OpenAI Ties Its Stock Market Debut to AI Safety Claims

OpenAI will not go public until it can make confident safety claims about its models, chief executive Sam Altman said Tuesday, with no date set.

He said a listing in 2026 is unlikely. He also said waiting too long to go public would be bad for the world. The Verge

The comments came during a Q&A with reporters after his DevDay keynote.

Altman gave the keynote at OpenAI's annual DevDay in San Francisco, California, on September 29, 2026, according to CNBC. He was scheduled to speak at 10 a.m. Pacific Time on Tuesday. The company describes DevDay as focused on new models, tools and agents, with product announcements and sessions for developers.

The remarks follow earlier signals. Altman had said an IPO now would be an "ill-advised moment" because of AI safety concerns, as reported by Fortune. Microsoft AI chief executive Mustafa Suleyman later said Altman had told Fortune's Alyson Shontell that OpenAI would probably delay its IPO.

Formal paperwork is already filed. OpenAI confirmed it submitted a confidential draft S-1 to the SEC, the standard form before a U.S. stock listing, and said timing for next steps was undecided, according to the company. A confidential filing lets a company prepare without committing to a date.

DevDay opened with protests, flyers and chants, as reported by The Verge. Protesters chanted "Sam Altman, get off it, put people over profit." The protest did not change the program.

The broader context here is the trade-off Altman described. Safety confidence comes first. Calendar timing comes second. Yet a long delay has its own costs.

In my view, that matters for teams building on OpenAI's platform. A safety-gated IPO gives OpenAI room to hold back models, limit access or change release timing without the disclosure rules of a public company. It also means employees, partners and customers keep working with less financial transparency than a listing would bring. That combination offers steadier roadmaps, with less visibility into governance.

Worth flagging for the longer term, this leaves a path where public ownership could widen access to frontier systems. If OpenAI can state its safety claims clearly enough to satisfy itself before listing, those same statements could help outside developers document risks, controls and limits for their own users. That would be a practical gain to watch for in future safety language.