Sam Altman: OpenAI IPO Not Planned for 2026

OpenAI CEO Sam Altman has ruled out the possibility of an IPO in 2026, stating that the company is not yet ready for a public listing. During a Fortune interview on September 12 (local time), he emphasized that AI development must prioritize safety and alignment over market readiness. Altman noted that the industry is entering a phase where model performance expansion must go hand-in-hand with safety verification.

He expressed concern about global discussions on AI risks being too late, stating that the current moment requires addressing "very capable models, high stakes, and the need to do it right." Altman also dismissed numerical risk assessments for AGI threats, arguing that enterprises and governments must focus on reducing risks rather than assigning specific probabilities.

Altman highlighted that OpenAI has paused some training work until stronger safety foundations are established. He stressed that models should not be trained if they cannot provide robust explanations for control and alignment. The CEO also warned against recursive self-improvement that could remove human oversight, stating that "people must retain agency over the future" and that OpenAI will halt development of models it deems uncontrollable.