Altman Won't Set an IPO Date, but Drops Hints on Hardware and Rivals

OpenAI CEO Sam Altman told CNBC on Tuesday he has no particular timeline for taking the company public, saying he wants to prioritize safety and OpenAI's mission while the technology's next generation remains in flux.

"I don't have a particular timeline in mind"

Speaking with CNBC's Kate Rooney at OpenAI's DevDay, Altman said an IPO could become realistic once the company has run "a few safety cases" and better understands how to handle the next level of AI. "I really think this is a time to put safety and mission first," he said, adding that he doesn't want the demands of being a newly public company to pull focus from that work.

A hardware tease, no launch date

Asked about hardware, Altman said OpenAI has "something that's worth waiting for" coming, without giving a timeline. "I am not worried about being first with new hardware. New hardware is very hard to do," he said, when asked about rivals' recent device and chip rollouts.

On agent safety and liability

Altman called Nvidia's newly released agent-safety platform, aimed at stopping AI agents from misbehaving, "a good thing," but said it's "not a full solution." He argued that aligning AI models is a scientific challenge, not one engineering measures alone can solve. On liability, he said he expects a "multi-level" framework to emerge that distinguishes faults in an AI model from a developer's use of it or a user's intentional misuse — comparing it to how responsibility is assigned in the auto industry. "If someone's driving drunk, it's their fault," he said.

On Meta's Muse

Asked about Meta's Muse and CEO Mark Zuckerberg's recent moves into agents and hardware, Altman said Muse "seems like a nice product," while saying he isn't threatened by Zuckerberg and expressing confidence in OpenAI's own roadmap.