OpenAI’s Sam Altman says it would be ‘ill-advised’ to go public in 2026

TechCrunch reports that OpenAI has filed confidentially for an IPO, yet CEO Sam Altman says the company will not go public in 2026. The comments came in a Fortune interview with editor in chief Alyson Shontell, asked amid fallout from the OpenAI–Hugging Face hack and wider AI-safety debate whether IPO plans still push the lab to move really fast.

Altman’s answer was blunt: OpenAI is not rushing. Given everything happening with safety, he said right now would be an ill-advised moment to go public. He framed readiness as both business readiness and readiness for what the moment is like in society with this technology. Pressed on 2026, he said not 2026 — the company has a lot of stuff to do.

We’re paying attention because TechCrunch also reconnects this to June New York Times reporting: bankers and lawyers were hired with a third- or fourth-quarter 2026 goal, then the company leaned toward 2027 on tech-stock volatility and its own financial challenges. Today’s quote turns rumor into CEO confirmation.

We’re treating TechCrunch’s angle as a markets-and-incentives story with a safety overlay — distinct from the control-risk emphasis in other writeups of the same interview. The load-bearing facts here are the confidential filing, the explicit rejection of 2026, and the dual readiness test: the business and the society.

That societal clause matters. Plenty of CEOs delay listings for rate windows or margin optics. Altman is saying the technology’s place in public life is part of the go/no-go. If he means it, the IPO is not only a finance event; it is a legitimacy checkpoint. If he does not, “society” becomes a convenient fog machine whenever markets look unfriendly.

The Hugging Face framing in Shontell’s question is the pressure test. TechCrunch places the IPO talk inside hack fallout and safety discourse, not as a standalone capital-markets aside. A lab that still feels IPO heat will keep shipping through turbulence. A lab that can honestly say it feels no rush can slow eval programs, widen disclosure, and fix agent containment without quarterly theater.

The June NYT thread TechCrunch cites — 2026 target drifting to 2027 on volatility and financial challenges — complicates the pure-safety narrative. Money stress and safety stress can both counsel delay. Our read: believe the “ill-advised” safety line as sincere preference, and still assume capital structure and market mood remain co-authors of the calendar.

Useful AI benefits from patient capital more than from a forced float into a trust deficit. I’m watching whether “when society is ready” gets defined with anything measurable, whether the confidential filing advances while the public timeline slips, and whether financial-challenge reporting stays in the same paragraphs as safety when the next banker leak lands.

Our desk take: confirm the delay, keep the societal bar, and do not let a confidential filing quietly reimpose speed pressure the CEO just disavowed.

Fortune’s Alyson Shontell interviewed Altman during Hugging Face hack fallout and broader safety talk. TechCrunch notes OpenAI’s confidential IPO filing, quotes his ill-advised and not-2026 lines, and cites June New York Times reporting that a 2026 listing goal had already been sliding toward 2027 amid tech-stock volatility and OpenAI’s financial challenges.