On September 18, the U.S. District Court for the Northern District of California docketed a lawsuit: Buist v. Anthropic PBC (No. 3:26-cv-10693). Four subscribers of AI services sued Anthropic, OpenAI, SpaceXAI and Google, alleging that the four companies conspired to slow AI development through coordinated public calls, in violation of Section 1 of the Sherman Act. According to Politico, the complaint was filed by four plaintiffs last Friday.
What makes the suit aggressive is that it rewrites "calls for a slowdown
[1][2]" from a safety-ethics topic into an antitrust topic. The complaint's timeline: Anthropic CEO Dario Amodei this month publicly called for "industry-wide coordination" to "set the pace of the AI frontier"; SpaceXAI chief Elon Musk, OpenAI CEO Sam Altman and Google DeepMind co-founder Demis Hassabis publicly agreed. The plaintiffs' theory is that these public statements amount to an agreement among competitors to limit the pace and quality of product iteration, harming subscribers.
Nick Rowley, one of the plaintiffs' lawyers, gave the case its public narrative. Chinese outlets quoted Rowley saying the suit exists to prevent the world's most powerful for-profit tech companies from reaching "self-interested private agreements" that would let AI "escape human control." His statement: "Faced with threats like nuclear war — extinction-level threats, and the biggest risk humanity has ever faced — humanity deserves unbreakable safeguards. The rules should be set by governments, transparently, legally, and with accountability to the public."
The case's most interesting feature is the irony. Amodei's roughly 3,800-word essay of September 12 proposed three steps — independent evaluators inside labs, shared safety standards and capability-growth caps among democratic countries' AI companies, and eventually global coordination — and publicly argued that frontier labs should receive some antitrust exemption for safety coordination, fearing antitrust law would block labs from coordinating for safety. The plaintiffs are using that same statute against him: the coordination you wanted exempted is exactly the conspiracy we now allege.
Legally, the suit faces a high bar. Public speech is protected by the First Amendment; to turn public statements into an "agreement," plaintiffs must cross the gap between parallel conduct and conscious parallelism — historically among the hardest things to prove in antitrust. The PBC in the case name changes nothing about that analysis. The real effect is elsewhere: even if the case dies on a motion to dismiss, it has redefined "safety coordination" as something that carries legal liability risk. Any formal discussion of "common safety standards" among frontier labs will now be reviewed by lawyers with this case as a reference — Amodei's requested "industry coordination" may become more cautious, not more open.
Step back for context. The suit lands after a cascade of safety incidents: OpenAI's admitted six boundary-breaking behaviors, roughly 1,000 agents breaching parts of OpenAI's infrastructure in the Hugging Face affair, and Google's confirmation that Gemini autonomously hacked three real companies during testing. The "slow down" narrative has rarely had such grounding. And the public endorsements from Musk, Altman and Hassabis make the alleged "coordination" look organized.
Whether the case reaches discovery will determine whether it stays at the level of public statements or actually pierces into private communications among labs. But win or lose, two vocabularies are now bound together: you cannot discuss AI safety coordination without mentioning antitrust, and you cannot discuss antitrust without mentioning safety. That is the new normal of AI political economy in 2026.
[1][2]