OpenAI's public equity pitch is no longer a thought experiment. It is becoming part of the political price of building the most valuable AI companies in America.
OpenAI has moved closer to the public markets just as Washington is getting more serious about taking a public stake in frontier AI. The company said on June 8 that it had confidentially submitted draft IPO paperwork to the SEC, giving itself the option to list while saying the timing remains undecided. Reports have put its valuation around $850 billion to $852 billion, with bankers including Goldman Sachs and Morgan Stanley involved in the process.
That would be a huge listing in any market. It also turns a policy debate into a cap table question. If the U.S. government ends up receiving even a modest stake in OpenAI, either through a voluntary arrangement or a law passed by Congress, the cost will land somewhere familiar: existing investors, future shareholders, or both.
The political pressure is not abstract. Senator Bernie Sanders has proposed giving the public a 50% ownership stake in the largest American AI companies through a one-time tax paid in company stock, not cash. The proposal is aimed at companies such as OpenAI, Anthropic and xAI, and it is built around the argument that AI systems are trained on society's collective knowledge and should return more of their gains to the public.
OpenAI is not endorsing that number. But Sam Altman has been pushing a smaller version of the same idea. As AP reported this week, Altman met privately with Sanders after the senator floated his plan and said he supported public equity in AI companies, though not at the 50% level. President Trump has also spoken favorably about a smaller public stake in major AI companies, framing it as a way for Americans to participate in the upside of what could become trillion-dollar businesses.
That is why the timing matters. OpenAI's IPO filing gives regulators and future investors a formal window into the company's structure, risks and obligations. If a government stake becomes part of the story before the listing, it will have to be explained in terms public-market investors can price. A 1% stake at an $850 billion valuation would be worth about $8.5 billion. A 5% stake would be worth $42.5 billion. Those are not symbolic numbers, even for a company with OpenAI's ambitions.
For founders and growth-round investors, this is the part worth watching closely. Any equity commitment made before an IPO becomes part of the ownership structure retail shareholders eventually buy into. If the shares are newly issued, everyone else is diluted. If they come from existing holders, the burden is more concentrated. Either way, the public wealth idea stops being a slogan once it enters an S-1.
The harder scenario for the industry is Sanders' version. A 50% stock tax would not simply transfer economic value. It would place the public, through a government-controlled fund, inside the ownership structure of the most important AI labs in the country. That changes how investors think about voting rights, future fundraising, governance and exit value. It also raises a question private AI companies have mostly avoided: how much of their upside will remain private if Washington decides the technology is too socially important to leave alone?
Anthropic is part of the same calculation. The Claude maker has also been moving toward public markets, and recent reports have put its valuation in the hundreds of billions, with some secondary-market estimates reaching far higher. A compulsory ownership transfer at that scale would reshape any eventual offering before investors even see the prospectus.
The regulatory relief angle is hard to ignore
The more practical reading of OpenAI's position is that a voluntary public stake could reduce pressure for harsher intervention. A government that owns a small piece of an AI company has a different incentive structure from one that only regulates it from the outside. That does not make antitrust scrutiny disappear, but it gives policymakers a way to claim that the public is sharing in the gains without immediately breaking up the companies producing them.
OpenAI has been preparing that argument for months. In April, the company released a policy paper calling for a public wealth fund, taxes on automated labor and other measures meant to spread the benefits of AI more broadly. The paper framed these ideas as early proposals rather than final answers. Still, it gave politicians a ready-made vocabulary for talking about AI wealth without relying only on corporate promises.
There is a reason this has traction now. AI companies are asking the public to tolerate more data centers, more energy demand, more automation anxiety and more market concentration. The public is asking what it gets in return. A small equity stake is one answer. Sanders' proposal is a much bigger one. The space between those two numbers is where the next phase of AI policy will be fought.
OpenAI is betting that showing up early with a version it can live with makes the harsher version less likely. Its confidential IPO filing gives the company optionality, but it also raises the stakes. When the public version arrives, investors will look for revenue, margins and infrastructure costs. They will also look for signs that Washington is becoming a shareholder, a tax collector, or both.
Also read: EngineAI takes its humanoid robot factory story to Hong Kong's public markets • BofA flagged seven of ten bear signals and Broadcom's miss confirmed the call • Barcelona startup THEKER closes Europe's largest robotics Series A with $85 million
