What happened
OpenAI and Anthropic are working together in Washington to shape a federal review system for advanced AI models, Crypto. News reported Tuesday. The proposal, as summarized in the report, would give federal reviewers a 30-day window to examine frontier systems before they ship to the public.
Neither company has released a joint statement of its own, and the exact scope of what counts as a frontier model is not spelled out in the reporting. The two labs are otherwise competing head-to-head for enterprise contracts, hyperscaler compute, and researcher talent, which makes a co-signed policy ask unusual on its face. Sam Altman's OpenAI and Dario Amodei's Anthropic have taken opposite public postures on regulation in the past, with Anthropic historically the more forward-leaning of the two on federal oversight.
The joint move, per the reporting, lands after months of White House signalling that a lighter-touch executive order framework may replace the Biden-era approach.
Why it matters
A 30-day pre-release review changes the release calendar. Frontier launches from OpenAI, Anthropic, Google DeepMind, and Meta have been the metronome of the AI trade since GPT-4 shipped in March 2023. A federal check-in that adds a month to that cycle affects three things at once: the pace of model announcements, the flow of enterprise deal talk that tracks those announcements, and the crypto-side narrative trades that piggyback on both.
It's also a hedge play for the two firms. Getting in front of the rules OpenAI and Anthropic will have to follow anyway lets them shape what a frontier model even is, which smaller labs and open-source projects will have to inherit. That is the part of the story worth flagging as editorial: this looks like safety diplomacy, and it partly is, but the practical effect is a compliance moat around the two firms that can afford a 30-day slip.