helium-3 reaction that its commercial economics depend on.

On the commercial side, the OpenAI negotiation is the swing factor: converting it from a reported term sheet into a binding, sited, gigawatt-scale agreement would reset expectations for the entire fusion sector, while its collapse would remove much of the story underpinning Helion's valuation. Sam Altman's recusal as board chair to manage that conflict of interest is a reminder of how tightly Helion is woven into the AI-power complex — a source of both its momentum and its risk. If Helion hits its dates, it becomes the company that made fusion a product; if it misses, it becomes the clearest test yet of whether commercial deadlines can outrun fusion physics.

Source