A fusion plant delivers grid electricity under a PPA

Where things stand (as of 14 Oct 2026)

Nothing qualifies yet. No fusion device anywhere has produced net electricity, even at the scale of a single machine. Five companies have power purchase agreements (PPAs) or offtake deals (FIA 2026 report). Only one of those plants could plausibly export power before 2032.

Helion Orion → Microsoft (Constellation acts as power marketer)

• Schedule. The plant is under construction in Malaga, WA. Helion's Orion page now promises only "initial operations beginning in 2028." In 2026 Helion took down its earlier claim that Orion would produce electricity by 2028. Its public-affairs director said at FusionX: "We will not, in no universe, have 50 megawatts hit full operation in 2028." She put full power at 2029 or 2030. She also said the Microsoft contract only requires Orion to "be connected to the grid and begin operations by the end of 2028" (Axios Pro, 8 Sep 2026; Axios Seattle, 28 Sep 2026; Fusion Report, 17 Sep 2026). Commissioning "could begin near the end of 2028," ramping in 2029 (NCW News, 27 May 2026).
• Technical status. Helion's FAQ (updated 18 May 2026) downgrades the Polaris prototype's goal from "net electricity" to "demonstrating electricity from fusion." It lists that demonstration as still-remaining work. The prototype was originally meant to show net electricity in 2024.
• Design not settled. Helion is building a "Tiny Merge" testbed to resolve questions about how its plasmas form and merge before the plant design is locked (GeekWire, May 2026).
• Physics doubts. Two 2026 peer-reviewed critiques (Nicolas; Lackner) argue its assumptions may be "wrong by orders of magnitude." Co-founder Slough also disputes its confinement physics (Scientific American, 19 May 2026).

Other plants are too late or don't qualify

• CFS's ARC (Google and Eni deals) targets the "early 2030s." Its grid-connection application was filed in April 2026, and that process may take 4–6 years (POWER; E&E; Utility Dive).
• Type One / TVA starts construction in 2028 at the earliest. TAE targets net energy gain in a pilot project only by 2031 (WNA, Oct 2026).
• China's BEST plans to finish construction by end-2027, with a power-generation demonstration "around 2030" (Sina/CCTV, 2 Oct 2026). It has no PPA and would not be net-positive. The Xinghuo fusion-fission hybrid would make mostly fission electricity, which the question excludes.

Reference class

Fusion net-electricity timelines have been missed without exception for decades. Helion has already slipped and redefined its 2024 milestone. Most developers (71% in the FIA 2026 survey) expect first commercial electricity sometime in the 2030s.

The bar here is high. It needs net-positive export at the meter after the plant's own consumption, under a PPA signed beforehand. It also needs confirmation from both the operator and the buyer.

Estimate

Helion pathway (~3.8%)

• Chance Polaris shows gross "electricity from fusion" by 2027: about 50%.
• Chance Orion is net-positive by end-2031: about 8% if that demonstration happens, about 1% if not.
• Chance Microsoft and Helion then confirm a delivery under the PPA, given net output: about 85%.

Other pathways

• About 1% for an ambiguous Helion/Microsoft "first delivery" claim that a resolver accepts.
• About 1% for everyone else combined (CFS, China, others).

Total: about 5% by end-2031.

Timing. I modelled quarterly hazards and accumulated them in Python:

• About 0 through Q3 2028, since no qualifying plant can run before then.
• 0.1% in Q4 2028, the contractual start-up.
• Rising from 0.15%/quarter to 0.3%/quarter through 2029, 0.35–0.5% in 2030, and 0.55–0.65% in 2031.
• Cumulative: 0.3% by end-2028, 1.2% by end-2029, 2.9% by end-2030, 5.2% by end-2031.

Strongest case against

The strongest case for a higher number:

• Helion has more than $1.5B raised and a $15.5B valuation.
• It holds state licences and a grid interconnection, construction is reportedly ahead of schedule, and it faces contractual penalties for missing dates.
• It reached 150M°C with deuterium-tritium fuel.
• Its direct energy recovery lowers the gain it needs, and investors may have seen private data.

Thin Manifold markets put Helion delivering 10 MW to Microsoft before 2031 at about 15%. But they look stale: the "50 MW before 2029" market still sits at about 11% after Helion's "no universe" statement. I discount them, but they keep me from going below about 3%.

Question. By the last day of each quarter, has a fusion power plant delivered electricity to a grid under a signed power purchase agreement?

Definitions.

• Fusion power plant: a facility that converts energy released by nuclear fusion into electricity, whether through direct conversion or a thermal cycle, anywhere in the world. A demonstration or pilot plant can qualify. Electricity generated by fission, auxiliary generators, or previously imported power does not qualify merely because it is exported from a fusion site.
• Grid: an operating electricity transmission or distribution network serving customers outside the generating facility. An isolated plant test circuit or internal electrical bus is not a grid; a customer-serving island grid or microgrid can qualify.
• Delivered electricity: actual, positive net electrical energy exported at the plant's grid connection over a documented delivery or settlement interval, attributable to fusion generation after accounting for plant electricity consumption during that interval. No minimum capacity, duration, annual net output, or profitability is required.
• Signed power purchase agreement (PPA): an executed, binding agreement between a seller and an identified buyer to purchase electricity generated by that plant. Physical and financially settled or virtual PPAs count if actual grid-delivered output is accepted or settled under the agreement. A letter of intent, memorandum of understanding, investment agreement, grant, or standalone purchase of environmental certificates does not count.
• Under the agreement: the qualifying export is recognized as a delivery or settlement of generated electricity under the PPA, which was executed before that export. Merely having a PPA for future output is insufficient.

Resolves YES if qualifying public sources identify the plant, confirm that fusion-generated electricity has actually been exported to a grid, and confirm that at least one such delivery was accepted or settled under an executed PPA. A joint statement can establish all these facts; neither publication of the full contract nor a disclosed meter reading is required if the statements explicitly establish qualifying delivery. Commissioning or test electricity counts only if it meets the same requirements.

Resolves NO if no qualifying confirmation has been published by the quarter-end. Signing a PPA, building or connecting a plant, achieving fusion ignition or plasma energy gain, producing electricity internally, announcing a future delivery date, or exporting non-fusion electricity does not suffice. Contracted supply fulfilled with electricity purchased from other generators does not count.

Resolving source. An official statement or public regulatory filing by the plant operator or delivering utility, corroborated by an official statement or filing from the PPA buyer confirming qualifying delivery under the agreement. A joint announcement by those parties qualifies. Named quotations in such an official announcement count; unattributed press reports and executive interviews do not. This is a narrower resolving-source requirement than Public confirmation (A2). Apply the Resolution window and Resolution date (A3) conventions; where multiple publications are needed, use the publication date of the last required confirmation. Earlier documents may establish contract terms but cannot alone establish a qualifying resolution. The operator, utility, and buyer may be any organizations, not only covered developers.