OpenAI postpones IPO due to AI security concerns

Could artificial intelligence get out of control? In view of unclear security risks, OpenAI boss Altman is postponing his IPO. Competitor Anthropic is sticking to its plans. Collaboration to defend against threats is preparing the industry.

ChatGPT developer OpenAI will no longer go public in 2026, according to its boss Sam Altman. Given the current security concerns surrounding artificial intelligence (AI), this would be a bad time for an IPO, Altman said in an interview with Fortune magazine. The company feels no pressure in this regard. Asked if that puts an IPO off the table in favor of 2027, Altman said it won't be 2026. There is still a lot of work to be done to meet safety requirements and figure out how the industry and governments can work together.

The New York Times newspaper reported in June that OpenAI was considering postponing a possible IPO with a valuation of $1 trillion to 2027.

According to the report, Altman also suggested that leading AI companies could be close to an agreement to slow development and work together to mitigate risks. The head of OpenAI rival Anthropic, Dario Amodei, had previously called for a more careful approach. The speed at which the capabilities of AI models are improved must be slowed down. Progress will still appear rapid. But the time gained must be used wisely. Altman agreed on X, saying this has been a major topic at OpenAI in recent weeks.

Anthropic sticks to IPO

At Anthropic, security concerns have not yet slowed down the IPO plans. According to insiders, the company wants to start marketing its IPO in mid-October at the earliest and complete it before the US congressional elections in November.

Concerns about potential harm from AI continued to grow this week after Anthropic researcher Jacob Coxon resigned. To justify this, he said that he saw an ever-increasing threat from AI and the possibility that it could kill on a massive scale by the end of the decade. The United Nations (UN) had also warned that AI could endanger the existence of humanity.

There have recently been a number of incidents in the industry in which AI agents from developers such as OpenAI have hacked external systems or attempted to access them. This has increased concerns about the increasing power of the models and a possible loss of control.