Four AI Subscribers Sue the Labs Over a Public Slowdown Pact

Four people who pay for AI chatbot subscriptions sued Anthropic, OpenAI, SpaceXAI and Google in the US District Court for the Northern District of California on Friday 18 September 2026. Buist v. Anthropic PBC alleges a horizontal agreement to slow the rate at which the four improve competing products, in breach of section 1 of the Sherman Act. It seeks treble damages, an injunction and a jury.

The complaint dates the agreement to 12 September 2026. That was the day Anthropic chief executive Dario Amodei published an essay, "We Must Pace the Frontier", calling for industry-wide cooperation on decelerating capability advances in favour of safety work. OpenAI's Sam Altman, SpaceXAI's Elon Musk and Google DeepMind co-founder and chair Demis Hassabis each responded publicly in agreement the same day, the Associated Press reported on 19 September. The four named plaintiffs are bringing the case on behalf of a proposed nationwide class of paid subscribers to ChatGPT, Claude, Grok and Gemini.

The waiver the defendants asked for themselves

Amodei's essay acknowledged the antitrust problem. He wrote that it would help for the US government to mediate "or at least enable" cross-lab discussions, and that it would need to "issue a narrow waiver for certain kinds of safety conversations". Altman replied on social media that OpenAI welcomes a "federal framework that sets consistent safety requirements" but that "we do not believe we need to wait for an anti-trust exemption or legislation to begin the work of providing this confidence", according to the Associated Press. The complaint's answer is flat: no such waiver exists, Congress has enacted no exemption, and no agency compelled the conduct.

A cartel case built from press clippings

The Next Web, which read the 29-page filing, reports that every factual source in it is a named executive's public statement or a press report, and that paragraph two describes the agreement as "proposed in public, accepted in public, and confirmed in public". There are no internal documents, no confidential witnesses and no meeting minutes. Bloomberg Law reported the filing. That is unusual for a cartel claim, and it cuts both ways. The evidence is easy for a court to verify, and the conduct alleged is conduct nobody tried to conceal.

The same reading exposes the weak points. The Next Web reports that the complaint identifies no slowed release, no cancelled model and no altered training run anywhere in its 29 pages, and that the claimed 80% share of US paid subscriptions is pleaded on information and belief rather than measured. The four acceptances are not equivalent either. Musk's was four words on social media, and Google's endorsed the direction while leaving implementation open.

The owned judgment is that the plaintiffs have picked the one theory that dodges the hardest question. By pleading the restraint as naked, they make the danger of unrestrained competition a matter for Congress rather than a defence a court may weigh. Whether a judge accepts that framing when the alleged agreement consists of four public essays and posts is untested, and the complaint concedes that the full effect on released products has not yet appeared. Representatives for Anthropic, OpenAI, Google and SpaceXAI did not respond to the Associated Press's request for comment on Saturday 19 September.