OpenAI Government Stake Hits 64% on Bipartisan Push for US Ownership

Trump and Sanders Agree on Something: The US Should Own a Piece of OpenAI

For the first time in this political cycle, the American left and right are advancing concrete ownership mechanisms targeting the same company in the same week. On June 6, the Trump administration confirmed active equity stake discussions with OpenAI, with the president calling a US government position in AI firms "a beautiful thing". Four days earlier, Senator Bernie Sanders proposed legislation that would impose a 50% tax on the stock of the largest AI firms, channeling the proceeds into a government-run sovereign wealth fund with direct ownership stakes.

These are not the same proposal. Trump envisions a negotiated equity deal, likely in exchange for favorable regulatory treatment or infrastructure access. Sanders wants a compulsory mechanism that applies across the sector. But the structural output is identical: US government ownership of OpenAI equity. That convergence is the catalyst. When both parties have independent reasons to pursue the same outcome, the probability of some version of it happening rises sharply, even if neither party's specific bill passes in its original form.

Cross-aisle alignment on economic policy this concrete is rare enough to be newsworthy on its own. Cross-aisle alignment on government ownership of a private technology company has no modern precedent. Prediction markets noticed immediately.

OpenAI Government Stake Odds Jump 14 Percentage Points: What the Market Is Pricing In

The implied probability of the US government taking a stake in OpenAI before the end of 2026 surged from 50% to 64% in just three days, a 14-percentage-point swing that represents the contract's sharpest move since inception. The 64% level is notable because it crosses the majority threshold: for the first time, prediction market participants believe a government stake is more likely to happen than not.

The timing of the move maps precisely to the dual announcements. Before Sanders' June 2 proposal, the contract sat at 50%, essentially a coin flip. Sanders' legislative language pushed it upward, but the decisive jump came after the Trump administration's June 6 confirmation. Two independent political actors, each with their own voter base, pursuing the same structural goal within the same week: that is the kind of redundancy that markets treat as reducing the probability of failure.

A 64% implied probability translates to roughly 2-to-1 odds in favor. To put that in context, the market is pricing a US government stake in OpenAI as more probable than not, but still acknowledges a one-in-three chance it doesn't happen. That's not a certainty trade. It's a conviction trade.

What a US Stake in OpenAI Could Actually Look Like, and Why the Structure Is Complicated

OpenAI is not a normal company. Its ongoing conversion from a nonprofit structure to a capped-profit entity creates unusual complexity for any government entry point. Microsoft already holds a substantial stake, providing a precedent for large institutional investors navigating OpenAI's governance, but a sovereign equity position would be structurally different from a corporate one. Government ownership brings regulatory entanglement, national security considerations, and the question of what "ownership" means when the entity's profit distribution is capped.

Three plausible deal structures have emerged in policy discussions. The first is a direct equity stake, the model the Trump administration appears to favor, where the government negotiates a minority position in exchange for regulatory clarity or infrastructure access. The second is Sanders' sovereign wealth fund approach, which would acquire stakes through tax-funded purchases rather than negotiation. The third is a licensing or revenue-sharing arrangement that functions like equity without formal ownership, sidestepping some governance complications.

OpenAI's valuation adds another variable. With Anthropic reaching $1 trillion in secondary market valuations and OpenAI itself valued at $880 billion, even a small percentage stake would represent tens of billions in taxpayer exposure. The Trump proposal's emphasis on returns flowing to American households as dividends suggests the administration is framing this as an investment, not an expense. That framing matters for political viability, but it also raises the stakes if OpenAI's competitive position continues to erode. ChatGPT's market share dropped from 69.1% to 45.3% between January 2025 and early 2026, with xAI's Grok surging to 15.2% and Google's Gemini reaching 9%. Buying into a company whose dominance is declining is a harder political sell than buying into an uncontested monopoly.

The Case Against 64%: Why the OpenAI Government Stake Could Still Fall Apart

The strongest bear case is simple: bipartisan enthusiasm does not equal bipartisan legislation. Trump and Sanders agree on the destination but disagree on every detail of the route. Trump wants a negotiated, voluntary deal. Sanders wants mandatory taxation. These are not complementary proposals; they are competing visions that could easily cancel each other out in Congress. If Sanders' bill gains traction, OpenAI has every incentive to negotiate with the White House quickly to preempt compulsory terms. But if neither proposal can pass on its own, the bipartisan "convergence" may amount to parallel press conferences that produce no legal mechanism by December 31, 2026.

The timeline is also tighter than it appears. The market resolves at the end of 2026, leaving fewer than seven months. No federal equity stake in a private technology company has been completed in that timeframe before. The closest precedent, the 2008 TARP equity injections, occurred under emergency legislation during a financial crisis, conditions that do not currently apply. Absent a crisis catalyst or executive action shortcut, the legislative and regulatory timeline for a novel government investment vehicle may simply not fit inside seven months.

Then there's OpenAI itself. The company has not publicly endorsed any government stake proposal. Its corporate restructuring is already complex, and adding a sovereign investor to the cap table during a nonprofit-to-for-profit conversion introduces legal and governance risk that the board may not accept voluntarily. If the deal requires OpenAI's consent, and OpenAI has reasons to stall, the clock becomes the bear's best friend.

At 64%, the market is pricing bipartisan political will as sufficient to overcome structural, legal, and timeline obstacles. That may prove correct. But the gap between political rhetoric and executed policy has swallowed bigger proposals than this.