What happened
On 18 September 2026, four paid subscribers filed Buist et al. v. Anthropic PBC, OpenAI OPCO LLC, SpaceXAI, and Google LLC in the U.S. District Court for the Northern District of California (No. 3:26-cv-10693), alleging the four labs made a Section 1 Sherman Act agreement to collectively slow the pace of AI development after Dario Amodei's 12 September 2026 essay urging a slowdown and CEOs' public alignment with it. The complaint argues coordination reduces the value paid subscribers receive. It follows the Senate (Hawley/Cruz) blocking a national-security antitrust exemption in the NDAA on 15 September that would have shielded safety collaboration.
Why it matters
This is the first major antitrust challenge to frontier-lab safety coordination and directly tests whether industry-wide AI slowdown/guardrail pacts constitute an output-restricting cartel. It creates legal risk precedent that affects how labs can publicly coordinate safety measures without triggering collusion exposure, and intersects directly with ongoing federal/state AI-governance debate.
Action needed
AI labs and industry bodies should monitor the motion-to-dismiss phase and avoid joint statements/commitments on development pacing that could be characterised as coordinated output restriction; counsel should reassess safety-collaboration channels and NDAA exemption advocacy.