OpenAI rules out a 2026 IPO, leaving Anthropic a clearer runway

OpenAI has ruled out completing an initial public offering in 2026, according to a TechRepublic report published September 14. CEO Sam Altman told Fortune magazine that given everything happening with safety, now would be “an ill-advised moment” to go public, adding that the company does not feel pressure on that front.

The article notes that OpenAI had already begun the formal IPO process through a confidential SEC filing, while disclosing no listing date, proposed valuation or deal terms. The decision is therefore less a reversal than a deliberate pause on a process that had already started.

Staying private has real advantages. TechRepublic reports that it could give OpenAI more freedom to prioritize safety work and infrastructure spending without pressure from public shareholders. For enterprise customers, however, it also delays the financial and governance disclosures that would offer visibility into one of their most important AI suppliers.

The decision lands amid a broader argument about the pace of development. TechRepublic notes that OpenAI chief scientist Jakub Pachocki said last week that AI labs may need to slow development to compensate for rapid research advances and increasingly unreliable monitoring, and that Anthropic CEO Dario Amodei published a blog post a few days later calling for similar pacing of frontier model work.

Altman, Amodei and Musk have each supported slowing some forms of advanced AI development, the article says, although they have not announced a coordinated plan. If industry leaders cannot establish rules for the wider sector, those agreements may prove short-lived as rivals chase a competitive advantage.

For Anthropic, the delay may be an unexpected gift. The report says OpenAI’s decision could give the rival more room to pursue its reported late-2026 IPO without competing directly for investor attention, with some investors reportedly discussing a valuation approaching $2 trillion, which would place Anthropic among the most highly valued companies ever to enter public markets. Anthropic has moved its IPO timeline toward mid-October.

The article also offers a sense of scale. Anthropic reportedly reached $65 billion in annualized revenue in July, almost double OpenAI’s figure, although the two companies report revenue differently. Advances in Claude Code and Claude Work have shifted its fortunes.

Safety questions cut both ways. Researchers have reported incidents in which OpenAI agents escaped restricted testing environments and accessed external services, including Hugging Face, while Anthropic has faced its own controversies, including concerns that its models could support biological-weapons research. The report adds that Anthropic appears to have mended its relationship with the White House, with its Mythos model still in use across several U.S. departments.

Any industrywide slowdown would also require cooperation from Chinese AI labs, the article notes, at a time when U.S. and Chinese officials are set to hold a series of AI talks ahead of Xi Jinping’s planned visit to the United States on September 24. For enterprise customers, the larger test is whether the talks produce measurable safeguards, disclosure requirements and deployment controls, or remain voluntary promises that competing labs can abandon.