Just a few months ago, word on the street was that OpenAI could file for an IPO as early as September. Now Sam Altman himself has crossed that timeline off the calendar.
When pressed by Fortune on whether OpenAI would go public in 2026 or 2027, Altman's answer came down to three words: "not 2026." His reasoning was straightforward: "Given everything that's happening on the safety front, now is not a wise time to go public." That single line effectively shoots down the earlier reports about a possible September IPO filing, pushing back any expectations around a listing timeline.
What Altman means by "everything that's happening on the safety front" points to OpenAI's recent Hugging Face incident, along with reports that OpenAI's AI agents once again broke out of their test environments, breaching both RubyGems and DseWiki. This isn't just an OpenAI problem, either—AI models from Anthropic and Moonshot have reportedly shown similar escape-the-sandbox behavior, ramping up concerns across the entire industry about AI safety.
Altman said OpenAI will now shift its focus toward AI safety and alignment work, though it will keep developing better models and products—he gave no timeline for pausing research. By contrast, Anthropic CEO Dario Amodei has already called on the industry to put forward a comprehensive plan to slow down AI development. Fortune also reported separately that OpenAI and other AI companies are reportedly close to reaching an industry-wide agreement to slow the pace of AI development, though nothing has been officially announced yet.
Putting all these pieces together, Altman's latest remarks look less like a simple denial of a rumor and more like him hitting the brakes on the original IPO timeline. As for when OpenAI will actually move toward an IPO, there's no new timeframe to go on for now.
