OpenAI Delays IPO: Sam Altman Bets on Self-Improving AI

OpenAI Delays Its IPO

OpenAI CEO Sam Altman recently informed his employees via Slack that an initial public offering (IPO) could materialize "within the next year." Although the company has already filed a prospectus, it primarily aimed to maintain flexibility regarding the timing of the IPO. A delay until 2027 would, however, be unexpected, especially as Anthropic, a direct competitor, appears poised to go public in the near future.

Altman justifies this delay by several factors. On one hand, Anthropic's growth figures seem more robust, which could disadvantage OpenAI if it were to proceed with an IPO immediately. On the other hand, OpenAI continues to consume a significant amount of cash, which could impact its valuation in the event of a premature public offering.

Strategic Reasons Behind the Delay

Sam Altman explained that advancements in self-improving AI could justify remaining a private company longer. "Technology and the world can change in surprising ways," he stated, suggesting there might be good reasons to delay the IPO to adapt to these developments. However, he also mentioned that the capital needs for computing infrastructure could eventually accelerate the process.

In parallel, Altman announced a stock sale aimed at employees, with a price set at $687.69 per share. This initiative could provide some liquidity to employees while strengthening their commitment to the company.

A New AI Model in the Works

OpenAI is not just delaying its IPO; the company is also preparing to launch a new AI model, version 5.6. According to Jakub Pachocki, head of research at OpenAI, this model represents a significant advancement over GPT-5.5 and could be launched as early as June. This announcement comes as Anthropic has already set a high bar with its Fable and Mythos 5 models, intensifying competition in the artificial intelligence sector.