Generated by Codex with GPT-5

Fusion power is usually described as a promise that keeps moving away. In “The Fusion Wager,” Alex Pasternack looks at a company trying to force that promise onto a commercial clock. Helion Energy, based in Everett, Washington, is building a machine called Orion near Malaga, Washington, and says it will deliver 50 megawatts of electricity to Microsoft data centers by 2029. That is not just an engineering milestone. It is a bet that fusion can become a manufactured energy product before the field has settled some of its hardest scientific questions.

The article’s central tension is simple: Helion has money, customers, urgency and an unusually direct technical path, but fusion still answers to physics. Private capital and the power demands of artificial intelligence have given fusion start-ups a new market story. Data centers need reliable, carbon-free electricity at huge scale, and companies such as Microsoft are willing to sign deals before any private fusion plant has proved it can put net power on the grid. Helion is therefore a test case for a broader question: can start-up speed change a field that has long advanced through national laboratories, giant machines and slow public funding cycles?

A Different Fusion Machine

Most familiar fusion designs try to confine plasma inside a tokamak, a doughnut-shaped magnetic chamber, or a stellarator, a more geometrically complex magnetic bottle. Helion is pursuing a less common design based on a field-reversed configuration. In plain terms, that means it forms compact rings of plasma at opposite ends of a linear machine, accelerates them toward each other, and makes them collide. The collision is supposed to heat and compress the plasma in extremely short pulses, creating fusion conditions for fractions of a millisecond.

That pulsed design matters because Helion also wants to avoid the usual route for turning fusion into electricity. Many fusion concepts would use heat to boil water, spin a turbine and run a generator, much like a conventional power plant. Helion instead aims to capture electricity directly from the expanding plasma as it pushes back against magnetic fields. If that direct recovery works at very high efficiency, the company would not need as large a fusion energy gain as some rival approaches.

The promise is elegant, but it comes with hard constraints. Fusion requires plasma hotter than the core of the sun, stable enough to react, and dense enough for those reactions to matter. Common fusion fuel based on deuterium and tritium brings neutron damage and tritium supply problems. Helion ultimately wants to use deuterium and helium-3, which would produce fewer damaging neutrons and fit better with direct energy capture, but helium-3 is scarce and the reaction is harder to ignite. The company says its pulse timing and machine design can make the physics work. Critics are not convinced.

The Crucial “If”

Pasternack presents Helion as both impressive and opaque. Its seventh-generation prototype, Polaris, is a large pulsed machine backed by massive capacitor banks and hardware that must fire with nanosecond precision. Helion says Polaris has reached 150 million degrees Celsius and has demonstrated fusion with deuterium-tritium fuel. It has also signed a power purchase agreement with Microsoft and announced a much larger development deal with Nucor.

But those achievements do not yet equal a power plant. The article emphasizes that Helion has not published enough peer-reviewed performance data for outsiders to fully judge its claims. That absence matters because the design depends on several demanding assumptions at once: that field-reversed plasmas can remain stable through violent compression, that the pulses happen quickly enough to avoid destructive instabilities, that useful energy can be recovered directly, and that a future helium-3 fuel cycle can be made practical.

The sharpest criticism comes from John Slough, a co-founder whose earlier research helped create the concept Helion commercialized. He argues that the very plasma configuration Helion relies on is likely to lose stability under the extreme conditions the company needs. Other physicists have questioned whether Helion’s assumptions about ion and electron temperatures make the deuterium-helium-3 plan too optimistic. Helion responds that those critiques miss the importance of its short pulse cadence. The dispute is not over whether fusion is real; it is over whether this particular shortcut can deliver electricity on a start-up deadline.

What The Wager Reveals

The article’s most useful point is that Helion may be changing fusion even if it misses its timetable. By treating fusion as a manufacturing problem, the company is pushing the field to think about components, supply chains, factory production and customers earlier than it otherwise might. Helion makes specialized capacitors and ceramics for its own machines. Other fusion companies face similarly industrial problems with magnets, materials, fuel handling and grid integration. Public funding still matters here; the Department of Energy’s recent \$135 million fusion investment is framed as support for shared barriers that no start-up can solve alone.

That makes the “wager” bigger than Helion’s contract with Microsoft. The wager is that fusion can leave the realm of heroic laboratory experiments and become an energy industry fast enough to matter for the grid in the 2030s. Pasternack does not present that outcome as either fantasy or inevitability. The physics remains unresolved, the public evidence is incomplete, and the commercial dates are aggressive. But Helion has made fusion’s next phase more concrete: not just proving that plasma can fuse, but proving that a machine can be built, repeated, maintained, financed and trusted to deliver power.