Sam Altman has begun pitching something that would have seemed far-fetched twelve months ago: the US government should own a piece of every major American AI company.

Bloomberg and the Financial Times reported today that OpenAI has begun preliminary discussions with Washington about a proposal where the US government would receive a 5% equity stake in OpenAI, and potentially in Anthropic, Google, and Meta as well. The model Altman is floating is based on the Alaska Permanent Fund, the state vehicle established in 1976 to invest Alaska’s oil revenues and pay annual dividends to residents.

The discussions are at an early, conceptual stage. Implementing any formal arrangement would likely require an act of Congress. But the fact that this conversation is happening at all, and at this level, is worth paying attention to.

What Is Actually Being Proposed

At OpenAI’s current valuation of $852 billion, a 5% government stake would be worth roughly $42.6 billion. Altman has reportedly raised the idea with President Trump, Commerce Secretary Howard Lutnick, Treasury Secretary Scott Bessent, and Senator Bernie Sanders.

The Alaska Permanent Fund comparison is instructive. The fund collects a share of the state’s oil revenues into an investment vehicle, which then pays annual dividends to Alaska residents. Altman is arguing that AI is the 21st century equivalent of oil, and that the public should share in the wealth it generates rather than having that wealth concentrate entirely in the hands of a small number of companies and investors.

The proposal, as reported, would not apply only to OpenAI. Altman is pitching it as an industry-wide arrangement, where every leading US AI developer contributes 5% equity to the same vehicle. Google, Meta, and Anthropic have not indicated they would agree to participate.

Why This Is Happening Now

The timing matters. This proposal was floated just days after the US government delayed the global export of GPT-5.6, OpenAI’s most capable new model. The export restriction was part of a broader discussion about which countries should have access to frontier AI models, and it signalled that Washington is paying much closer attention to AI company decisions than it was even a year ago.

OpenAI’s proposal looks, in part, like a political play. If the US government holds a financial stake in AI companies, those companies become harder to regulate harshly or break up. A government co-investor has different incentives than a government regulator. Altman is betting that turning Washington into a financial stakeholder makes the relationship less adversarial and buys the AI industry more operating room.

Whether that calculation is correct is a separate question.

The Governance Problem This Creates

There is a real tension at the centre of this proposal. A government that owns 5% of OpenAI has a financial interest in OpenAI succeeding. That interest does not automatically align with the public interest in AI safety, competition, or data privacy.

Antitrust enforcement becomes awkward when the enforcer is also a shareholder. Safety regulations that slow AI development become financially costly to the regulator. The cleanest version of AI oversight comes from a government that has no financial skin in the game. A sovereign wealth stake complicates that.

Other jurisdictions are also likely to respond. If the US government holds equity in OpenAI, regulators in the EU, UK, Australia, and elsewhere will face pressure to either demand analogous arrangements or treat US AI providers differently. For enterprise buyers in those markets, the data sovereignty picture gets murkier, not clearer.

None of this means the proposal is bad. It might be genuinely better than the alternative, where AI wealth concentrates privately while the government tries to regulate a technology it has no direct stake in understanding. But the tradeoffs are real and worth understanding.

What This Means for Business

For businesses currently building on top of AI infrastructure, the near-term reading is actually stable. If Altman’s proposal gains traction, it significantly reduces the chance of sudden, sharp US regulation of AI companies in the next two to three years. A government that is financially aligned with AI company success is unlikely to do anything that materially damages that success.

That is not a guarantee of stability. But it is a shift in the regulatory calculus.

For enterprise buyers specifically, a few things are worth watching:

US providers become politically safer in the near term. Companies like OpenAI and Anthropic, if they are linked to a US government wealth vehicle, become more entrenched as infrastructure. The risk of an aggressive regulatory response that disrupts service contracts or forces data changes is reduced.

Overseas complications grow. If you are procuring US AI services and your business operates significantly in Europe or Asia, the data sovereignty question becomes more complex, not less. A government stake in AI companies is the kind of development that gives overseas regulators reason to require different treatment or domestic alternatives.

The IPO path for OpenAI changes shape. A government co-investor in OpenAI’s equity structure would be an unusual situation for public markets. How that resolves will depend on the structure of any deal, but it adds a dimension to the governance story that enterprise buyers should factor into long-term platform decisions.

The biggest takeaway is not which direction this proposal goes. It is that the relationship between US AI companies and the US government is shifting from a hands-off permissive model toward something more structured. Businesses that assume the current permissive environment is permanent are likely underestimating how quickly that could change, in either direction.

The window for building AI capability inside your organisation has not closed. But the policy environment around AI is no longer background noise. For business leaders evaluating AI strategy, understanding the governance landscape is becoming as important as understanding the technology itself.

---

If you’re deciding where to start with agents, start here. The free Working With Claude field guide walks through the ecosystem, Claude Code, and a real rollout plan. Get your copy.