There is a principle, and then there is a competitor taking your customers. Anthropic is learning, publicly, that these two things do not always coexist.
Anthropic is considering rolling out a new AI model to counter OpenAI’s momentum, according to Reuters, citing three sources, in a move that follows CEO Dario Amodei’s call for the global AI community to slow the pace of releasing new capabilities to address safety concerns. The gap between that call and this reported reversal: nine days.
“We must slow the pace at which we improve the capabilities of AI models,” Amodei wrote in an essay published on September 12. The essay proposed a three-step plan aimed at slowing development without sacrificing commercial advantage, and Amodei said Anthropic had “unilaterally” committed to the first step, which would embed third-party evaluators with employee-level access to verify safety practices. It drew public support from OpenAI CEO Sam Altman and Elon Musk. Nine days later, Reuters reported that Anthropic may ship anyway.
The competitive pressure is real. OpenAI released GPT-6 Astra on September 3, touting gains in computer use, software engineering, cybersecurity, and professional work. Reports citing Ramp data say Astra quickly captured roughly 13% of enterprise AI spending, overtaking Anthropic’s Claude Fable, which sits at around 8%. Reports also say Astra helped OpenAI claim the top spot on OpenRouter, a widely used platform routing developer traffic across AI models, marking the first time in over two and a half years that OpenAI has led Anthropic on that metric.
For investors studying Anthropic ahead of any listing, the more important question is what this episode reveals about the business itself.
Anthropic still holds a substantial revenue advantage. Bloomberg, as reported by TechCrunch, said Anthropic’s annualized revenue run rate surpassed $65 billion at the end of July, up from $9 billion at the end of 2025, while OpenAI’s revenue run rate was around $40 billion. The revenue lead is genuine. But the deeper risk is concentration: large customers can change course, and a run-rate figure can obscure how quickly that can show up in headline growth.
The deeper issue for a prospective shareholder is identity. Anthropic built its valuation, its enterprise relationships, and its recruitment story on being the safety-first alternative. Anthropic said it confidentially submitted a draft S-1 to the SEC on June 1. The company was valued at about $965 billion in a late-May funding round, Reuters reported. Whatever the timing, investors buying into a listing are pricing a specific identity: a company that earns premium positioning through credible restraint. If that restraint buckles whenever a rival ships a model, the premium deserves scrutiny.
A disciplined long-term investor would not dismiss Anthropic outright. The revenue trajectory, the enterprise relationships that remain, and the genuine technical capability in Claude’s model family are real assets. But the Mogul framework demands honesty about what competitive moats are actually made of. Safety positioning is a moat only if it holds under pressure. This week, it did not obviously hold.
Reuters, citing three people familiar with the matter, says Anthropic is weighing a launch partly in response to OpenAI’s momentum and is evaluating the safety of its next model. Anthropic declined to comment. That distinction matters: the report establishes that Anthropic is considering a release, not that it has approved one. A final decision to wait would partly rehabilitate the principle. A decision to ship would confirm that enterprise market share governs, and that Amodei’s essay was, at minimum, aspirational.
Either outcome is instructive. The companies that compound wealth over decades do so because their competitive advantages are structural, not situational. Anthropic’s answer to this moment will tell investors more about the durability of its moat than any revenue figure will.
