Sam Altman rules out an OpenAI IPO this year and says uncontrollable AI is possible

Key points

• Sam Altman confirmed in a September 2026 interview with Fortune that OpenAI will not hold an IPO this year.
• Altman said it is "absolutely" possible to build an AI that humans can no longer control.
• He vowed to pause AI training rather than allow development to cross that line.
• The 45-minute interview also covered a hacking incident at Hugging Face, an AI model-sharing platform.

Sam Altman sat down with Fortune magazine this week and said two things that deserved to be said plainly: no stock market debut in 2026, and yes, a truly uncontrollable AI could be built.

On the IPO, he was direct. "We're not rushing into an IPO," Altman told Fortune, adding that going public right now would be "an ill-advised moment" given ongoing safety work. "We've said for a long time, we'll do it when we're ready. I would say not 2026." We reported on 12 September that OpenAI had already begun formal IPO paperwork with regulators, which makes the delay a deliberate choice rather than a lack of preparation.

What did he mean by "beyond human control"?

Altman said it is "absolutely" possible to build an AI system that humans can no longer oversee or stop. He didn't describe it as inevitable, but he didn't dismiss the idea either.

His commitment was to act before reaching that point. If training a new model looked dangerous, he said OpenAI would pause rather than push through. His exact words: "There are risks we should not be able to incur on behalf of humanity."

That framing matters. Altman is essentially saying that OpenAI's own financial interests would not override a safety stop. Whether the company would actually pull that brake is a separate question, and one worth watching.

The interview also touched on recursive self-improvement, the idea that an AI could rewrite its own code to become smarter without human input, and a recent hacking incident at Hugging Face, a platform where researchers share AI models. Altman did not offer detailed comments on the breach.

What does this mean for ordinary people?

For anyone considering OpenAI shares through secondary markets, the 2026 IPO door is closed. The company has raised large sums from private investors, so it isn't starved for cash, but a public listing would let ordinary investors buy in directly. That remains off the table.

The safety comments are the more consequential part. OpenAI builds the technology behind ChatGPT, which hundreds of millions of people use. Altman's admission that runaway AI is a real possibility, not a science-fiction worry, signals that the people closest to this technology take the risk seriously. His promise to hit pause is reassuring on paper. Holding him to it is the hard part.