Fusion plant before the physics

On July 30, 2025, Helion Energy began site work on a fusion power plant called Orion, on the Columbia River in Washington. The order of events is the story. A signed contract came first: a 2023 agreement to sell Microsoft electricity from 2028, the first commercial supply deal for fusion power. Now there is a construction site. What does not yet exist is a machine that has made more energy than it consumes.

"We focus on the electricity piece, making electricity, rather than the pure scientific milestones." — Helion CEO David Kirtley, asked about scientific breakeven

Helion's prototype, Polaris, reached 150 million degrees by early 2026 — its own engineers call that about three-quarters of the way to commercial conditions, and it has never demonstrated net electricity. Orion is a larger, separate machine that has to clear a bar the prototype hasn't. Asked about scientific breakeven, the milestone that would prove the approach, Helion's CEO redirects to making electricity instead. The 2028 date is a contractual promise, not a forecast; outside physicists call it aggressive — no fusion device anywhere has yet produced net electricity at all.

The sharpest doubt comes from inside. John Slough, the researcher whose work founded the company and who has since split from it, says he can no longer see anything in the physics that supports Helion's design — that firing plasmas together to compress them drives the kind of instability that bleeds away the magnetic field before fusion can do its work. A first commercial fusion contract is a genuine first. Whether a building can be willed ahead of the science it depends on is the bet that 2028 will settle, one way or the other.