News · 2026-09-13

Altman says an OpenAI IPO in 2026 would be ill-advised given everything happening with safety

OpenAI chief executive Sam Altman said the company will not go public in 2026, telling Fortune on 12 September 2026 that “given everything happening with safety, right now would be an ill-advised moment to go public.” The remark ties OpenAI's listing plans directly to the industry's week-long argument about slowing frontier AI, while its closest rival, Anthropic, is reported to be moving ahead with its own offering.

Key facts

• The commitment: “I would say not 2026,” Altman said, according to Fortune.
• When: the interview was conducted on Friday 11 September and published by Fortune on 12 September 2026.
• Context: TechCrunch reports OpenAI has filed confidentially for an IPO; Bloomberg reported a day earlier that OpenAI slowed parts of model development over safety.
• Primary sources: Fortune's report and TechCrunch's write-up.

What Altman said

According to Fortune's account, Altman said the company does not “feel pressure” to list, and explained the delay in terms of the moment rather than the business: “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.” He added a line that reads like an argument against public-market discipline: “We need to be able to make decisions that are not obviously in the interest of our business and our shareholders.” TechCrunch's transcription has him saying: “We're not rushing into an IPO.”

The comments came a day after Anthropic's chief executive asked the industry to pace frontier AI and Altman replied that OpenAI would match the call for embedded outside evaluators. Earlier in the week OpenAI had said it will slow or stop development if needed, without naming a threshold.

Has OpenAI actually slowed down?

Fortune's interview does not itself say OpenAI slowed training; it reports that the company has discussed pauses as it reaches new capability levels. Other reporting goes further. Bloomberg, in a story carried by Business Standard on 11 September, reported that “OpenAI said it has slowed parts of model development and paused certain internal AI training recently due to safety concerns.” In July, Altman said OpenAI paused training after a sandbox security incident.

Why an IPO and safety collide

A public company owes its shareholders a steady story of growth, and quarterly results punish anything that looks like voluntarily holding back a product. Imagine a carmaker announcing it will deliberately slow production of its best-selling model because it is not sure the brakes are good enough, in the same quarter it asks investors for billions. Altman is arguing OpenAI cannot credibly do both at once.

CNBC's Deirdre Bosa made the point directly on X: “The more OAI and Anthropic argue that commercial pressure can push AI companies to move too fast, the harder it is for either to go public.” Public markets, she wrote, “just add more of that pressure: growth expectations, quarterly reporting, a stock price moving every second.” Investor Brad Gerstner replied: “Disagree. Anthropic will IPO. The market knows how to price risk - see SpaceX.”

Anthropic's path

Anthropic confidentially filed its IPO prospectus in June, according to CNBC. Reuters reported on 5 September, via CNBC, that the prospectus is now not expected to become public until late September, with marketing to investors starting mid-October at the earliest. A search of the SEC's EDGAR filing system on 13 September found no public registration statement from Anthropic.

The extinction-risk question

Reuters' report of the interview focused on what Altman said about catastrophic risk. The remarks it quotes do not name a figure he endorses: “Whether it's 10 or eight or six, the point is, we all have a tremendous amount of responsibility.” That framing, a shared responsibility regardless of the exact odds, is the same one Altman used to explain why a public listing can wait.

The caveat

“Not 2026” is a short delay, not a change of plan, and it arrives when three and a half months remain in the year. Altman's safety framing is also convenient: it recasts a timing decision that could have many financial causes as a principled one, and nothing in the interview specifies what safety milestone would make a listing appropriate.

Primary source, verified: read the paper →

Ground Truth. (2026, September 13). Altman says an OpenAI IPO in 2026 would be ill-advised given everything happening with safety. Ground Truth. https://groundtruth.day/news/altman-says-an-openai-ipo-in-2026-would-be-ill-advised-given-safety.html

@misc{groundtruth:altman-says-an-openai-ipo-in-2026-would-be-ill-advised-given-safety,
  title  = {Altman says an OpenAI IPO in 2026 would be ill-advised given everything happening with safety},
  author = {{Ground Truth}},
  year   = {2026},
  month  = {sep},
  url    = {https://groundtruth.day/news/altman-says-an-openai-ipo-in-2026-would-be-ill-advised-given-safety.html}
}

Topics: openai · ipo · ai-safety · anthropic · finance · sam-altman