OpenAI's IPO has a Sam Altman problem

I think OpenAI can sell a blockbuster IPO. The harder sell is Sam Altman as a governance risk the market is supposed to shrug off. Reuters reported that the company is preparing to confidentially file for a U.S. listing in the coming weeks. That matters less to me as a spectacle than as a disclosure test. Public investors do not just need to know whether demand is there. They need to know who holds power, who gets paid and who benefits when the most hyped company in artificial intelligence signs deals around itself.

The valuation headline is almost the easy part. OpenAI was last valued privately at $852 billion, and Fortune argued that a filing could support an IPO worth as much as $1 trillion while forcing answers about control, compensation and cash burn. I do not find the trillion-dollar story absurd. Public markets have backed money-losing technology companies before. What they tend to demand, at least eventually, is a cleaner incentive map than late-stage private capital does when founder mystique is still carrying half the narrative.

That is why this is not a routine founder-led software float. Reuters later reported that Altman held more than $2 billion of stakes in companies that dealt with OpenAI, while the current public-benefit-corporation structure leaves Microsoft with 27 per cent and the OpenAI Foundation with 26 per cent. None of that is proof of wrongdoing. It is proof that the ownership chart, the board’s leverage and the related-party protections are not footnotes. They sit at the centre of the equity story.

Karen Kwok said it cleanly in a Reuters Breakingviews column:

“A charismatic founder can keep a scrappy young company afloat. A domineering one can push a maturing giant astray.”

— Karen Kwok, Reuters Breakingviews

That line gets closer to the real public-market issue than most IPO hype does. Private investors will often live with blurry lines if they think they are buying access to the next platform giant. Listed shareholders are supposed to get something stricter: cleaner disclosure, sturdier guardrails and some believable limit on how far one executive’s outside interests can shape the valuation. Bloomberg’s reporting after the Elon Musk trial points the same way. Winning one legal fight does not settle the questions that would show up in an S-1.

The bull case

I can see the bull case. If OpenAI keeps revenue growth high, holds its lead in consumer AI and stays the default partner for enterprise adoption, plenty of investors may decide governance is something to price rather than fear. Altman told Fortune that “filing for an IPO is different from being ready to go public”. Fair enough. A confidential filing can impose discipline before it raises a dollar. It can force a company to explain itself before it asks the market to pay a premium multiple.

I still come out closer to the sceptics. If OpenAI really needs something like $600 billion of computing power to sustain its ambitions, as Reuters Breakingviews argued, future shareholders would be financing extraordinary scale while also trusting an unusually concentrated power structure. Either bet on its own would be demanding. Put them together and the valuation exercise gets a lot messier than the standard fast-growth, losses-now, profits-later tech float.

What the filing has to answer

So I do not care much whether Altman is brilliant. The questions that matter are narrower and more boring, which is exactly why they matter. Where does his authority stop? How will compensation work? What related-party restrictions apply? What happens to minority shareholders if OpenAI’s priorities diverge from those of Microsoft, the Foundation or Altman-linked interests? If the filing answers those crisply, the market may absorb the complexity. If it hedges, the IPO deserves to clear at a discount to the mythology.

That is my basic view. Public investors should worry less about whether OpenAI can get the deal done and more about what kind of company they are being asked to own. Scarcity can fill an order book. Narrative can support eye-watering multiples. A public listing is not supposed to be a fan vote. If OpenAI wants a premium valuation, the filing has to make Altman look less like the whole operating system and more like a chief executive working inside a structure other shareholders can trust.