News · 2026-07-04

OpenAI Is Reportedly in Early Talks to Give the US Government a 5% Stake

OpenAI is reportedly in early discussions about giving the US government a roughly 5% stake in the company. The talks, first reported by the Financial Times and covered by CNBC on July 2, 2026, are described as preliminary - there is no agreed deal, and any arrangement would need new legislation from Congress.

Key facts

• A 5% stake would be worth about $42.6 billion at OpenAI's roughly $852 billion valuation from its March 2026 funding round.
• The proposal floated is a sovereign-wealth-fund-style "public wealth fund" holding about 5% of each leading US AI lab - OpenAI, Anthropic, Google, and Meta.
• Reported July 2, 2026 by the Financial Times, corroborated by Reuters and the Guardian; it would require congressional legislation.
• It surfaces ahead of OpenAI's planned September 2026 IPO.

The report lands at a strange moment for one of the most valuable private companies in the world. OpenAI is under two kinds of pressure at once: political scrutiny over how much power a handful of private AI labs now hold, and its own need to keep raising enormous sums of capital as it races toward a public listing. Giving the government equity is one way to defuse the first problem while OpenAI works through the second.

To understand why this is even on the table, it helps to know what a stake actually means. When you own equity in a company, you own a slice of everything it is worth and a claim on a slice of its future profits. Owning 5% of OpenAI would not mean the government runs the company or makes its product decisions; it would mean the US Treasury holds a financial interest that rises and falls with OpenAI's fortunes. What is reported here is not a check the government writes to OpenAI, but a stake it would simply be granted or allowed to hold - the kind of arrangement more familiar from sovereign wealth funds, the state-owned investment pools that countries like Norway and Singapore use to invest national savings, than from a typical Silicon Valley cap table.

According to the Financial Times' sources, Sam Altman has framed the idea as the "best way to share the upside of AI with the public." The logic is that if AI turns out to be as transformative and profitable as its builders claim, ordinary taxpayers should get some of that windfall directly, rather than watching all the gains flow to a small group of private investors and employees. Think of it less like a government bailout and more like a national pension fund taking a slice of a fast-growing company's future in exchange for that company getting political cover and a friendlier relationship with the regulators who will increasingly shape its business.

The mechanics matter here, because the version reported is bigger than a single deal. A 5% stake distributed across OpenAI, Anthropic, Google, and Meta would turn the federal government into a shareholder of the entire frontier AI industry at once, not just one company. That is a far more structural move than a one-off investment. It would tie US fiscal interests directly to AI company valuations, which raises obvious questions about conflicts of interest: a government that owns a piece of the companies it is supposed to regulate has a financial reason to want those companies to succeed. It is also why the proposal would need an act of Congress rather than a private negotiation. This is not something OpenAI's board could simply approve on its own; it would require lawmakers to write and pass a new framework for how the government takes and holds such stakes.

Why it matters: if this moves from talk to reality, it would be one of the largest government equity positions in private tech history, and it would set a template that other countries and other labs would likely study closely. As AI companies keep bumping up against national security concerns, antitrust scrutiny, and questions about who ultimately controls the most powerful models, the boundary between private enterprise and public oversight is being redrawn in real time (see GPT-5.6 launches under government vetting for a related thread on the US government's growing hand in frontier AI oversight). A public ownership stake would be a dramatic marker of just how far that boundary has already shifted.

The honest caveat: this is reported early-stage talk, not a settled deal. It requires legislation that does not yet exist, and the political will to pass it is far from guaranteed. Other labs are already distancing themselves - a source at Anthropic told reporters the company has "not discussed government stakes," while Google, Meta, and the White House all declined to comment. That pattern of denials and no-comments is itself worth reading carefully: it suggests the idea, if real, is far from consensus even among the companies it would affect. Treat this as a trial balloon worth watching, not a done deal.

Primary source, verified: read the paper →

Ground Truth. (2026, July 4). OpenAI Is Reportedly in Early Talks to Give the US Government a 5% Stake. Ground Truth. https://groundtruth.day/news/openai-floats-a-5-percent-us-government-stake.html

@misc{groundtruth:openai-floats-a-5-percent-us-government-stake,
  title  = {OpenAI Is Reportedly in Early Talks to Give the US Government a 5% Stake},
  author = {{Ground Truth}},
  year   = {2026},
  month  = {jul},
  url    = {https://groundtruth.day/news/openai-floats-a-5-percent-us-government-stake.html}
}

Topics: OpenAI · policy · government · Sam Altman · business