Altman: OpenAI Should Not Go Public in 2026, AI Safety Work Not Yet Complete
nashnova research
OpenAI CEO Sam Altman said explicitly that the company will not IPO in 2026, citing unfinished AI safety work — his clearest public rejection of the timeline yet, deflating Wall Street's expectation of a landmark dual listing with Anthropic.
What exactly did Altman say?
In a *Fortune* interview published September 12, Altman stated: "Now is not a wise time to go public. Not 2026, yes."
His reason was singular: AI safety work is not done.
This means → OpenAI's internal bar for "safe enough to list" is far higher than the market assumed. The IPO is gated by safety progress, not market conditions.
What was the market betting on?
Wall Street had widely expected OpenAI and Anthropic to go public in 2026, back to back — a defining capital event for the AI industry.
Altman's statement knocked out half that thesis in one sentence.
In plain terms = the market was pricing in two AI giants ringing the bell the same year. Now one has explicitly said no, and the other hasn't spoken.
Where does the pressure shift?
The spotlight turns to Anthropic CEO Dario Amodei, who on the same day published a call to slow down AI capability development.
This reflects a growing "hit the brakes" consensus inside the AI industry — not just OpenAI, but its closest rival is also voluntarily easing off the accelerator.
The backdrop: an Anthropic researcher resigned publicly to warn of AI-loss-of-control risks, and debate over the pace of frontier-model development continues to intensify.
When can OpenAI actually go public?
Altman offered no new timeline — only that it depends on the pace of safety work.
This means → the IPO is no longer a "pick a date" decision. It is a "pass the bar first" decision — and OpenAI sets the bar itself.
In plain terms = the listing date is not about market windows. It is about when they believe AI is safe enough.
市场有风险，内容仅供研究参考，不构成投资建议。