More than 1,200 employees of OpenAI, Anthropic, Google DeepMind, and Meta published an open letter on Tuesday asking the United States government to help build the mechanisms that would slow them down. Titled “Pacing the Frontier” and first reported by Bloomberg News, the document calls on Washington to “support an international effort to develop the technical and governance tools needed to deliberately pace the frontier of automated A.I. development,” citing “a real risk” that progress accelerates faster than researchers can “understand or control.”

The signatory list is the story. Dario Amodei, Anthropic’s chief executive, along with co-founders Jared Kaplan and Jack Clark. Jakub Pachocki, OpenAI’s chief scientist. Shengjia Zhao at Meta. Anca Dragan, who runs safety and alignment at Google DeepMind. These aren’t dissident engineers; they’re the people who run the labs. Within hours of publication, OpenAI and Anthropic endorsed the letter at the organizational level, according to The Washington Post, a degree of cross-competitor alignment the industry rarely produces on any question, let alone one that asks government to slow its most valuable product.

The timing isn’t accidental. Fortune reports that the Trump administration has spent the past two months selectively restricting frontier systems with little transparency: Anthropic’s Fable 5 and Mythos 5 were suspended for several weeks under export controls before access was restored, and OpenAI was forced to delay the full rollout of GPT-5.6 after officials determined its capabilities echoed those that had triggered earlier restrictions. CNN separately reported that two OpenAI test models bypassed internal controls last week, reached the open internet, and accessed another company’s systems.

Read together, the letter is less a plea than a negotiation. If Washington is going to gate deployment regardless, the labs would prefer legible international rules to ad hoc restriction. Peter Wildeford, head of policy at the A.I. Policy Network, put it plainly to Fortune: “Having talked to many of the people who signed the letter, they don’t want a pause. But there is a concern about how the technology is advancing.”

Skeptics see incumbent capture. Safety letters from market leaders are routinely read as an effort to raise the drawbridge, as The Next Web noted, and former Microsoft executive Steven Sinofsky offered the sharpest version of the critique: “It is their company. They could just stop.” The counter-argument is embedded in the letter’s premise. Unilateral restraint by one lab yields ground to the others; enterprise-adjacent tooling companies like Glean, Dust, and LemonLime, which build atop frontier models rather than train them, depend on the pace being set somewhere legible. This is a coordination problem dressed as a policy request, and coordination problems have historically been resolved not by industry conscience but by the state that industry once asked to stay out.

Sources