Sam Altman Says an OpenAI IPO in 2026 Would Be Ill-Advised

Introduction

OpenAI is taking 2026 off its IPO calendar. In a Fortune interview, CEO Sam Altman said the company would not go public next year and described the current moment as an ill-advised time for an offering. His reasoning was tied less to a lack of investor interest than to the safety questions surrounding increasingly capable AI systems.

Key points

• No IPO in 2026: Altman said OpenAI will not pursue a public listing next year. The company is not rushing into the public markets and will decide when it is ready.
• Safety is part of the timing decision: He said that, given everything happening around AI safety, going public right now would be ill-advised. He also said OpenAI does not feel pressure to move immediately.
• Loss of control is considered possible: Asked whether it is possible to build an AI that goes beyond human control, Altman answered yes. He said OpenAI would take steps to prevent that outcome, including pausing training if needed.
• Some risks should not be accepted on humanity’s behalf: Altman argued that there are risks people should not be able to incur on behalf of humanity. The remark frames certain AI risks as larger than an ordinary corporate risk-management issue.
• The interview covered more than finance: The roughly 45-minute conversation also touched on the Hugging Face hacking incident, recursive self-improvement, and the prospect of advanced systems becoming difficult to control.

Why it matters

For OpenAI, delaying a public listing preserves strategic and governance flexibility. A public company must deal with recurring investor communication, market expectations, and external scrutiny. Those demands can become especially significant while a company is still changing its approach to model development, training safeguards, and long-term governance.

Altman’s comments do not mean OpenAI has abandoned the idea of going public. They also do not establish a new target year. “When we are ready” remains an open-ended condition, leaving the company room to prioritize technical and organizational preparation over a fixed financial timetable.

The more consequential part of the interview may be the connection between capital-market timing and safety. Altman’s reference to pausing training presents safety as a possible operational constraint rather than only a principle stated in policy documents. It raises a practical question: would a leading AI company accept delays, added costs, or a competitive disadvantage if a training process appeared to create unacceptable danger? The supplied interview summary does not specify what threshold would trigger such a pause, so it would be premature to infer that OpenAI already has a public, measurable standard.

For now, two points are clear. OpenAI has ruled out a 2026 IPO, and Altman is presenting safety as a central condition for the company’s next stage. As OpenAI continues developing more capable systems, the relationship between public-market ambitions and safety governance will remain an important way to evaluate its strategy.

Source: The Verge AI