Helion quietly rewrote its 2028 promise: Orion will 'begin initial operations', not make electricity

Helion has removed from its website the statement that Orion, its first commercial fusion plant in Malaga, Washington, will begin producing electricity by 2028. The page now says the plant will "begin initial operations in 2028." Jackie Siebens, Helion's public affairs director, told Axios that "we will not, in no universe, have 50 megawatts hit full operation in 2028," putting full power in 2029 or 2030. Helion has also changed its stated goal for Polaris, its current prototype, from producing "net electricity" to "demonstrating electricity from fusion." Its 2023 power purchase agreement with Microsoft requires Orion to be connected to the grid and begin operations by the end of 2028; whether connecting to the grid means delivering electricity is not clear from the public terms. Axios first reported the hedging in early September. The tracker's entry for Helion's plant licences now carries the change.

Why it matters

Helion's 2028 date has been the most cited timeline in fusion, because it was the only one attached to a signed contract with a named buyer. Microsoft agreeing to buy fusion power by 2028 made "fusion is always 30 years away" look out of date, and it anchored a good part of the sector's fundraising story.

The edit matters less for the year than for what the words now promise. "Produce electricity by 2028" is a falsifiable claim about physics. "Begin initial operations in 2028" can be met by a plant that is grid-connected and running but has not delivered net power — which may also satisfy the Microsoft contract as written. The Polaris change is the more important of the two. Net electricity from Polaris was the proof Helion's approach works at all; "demonstrating electricity from fusion" is a lower bar that a machine could clear while consuming more energy than it returns. A company softening the experimental claim and the commercial date in the same season is telling you the experimental result is not in yet.

None of this is a failure. Helion has built and licensed more than any rival, cleared full plant licensing first, and raised accordingly. But the fusion timeline the industry cited most is now a 2029–2030 one with a softer physics claim, and the change was made by editing a webpage rather than announcing it.

What to watch

Whether Helion publishes a Polaris result and which bar it claims, what Orion delivers to the grid in 2028, and whether Microsoft comments on what the contract requires.