OpenAI Wants to Give the US Government a 5% Stake. The Pitch Compares It to Alaska’s Oil Fund — But AI Isn’t Oil.

By John

OpenAI has proposed handing the U.S. government a 5 percent equity stake in the company — worth roughly $42.6 billion at the $852 billion post-money valuation the company closed in March 2026. Sam Altman’s public rationale isn’t “the government deserves a cut because it’s powerful.” It’s a specific, borrowed model: the Alaska Permanent Fund, the sovereign wealth vehicle that invests Alaska’s oil revenue and pays annual dividends to every state resident. Altman wants leading AI companies to each carve off 5 percent of equity into something similar, so the public shares directly in AI’s financial upside the same way Alaskans share in oil wealth.

It’s a clean pitch. It’s also built on an analogy that falls apart exactly at the point that matters most: where the underlying value actually comes from.

The Alaska Model, and Why Altman Picked It

The Alaska Permanent Fund works because oil is a finite, physically extracted resource that sits under state-owned land — the state has a straightforward legal and moral claim to a share of value pulled out of the ground within its borders, and the fund exists to convert that claim into direct payments to residents rather than letting it evaporate into general government spending. It’s popular, durable (running since 1976), and politically easy to defend: nobody disputes that the oil belongs, in some real sense, to the state before a company extracts and sells it.

Altman’s framing borrows the *popularity* of that model — broad-based public benefit from a valuable resource — without the underlying legal claim. Alaska’s oil is a physical asset the state has sovereign rights over. OpenAI’s $852 billion valuation is built from software, compute infrastructure OpenAI purchased and operates, and models trained largely on data scraped from the public internet and licensed content — none of which the federal government holds anything like Alaska’s ownership claim over. The proposal isn’t “the government owns 5% because it’s entitled to the underlying resource.” It’s “OpenAI is choosing to give the government equity,” a fundamentally different, and much more discretionary, transaction dressed in the language of a resource-sharing model that doesn’t actually map onto how AI value gets created.

What OpenAI Actually Gets Out of It

Altman reportedly began pitching this concept directly to the Trump administration in early 2025 — well over a year of quiet groundwork before the proposal became public via reporting in July 2026. The stated goal, per OpenAI’s own framing, is defusing mounting political pressure in Washington. That’s worth taking at face value: a company operating at OpenAI’s scale, training on internet-scraped data, sitting adjacent to ongoing debates about AI’s labor market and safety implications, and courting enormous continued capital investment, has real incentive to trade a slice of equity for a friendlier regulatory and political relationship with the administration that will shape whatever rules eventually apply to it.

Altman has engaged directly with President Trump, Commerce Secretary Howard Lutnick, and Treasury Secretary Scott Bessent — a level of direct executive-branch access that’s itself a form of value for a company navigating AI policy, separate from whatever the equity stake accomplishes on paper.

The Parts That Still Aren’t Settled

Several things keep this firmly in “proposal,” not “policy.” Talks are described as conceptual, not concluded. Any actual implementation would likely require an act of Congress — meaning this isn’t a deal Altman and the administration can simply sign, but a structure that would need to survive the legislative process, with all the negotiation, opposition, and modification that implies. And Altman’s framing explicitly calls for *other* leading AI companies to make the same 5 percent commitment, not just OpenAI — but it’s unclear whether Anthropic, Google, Microsoft, or other major labs have any interest in following suit. A single company volunteering equity is a very different proposition than an industry-wide norm, and right now this is clearly the former dressed as a pitch for the latter.

Why the Framing Matters More Than the Number

If this deal happens roughly as proposed, the actual mechanics — a sovereign vehicle holding AI equity, paying some form of public dividend — would be a genuinely novel instrument in how governments capture value from technology companies, closer to a strategic industrial stake than a tax. But the Alaska comparison being used to sell it is doing a lot of rhetorical work to make a discretionary, company-initiated equity gift sound like a resource-royalty arrangement the public is owed. Whether that framing survives contact with Congress, and whether other AI labs actually follow OpenAI’s lead, will determine whether this becomes an industry norm or a one-off arrangement that mostly bought OpenAI political goodwill at a moment it needed it.