Washington is no longer just asking how to regulate AI. It is now asking whether the public should own a piece of the companies building it.
President Trump has opened the door to one of the more unusual ideas now circling the AI industry: a government stake in companies such as OpenAI. There is no signed deal, and the proposal is still early. But the fact that it is being discussed at the White House changes the conversation around AI power, public benefit and who gets paid when the technology becomes central to the economy.
According to Reuters, Trump told reporters on June 5 that his team would look into the idea of AI companies giving the American public a stake in their firms. He also said he planned to meet with AI executives as soon as next week. CNBC has separately reported that OpenAI and the administration have been in ongoing discussions about a possible government equity stake in the company. That does not mean OpenAI is about to hand Washington a board seat. It does mean the most valuable private AI company in America is now part of a policy debate that used to belong mostly to think tanks and campaign speeches.
The timing matters. OpenAI has been valued at about $852 billion in its latest funding round, according to recent reports, and the company has spent the past year arguing that the wealth created by advanced AI should reach people who do not already own technology stocks. Sam Altman has pushed versions of a public wealth fund, a structure that could hold equity or other long-term assets tied to AI growth and distribute benefits more broadly. That idea sounds clean in a policy paper. It becomes much more complicated when the buyer, partner or recipient is the federal government.
The Trump administration has already shown a willingness to use ownership as an industrial policy tool. The White House points to its roughly 10% stake in Intel as proof that public capital can be tied directly to strategic industries. AI is a bigger and more politically sensitive target because the companies involved are not only building consumer software. They are building the models that could shape defense planning, cyber operations, scientific research and the way federal agencies make decisions.
That brings the equity discussion together with another move from the same week. On June 5, Trump signed a national security memorandum directing the Pentagon, intelligence agencies and other national security bodies to accelerate the adoption of advanced AI. The memo calls for faster onboarding of commercial and open-source models from multiple vendors within 120 days, while also requiring updates to policies on autonomous weapons and annual reviews of AI guidance. This is not just a procurement memo. It is a demand signal to every model lab, defense software company and cloud provider that the federal government wants access to frontier systems faster than the usual contracting cycle allows.
The guardrails are part of the message too. The White House fact sheet says national security agencies should use AI in ways that preserve accountability, respect the chain of command and avoid unlawful surveillance of Americans. AP also reported that the directive restricts uses such as censorship, ideological bias and autonomous weapons without human oversight. Those limits are important, but they are not the same as independent oversight. They describe what agencies should avoid. They do not yet answer how the public will know whether those lines are being honored inside classified systems.
The public benefit question
A government stake in OpenAI would be sold as a way to let Americans share in the upside of an industry that may create enormous wealth. That argument has political appeal because many voters can already see the costs of AI more clearly than the benefits. Data centers are raising questions about power and water use. White-collar workers are watching software take on tasks that once looked safe. Schools, courts and media companies are still trying to work out what it means when machines can produce convincing text, images and code at scale.
There is also a harder question for investors and founders. If the government owns part of a leading AI company, does it become a neutral regulator, a strategic customer, a financial beneficiary or all three at once? OpenAI already sells tools into public and private markets. If Washington is both a potential shareholder and a major buyer, competitors will ask whether procurement decisions can remain fair. That question will only get sharper if Anthropic, Google, Microsoft, xAI or SpaceX are drawn into similar conversations.
The politics are strange, which is why the story is worth watching. Bernie Sanders has been talking about public ownership in AI from the left, including a much more aggressive proposal for a 50% stake in major AI companies to create a sovereign wealth fund. Trump is approaching the issue from a different direction, closer to dealmaking and national competitiveness. Altman has his own reason to engage, because OpenAI needs public legitimacy as much as it needs capital and compute. Different roads are leading to the same place: AI companies are becoming too economically important for Washington to treat them like ordinary vendors.
For the market, the immediate implication is not that OpenAI will be partly nationalized next week. The real signal is that the federal government is moving deeper into the AI stack at the same time it is accelerating military adoption. Procurement, safety testing, public wealth proposals and equity stakes are no longer separate debates. They are becoming one negotiation over control, value and trust. The next thing to watch is whether Trump’s meeting with AI executives produces a concrete structure, or whether this remains a trial balloon meant to pressure companies into offering the public something before Congress writes the rules for them.
Also read: Shelbyville’s data center fight shows AI has a local trust problem • Sriram Krishnan is taking Trump's AI fight outside the White House • China tells fund managers to fund innovation without chasing hype
