OpenAI Says It Will Skip a 2026 IPO Over AI Safety Concerns

OpenAI may eventually go public, but Sam Altman says it will not do so in 2026. In a Fortune interview, the CEO called the present moment an “ill-advised” one for an IPO, saying the company still has work to do on AI safety, alignment, and the social response to more capable systems.

The decision takes a firm date off the table without setting a replacement. Altman said OpenAI will go public when its business is ready and when the wider social context around the technology is ready, too. Asked whether that meant 2027, he answered only that 2026 was out.

The readiness test

Altman’s explanation joins two questions that are often kept apart: whether OpenAI is financially prepared for public markets, and whether the technology is safe enough to advance quickly. He said safety standards are not yet at a point where capabilities should be pushed much further, and said AI beyond human control is “absolutely” possible.

That does not amount to a declared halt in development. But Altman said OpenAI has discussed pauses at new capability levels to make room for more safety and alignment work. He also said the company has discussed working with other leading AI companies and foreign governments on those issues.

A private structure built for discretion

Altman linked the timing decision to OpenAI’s unusual governance structure, split between nonprofit and for-profit entities. He said the structure exists so the organization can make decisions that are not obviously in the interests of the business or its shareholders when its mission demands otherwise.

What Altman says must be ready

• OpenAI’s safety and alignment work.
• The business itself.
• Society’s ability to contend with increasingly capable AI models.

The private-market delay is consequential because it removes a near-term deadline, but it does not define how OpenAI will judge progress. Altman’s comments establish a direction: the company is willing to defer an IPO for safety work. They do not specify the safeguards, external oversight, or conditions that would make OpenAI ready to resume a public-listing process.

A delay that raises the bar

For now, Altman is asking stakeholders to accept a slower financial timetable while OpenAI addresses risks he describes as central to the company’s mission. The next question is whether the safety case becomes measurable: pauses, shared standards, and coordination were all discussed, but none was presented as a completed, detailed plan in the interview.