What Is the Pacing the Frontier Letter, and Why Did Anthropic Sign It?

What Is the Pacing the Frontier Letter, and Why Did Anthropic Sign It
Takeaways

    • Pacing the Frontier is a July 2026 letter in which 1,224 frontier-lab employees ask Washington to build tools to slow automated AI research.
    • Anthropic’s founding leadership signed in force, the same summer the company refused to join the rival Open Weights letter backed by roughly 50 firms.
    • The request is aimed at the US government and names zero European institutions, leaving the EU AI Act and Europe’s open-weight bet outside the room.

In the same July that Anthropic stood alone among the major closed labs in refusing to sign the Open Weights letter, its founders put their names to a very different document. Pacing the Frontier is a statement from 1,224 employees of frontier AI companies asking the US government to support an international effort to build the tools needed to “deliberately pace the frontier of automated AI development.”

Behind the careful phrasing is a simpler admission. Nobody will slow down first, so they want Washington to make everyone do it at once.

The signatory list is the story. Dario Amodei signed, alongside Anthropic co-founders Jared Kaplan, Jack Clark, Benjamin Mann and Chris Olah. So did OpenAI chief scientist Jakub Pachocki, Google DeepMind’s Anca Dragan and Meta’s Shengjia Zhao. The same OpenAI and Google that backed open weights now co-sign a call to pace the frontier. The labs want diffusion and a brake pedal at once.

Why Should Europe Care About the Pacing the Frontier Letter?

Europe should care because the request goes to Washington, not Brussels. Europe spent three years building the EU AI Act as its governance play, yet this letter names no European institution and asks a US administration to lead. A US-run pacing regime means governance tools that can throttle frontier progress, and anything built to throttle the frontier can throttle its diffusion too. That diffusion is the ground Europe is standing on: Mistral as its own sovereign bet, Kimi K3 and GLM-5.2 as the Chinese models Hugging Face reached for when the closed American systems failed.

This is rival labs jointly admitting that self-improving research could outrun oversight, and asking for guardrails before a crisis, not after. Some signatories say they want to build pacing mechanisms voluntarily.

Still, the optics are hard to miss. The people driving fastest have signed a petition for speed bumps.

Author: Juno H. Sheng

See Also

Why Did Anthropic Refuse to Sign the Open Weights Letter?

Is Kimi K3 the Model That Breaks Europe’s Sovereign AI Bet?

Does Anthropic Want to Ban Open-Weights Models? Dario Amodei Says No

Share this article

Latest news

Subscribe to our newsletter

More News