The Pulse
Altman Makes OpenAI’s IPO Wait on Safety Claims
Sam Altman said OpenAI needs stronger safety assurances before pursuing an IPO, but warned that waiting too long to go public could also be harmful. At DevDay on September 29, he gave no timeline and said safety and alignment must advance a
AI.info Team · Written with AI ·
Sam Altman put two competing demands on OpenAI’s future IPO at DevDay on Tuesday, September 29: the company needs to make stronger safety claims as its models grow more capable, but waiting too long to list could also be harmful. He offered no date for a public offering. Instead, he described safety work as a condition for deciding when OpenAI is ready to face public markets.
The tension is practical as well as rhetorical. Altman said he does not want the pressure of public investors bearing on decisions while the company adjusts to increasingly capable models and new safety requirements. Yet he also said a company as consequential as OpenAI should not remain private indefinitely.
Safety claims come before a listing
Speaking with reporters after his keynote, Altman tied the question of going public to OpenAI’s ability to make credible assurances about its systems. He said the company intends to keep advancing AI, while recognizing that more capability is coming and must be matched by stronger evidence about safety.
“We intend to continue with AI progress … but as the models have had this surge forward in capability, and we see more of that ahead of us, we have got to be able to make confident safety claims.”
Sam Altman, chief executive of OpenAI, speaking after DevDay, as quoted by The Verge
Altman characterized OpenAI’s approach as putting safety and alignment ahead of capability gains, rather than calling for a blanket halt to development. He said a transition to very capable models brings a different set of safety requirements, and that he did not want investor expectations to add pressure while the company works through that shift.
Wall Street could change the incentives
Altman raised the possibility that a public company could face pressure to satisfy shareholders even when safety work calls for a slower course. He described a risk of disappointing Wall Street investors “in the names of safety or whatever else.” The concern, as he framed it, is not that an IPO itself makes a model less safe, but that public-market demands could complicate choices about pace and deployment.
His comments build on a statement earlier in September. In an interview with Fortune published September 12, Altman said OpenAI would not go public in 2026, calling the timing “ill-advised” given safety concerns. He said the company had substantial work ahead on safety, alignment and cooperation between the AI industry and governments. At DevDay, he did not replace that decision with a new target year.
Altman also argues against waiting indefinitely
The caution came with a counterpoint: Altman said it could be “bad for the world” if OpenAI waited too long to go public. He also said he considers public companies valuable, making clear that safety is not an argument against a listing in principle. His remarks instead cast timing as a balance between the demands of public ownership and the company’s ability to make responsible decisions as its models advance.
That leaves OpenAI without a disclosed IPO timetable or a public set of safety milestones that would trigger one. The next step, in Altman’s account, is to build confidence in how the company can operate with more capable systems and make safety claims it can stand behind. For now, he says, that work takes priority over setting a date for Wall Street.