Sam Altman's two biggest bets are starting to merge. Helion Energy, the fusion startup where Altman has personally invested over $375 million, is in talks to sell 12.5% of its power output to OpenAI, the company Altman runs.
The deal, reported by TechCrunch, would give OpenAI a direct claim on fusion-generated electricity - assuming Helion can actually deliver it. Helion has previously signed a power purchase agreement with Microsoft targeting 2028, but no fusion company has yet produced net energy gain at commercial scale. That's a big "if" sitting underneath this entire negotiation.
The conflict-of-interest questions here are obvious. Altman is the CEO of the buyer and a major investor in the seller. OpenAI's board has already navigated messy governance situations, and adding an energy deal with Altman's personal portfolio company doesn't simplify things. Helion would presumably need to demonstrate that the terms are market-rate and that Altman recused himself from both sides of the negotiation.
But the strategic logic is real. AI data centers are power-hungry and getting hungrier. OpenAI, Google, Meta, and Microsoft are all scrambling to lock down energy sources - nuclear, solar, and now fusion. Training a single frontier model can consume as much electricity as a small city uses in a month. Any company that can secure cheap, abundant power has a structural cost advantage over competitors who are bidding up prices on the same grid.
Fusion power, if it works, would be nearly limitless and carbon-free. That's the pitch. The reality is that commercial fusion has been "20 years away" for about 60 years running. Helion says it's different, pointing to its seventh prototype and its Microsoft contract. Investors, including Altman, have put roughly $600 million into the company.
For now, this is a letter of intent, not a power bill. But it signals where the AI industry thinks the real bottleneck is heading: not chips, not data, not talent - watts.