Altman rules out an OpenAI IPO before 2027

Sam Altman says OpenAI won't go public until at least 2027. In an interview published September 12, he called this an "ill-advised moment to go public," pointing to safety and alignment work the company still has ahead of it, and said he's "happy to be able to do that as a private company," per Fortune. Bloomberg, Axios and CNBC all ran the story the same day.

That's a specific thing to watch: a safety concern moving a real financial decision, not just a blog post or a conference panel. OpenAI has talked about safety timelines before. An IPO is money on the table, and Altman is saying no to it over this.

It also fits a pattern. Altman has separately told OpenAI staff the company would consider slowing frontier development if other labs did the same, a stance that only works if competitors actually follow. Staying private for another year or more removes the quarterly disclosure and shareholder pressure that usually forces a company to explain safety delays in public. It also removes the audience that would ask why.

An IPO means a prospectus, audited financials, and a board that answers to shareholders on a fixed schedule. None of that exists while OpenAI stays private, so Altman's own account of what's slowing things down has no outside check attached to it. If 2027 arrives and OpenAI is still private, that's worth asking about directly.