Ultraman claims that OpenAI may go public within a year, with reports of launching the new GPT-5.6 model this month.

According to Beating Monitoring, OpenAI CEO Sam Altman predicted in an employee Slack message sent on Monday that the company is expected to go public within the next year, though the exact timeline will depend on the funding needs for its computing infrastructure and the pace of AI recursive self-improvement as it evolves.

Altman noted that if technology can achieve rapid recursive self-improvement—meaning AI can autonomously develop the next generation of AI—the world and the forms of technology could change in unexpected ways, making it more advantageous to keep the company private at that time. On the other hand, however, OpenAI’s enormous funding requirements for purchasing computing resources and building infrastructure could push the listing process forward. The company is currently negotiating to lease a large data center in Ohio and needs to raise thousands of billions of dollars in funding for NVIDIA chips and rent.

Alongside its plans to go public, OpenAI is preparing to release a brand-new AI model this month, code-named 5.6. In an employee letter, Chief Scientist Jakub Pachocki said that compared with the current flagship model GPT-5.5, the new model will deliver significant improvements.