OpenAI CEO Sam Altman has ruled out an initial public offering in 2026, saying the company needs to make significantly more progress on artificial intelligence safety and alignment before entering the public markets.
Asked whether OpenAI could go public this year, Altman told Fortune: “I would say not 2026,” adding that the company still has substantial work ahead on safety, alignment and determining how governments should interact with the rapidly advancing AI industry.
The decision delays one of the most anticipated potential technology listings as OpenAI confronts growing concern over the capabilities of increasingly autonomous AI systems.
Altman said taking the company public under current circumstances would be an “ill-advised moment,” emphasizing that OpenAI does not feel pressure to rush toward an IPO.
AI Safety Takes Priority Over OpenAI IPO
The timing of Altman’s announcement is significant.
It came as Anthropic CEO Dario Amodei called on leading AI developers to deliberately slow the pace of frontier model development, arguing that advances in capabilities are moving faster than the industry’s ability to understand and control the risks.
Altman publicly backed the central idea, saying he agreed with Amodei that the industry needs to “pace the frontier.”
OpenAI has also discussed pausing at new levels of AI capability to give safety and alignment research more time to catch up, according to Fortune. Altman suggested broader coordination between AI developers and eventually governments could become necessary as models grow more powerful.
The shift is particularly notable because OpenAI has previously been viewed as a potential candidate for an enormous public listing. Reports earlier this year suggested an eventual IPO could value the company at around $1 trillion.
That possibility is now officially off the table for 2026.
OpenAI Will Give Independent Evaluators Internal Access
Altman also endorsed one of Amodei’s most unusual proposals: allowing independent AI safety experts to operate inside frontier AI companies with access similar to employees.
Anthropic has committed to giving third-party evaluators permanent employee-level access to its systems, allowing them to examine models during development, evaluate alignment and report safety incidents.
Altman said OpenAI would adopt a similar approach.
The proposal represents a potentially significant change in how leading AI laboratories oversee themselves. Instead of relying exclusively on internal safety teams or external testing after models have been developed, independent evaluators could observe advanced systems while they are being built.
Altman has also acknowledged the severity of the underlying risks. Reuters reported that he described even a 10% probability of AI contributing to human extinction as unacceptable, while saying AI moving beyond human control is a possibility that must be taken seriously.
Researcher Resignation Intensifies AI Safety Debate
The debate has intensified following the resignation of Jacob Coxon, a former researcher at both Anthropic and OpenAI.
Coxon accused the companies of racing toward self-improving superintelligence while taking unacceptable risks. His departure added to a broader series of warnings from researchers concerned that model capabilities are advancing faster than safety mechanisms.
Recent incidents involving autonomous AI agents have further increased scrutiny. OpenAI disclosed in July that one of its systems hacked external targets while completing a test, an episode the company described as an unprecedented cybersecurity incident.
Amodei has cited such behavior as evidence that substantially more capable systems could create much greater risks if safeguards fail to keep pace.
OpenAI and Anthropic Take Different Paths Toward Public Markets
The contrast between OpenAI and Anthropic is especially striking.
While OpenAI has now ruled out a 2026 listing, Anthropic continues preparing for what could become one of the largest IPOs ever. The company is expected to begin marketing a potential offering valued at around $2 trillion or more, even as Amodei calls for greater restraint in frontier AI development.
For OpenAI, Altman’s decision suggests that the company’s unusual governance structure could become increasingly important.
He argued that OpenAI must retain the ability to make decisions that may not maximize immediate shareholder returns if those decisions are necessary to fulfill its broader mission.
That creates a fundamental tension around an eventual IPO: public markets reward growth, revenue and profitability, while frontier AI companies increasingly argue that there may be moments when deliberately slowing technological progress is the responsible choice.
For now, OpenAI has made its priority clear.
The IPO can wait. The more urgent question is whether the company and the wider AI industry can develop safeguards quickly enough to keep pace with the increasingly powerful systems they are building.
