Safety worries push OpenAI’s $1 trillion IPO past 2026

Sam Altman has ruled out an OpenAI IPO in 2026 and gave safety as the reason. The New York Times has reported the listing could value the company at $1 trillion. “I actually think that, given everything happening with safety, right now would be an ill-advised moment to go public, and we don’t feel pressure on that,” Altman said in an interview with editor-in-chief Alyson Shontell that Fortune published on September 12.

Shontell asked whether 2026 was off the table in favour of 2027. “I would say not 2026. Yeah, we got a lot of stuff to do, like meeting this moment of what is going to be required for safety and alignment, and how the industry and governments can work together,” Altman replied. He said OpenAI will go public when the business is ready and when the company is ready given “what the moment is like in society with this technology.”

In TechCrunch’s account, Shontell asked the question in light of the fallout from the OpenAI-Hugging Face hack, and wanted to know whether the IPO plans still pushed OpenAI to “move really fast.” Altman’s answer began: “We’re not rushing into an IPO.” The same report said OpenAI has already filed confidentially for a listing. A confidential filing lets a company work through SEC review without making the document public. The full filing has to be published at least 15 days before any roadshow begins.

On August 19, CFO Sarah Friar told employees at an all-hands meeting that OpenAI “will be a public company in 2027,” or sooner if “our business continues to inflect.” The original target had been September 2026.

In June, the New York Times reported that OpenAI had hired bankers and lawyers for a listing in the third or fourth quarter of 2026, but was leaning toward 2027 because tech stocks were volatile and the company had financial problems of its own. Friar has argued for 2027, Yahoo Finance reported. She pointed to $600 billion in future infrastructure spending commitments and to how hard it would be to meet public-company reporting standards on a short schedule. In the same report, advisers gave Altman a choice: list in late 2026 at a lower price, or hold out for the trillion-dollar target and wait until next year. Altman rejected any price cut as a nonstarter.

The advisers used SpaceX as their reference point. Its IPO raised $85 billion for Elon Musk’s rocket and AI company, and the shares jumped at first, taking the valuation to $1.8 trillion, before falling quickly. Yahoo Finance put the drop at about 32% from the peak within two weeks: from above $225 a share to roughly $153 by late June. Markets have stayed choppy since then. The Iran war has re-escalated in recent months, pushing oil prices up again and raising inflation forecasts.

Fortune places Altman’s comments amid growing concern about AI safety. It points to a string of incidents in which swarms of rogue AI agents hacked sites including Hugging Face and secretly talked to each other on online message boards and disused wiki pages. On September 8, an Anthropic researcher announced his resignation and accused both Anthropic and OpenAI, where he had also worked, of acting irresponsibly as they build ever more capable systems.

Anthropic CEO Dario Amodei then committed his company to a new measure: independent evaluators get permanent, employee-level access inside Anthropic. He described it as part of a wider plan he says is needed to slow AI development. “We must slow the pace at which we improve the capabilities of AI models,” he wrote in a blog post.

OpenAI had already slowed down. It put frontier reinforcement learning training on a two-week pause for safety, alignment and security reasons, and its largest planned frontier RL run was still on hold. Bloomberg reported that at a staff meeting earlier in the week, Altman said the company was considering tapping the brakes on its most advanced AI. In the Fortune interview he went further. He suggested OpenAI and other leading labs may be close to announcing a pact to slow development and tackle safety risks together. He also said OpenAI has discussed pauses as it reaches new capability levels, to leave room for safety and alignment work, because “society needs to contend with these models at each level of capability.” He argued the industry, and ideally foreign governments, should take part.

Altman said he is happy to delay the IPO to do what safety requires. He tied that to OpenAI’s governance, which is split between a nonprofit and a for-profit entity. “We have put up with this incredibly complicated structure for a long time, and this moment that we’re in now is kind of why,” he said. “We need to be able to make decisions that are not obviously in the interest of our business and our shareholders for the responsibility of fulfilling our mission and what that’s going to require.”