Sam Altman Says an OpenAI IPO Would Be ‘Ill-Advised.’ Safety Isn’t the Only Reason

“I actually think that given everything happening with safety, right now would be an ill-advised moment to go public.” Sam Altman delivered that verdict from OpenAI’s San Francisco headquarters, and with a single sentence he pushed the most anticipated technology listing of the era over the horizon. The line was calm, deliberate and eminently quotable. It was also only half the story.

The reading Altman wants you to take is the honest-sounding one: a company building systems this powerful should not answer to public shareholders while it is still learning how to keep them safe. It is a defensible position, and he argued it plainly. But chief executives rarely make a decision this large for a single reason, and the reason they say out loud tends to be the one that flatters them. The more revealing question is what staying private also buys him.

Start with what is on the record. OpenAI filed confidentially for a public offering earlier this year, amid talk of a valuation near $1 trillion, before quietly pausing the effort. Pressed on timing in an interview with Fortune, Altman was unambiguous: “I would say not 2026, yeah. We’ve got a lot of stuff to do.” The company is now understood to be looking toward 2027, and Altman insists the clock is his to set. “We’re not rushing into an IPO,” he said. “When we’re ready, which is when the business is ready, when we feel ready from what the moment is like in society with this technology.”

That last clause is where caution shades into strategy. A public listing would hand a claim on OpenAI to shareholders whose mandate is returns — and Altman all but conceded the friction. He has pointed to the company’s unusual nonprofit-over-for-profit structure, which he says obliges OpenAI to make choices “not obviously in the interest of our business.” Rendered in market terms, that is a chief executive saying he intends to keep making decisions public investors would punish. “We don’t feel pressure on that,” he added. Staying private is how you protect that freedom — and how you keep your books closed while the business burns cash and settles on a model.

The contrast next door sharpens the point. Anthropic, OpenAI’s closest rival, is reportedly preparing to go public before the year is out, at a valuation some believe could reach $2.3 trillion. Two labs building the same kind of technology have read the same moment and walked through opposite doors — one opening its accounts to the market, the other bolting them shut and calling it prudence. They cannot both be simply right about what safety demands of a public company.

None of which makes the safety concern a fiction. Altman spoke against a genuinely uneasy backdrop: fallout from a security incident involving OpenAI’s systems, an Anthropic researcher putting the odds of an extinction-level outcome above one in ten within a decade, and Altman himself hinting that the leading labs may be close to a pact to slow development. A firm that might soon agree to tap the brakes is an awkward fit for a market that pays for acceleration. The safety story and the strategic story happen to point the same way.

So the delay is real, and so is the thinking behind it — just not all of the thinking. For now the most closely watched startup in the world will keep answering to a board and a mission rather than a ticker, and outsiders will keep pricing it by rumor. When OpenAI finally opens its books, the valuation will matter less than the question Altman is postponing: whether a company that sometimes needs to act against its own commercial interest can survive contact with people who bought the stock to make money.

Tags: OpenAI, Anthropic, Sam Altman, AI safety, IPOs