Paid AI subscribers sue four labs over their slowdown pledge
Four people who pay for ChatGPT, Claude, Grok and Gemini are suing the companies that sell them, arguing that the labs' public push to "pace" frontier development is an illegal agreement to restrain trade.
The class action — Buist v. Anthropic PBC, No. 3:26-cv-10693 — was filed Friday in the U.S. District Court for the Northern District of California and names Anthropic, OpenAI, SpaceXAI and Google. The plaintiffs are four consumers who hold paid subscriptions to the four services, and they are asking the court to certify a nationwide class of everyone else who does. Their claim is not that slowing down is wrong in itself: the complaint says a lab deciding to move cautiously is its own business. What antitrust law forbids, they argue, is competitors taking the "shortcut" of substituting "collective restraint for individual accountability."
The complaint says the four defendants control roughly 80% of the paid consumer subscription market for frontier AI models, and that an agreement among the closest rivals in the field that their progress "should be slower than competition would otherwise produce" is anticompetitive on its face. The plaintiffs want an injunction halting the coordination and a declaratory judgment that the antitrust laws were violated. Notably, they are not asking for money — no damages are specified in the complaint.
The filing traces the alleged agreement to a specific sequence. On September 12, Anthropic chief executive Dario Amodei published an essay urging the industry to pace itself; within days Elon Musk and OpenAI chief executive Sam Altman publicly endorsed the sentiment. The complaint also reaches back to July, pointing to a statement signed by senior employees at several leading labs that acknowledged the "intense competitive pressure not to unilaterally slow" development, and to a July working group meeting on an industry standards body that the plaintiffs characterize as the groundwork for collectively throttling output.
The uncomfortable detail for the labs is that Amodei saw this coming. His essay acknowledged the antitrust problem directly, writing that it would help if the U.S. government would mediate the cross-lab talks "or at least enable" them — the government wouldn't have to take part, but would need to "issue a narrow waiver for certain kinds of safety conversations." Altman responded that OpenAI welcomes a "federal framework that sets consistent safety requirements" but does not believe the industry needs to wait for an antitrust exemption to start. Instead of a waiver, the labs got a complaint that quotes the request. Nick Rowley, the plaintiffs' lead attorney, put the stakes bluntly: "AI will quickly spin out of human control and could kill us all if we allow AI safety and protocol ... to be controlled by private self-serving agreements between the world's most powerful 'for profit' technology companies."
Antitrust cases built on public statements and inferred agreement are hard to win — courts have historically wanted evidence of an actual deal, not a shared mood, which is why the July working group matters more than the essay. If discovery turns up a plan rather than a discussion, the case has teeth. None of the four companies responded to requests for comment, and the political cover the labs sought is not arriving either: Senator Josh Hawley said at a recent hearing that there is "no world" in which he would grant the most powerful companies in history an exemption from antitrust law to collaborate. We covered the legal question the labs were asking — OpenAI asked Congress if an industry-wide AI slowdown is even legal — earlier this month.
What to watch: whether a defendant moves to dismiss on the pleadings, and whether notes from the July standards-body meeting surface in discovery.
Should an agreement to slow down for safety be legal, or is that a cartel by another name? Tell us in the comments.
