OpenAI’s long-anticipated market debut was already gathering momentum when the company confidentially filed for a potential IPO in June. But the prospect of public-market scrutiny now appears to be colliding with a more basic question: whether the company can safely keep accelerating its most powerful systems.¹

The warning signs sharpened in the weeks that followed. Jacob Coxon, a researcher who had worked at both OpenAI and Anthropic, left Anthropic and said he was quitting the industry over safety concerns, writing that the people building AI “earnestly believe that it could kill us all by the end of the decade.” His departure fed calls to slow the race for frontier models.¹

Against that backdrop, Altman told Fortune on Friday that OpenAI would not list in 2026. “Right now would be an ill-advised moment to go public,” he said, adding that the company had “a lot of stuff to do” on safety, alignment and cooperation between industry and governments.²

That is a strikingly practical argument against an IPO: staying private, Altman suggested, gives OpenAI latitude to make choices that may not flatter near-term business interests. “We need to be able to make decisions that are not obviously in the interest of our business and our shareholders,” he said, invoking the company’s mission and the responsibility it entails.¹

Altman’s position is not a rejection of progress, but a demand to pace it. He said OpenAI has discussed pausing training runs at new capability thresholds to give researchers time to improve safeguards—and said it would do more of that. He also drew a firmer line: “I do not think we should train models where we cannot make a safety case” for their controllability and alignment.¹

The delay therefore pushes one of Silicon Valley’s biggest IPOs beyond 2026, possibly to 2027 or later. More importantly, it turns OpenAI’s listing timetable into a test of whether the AI industry can put restraint ahead of the next valuation milestone.²