Anthropic CEO Dario Amodei wants the AI industry to slow down, but his own company may be preparing to step on the accelerator again. According to Reuters, Anthropic is considering releasing a new AI model to counter growing momentum from OpenAI, just days after Amodei called for the industry to deliberately moderate the pace of AI development.
Anthropic has a rather large competitor-shaped problem
All this makes for a rather awkward situation. Amodei recently argued that AI capabilities are advancing so quickly that safety measures need more time to catch up. In his September 12 essay, he wrote, “We must slow the pace at which we improve the capabilities of AI models.” He proposed measures including permanent access for independent third-party evaluators, coordinated safety standards, and international cooperation.
The reason Anthropic may still be considering another model is fairly straightforward: OpenAI’s GPT-6 Astra is gaining traction with businesses, putting pressure on Anthropic’s position in the enterprise AI market. Reuters reports that three sources familiar with the matter say Anthropic is weighing a new release, although the company has not publicly confirmed that it is preparing one.
That competitive pressure matters even more with Anthropic preparing for a potential IPO. Investors are reportedly reassessing the company’s position as OpenAI gains enterprise momentum, creating an uncomfortable choice: slow down development for safety, or keep pace with rivals that aren’t necessarily hitting the brakes.
The AI race isn’t exactly built for taking a breather
Amodei’s concerns have received support from other major AI figures. OpenAI CEO Sam Altman and SpaceX CEO Elon Musk backed calls for a more deliberate pace, while Google DeepMind’s Demis Hassabis has also acknowledged the need for greater coordination around AI safety.
But the competitive reality remains stubborn. If one company slows down while another keeps releasing increasingly capable models, the incentive to keep pushing becomes difficult to ignore. Anthropic’s reported dilemma therefore captures the bigger problem facing the entire industry: everyone may agree that the AI race needs a speed limit, but nobody seems particularly keen on being the first to take their foot off the accelerator.
RELATED COVERAGE:
- Anthropic previews new standard to streamline AI-to-machine connections
- Claude chat and Cowork just became one, and you’ll never have to pick a lane again
- Claude can now use your computer while you get other things done
- Claude can now join your Slack channels and work alongside your team
- Claude now remembers what you told it, wherever you use it
