Four of the world’s most prominent artificial intelligence developers now face a federal antitrust complaint accusing them of working together to hold back the pace of model improvement. The proposed class action, filed in the US District Court for the Northern District of California, names Anthropic, OpenAI, SpaceXAI, and Google.
It argues that an understanding among rivals to ease off capability growth violates Section 1 of the Sherman Act and leaves paying subscribers with weaker products than unfettered competition would deliver.
The plaintiffs are four consumers who subscribe to Claude, ChatGPT, Grok, or Gemini.
They seek to represent a nationwide class of other paid users.
Their core claim is straightforward: when competing labs decide together that their systems should advance more slowly than each would choose on its own, they restrict output.
That restriction, the filing says, is the kind of collective restraint the antitrust laws have long forbidden, regardless of the novelty of the technology.
According to the complaint, the public phase of the alleged pact unfolded on September 12, 2026.
That morning Anthropic chief executive Dario Amodei published an essay titled “We Must Pace the Frontier.”
In it he argued that labs should jointly set limits on the speed of unchecked progress so that no single firm would lose commercial ground by acting alone.
Within hours, SpaceXAI’s Elon Musk,
OpenAI’s Sam Altman, and Google DeepMind co-founder Demis Hassabis each signaled support.
Musk wrote that Amodei was right.
Altman said he agreed that the frontier needed pacing and committed OpenAI to an early step in the proposal. Hassabis called the essay the right path forward.
The suit maintains that the alignment did not begin that day. It points to a July 2026 statement signed by senior figures at several frontier labs.
That document acknowledged intense competitive pressure against any company that slowed by itself and urged governments to back a broader effort to moderate automated development.
The complaint treats those earlier contacts, plus subsequent working-group meetings among Anthropic, OpenAI, and Google, as the start of coordination roughly two months before Amodei’s essay appeared.
Plaintiffs emphasize that they do not challenge any company’s independent choice to invest more in safety or to release models more cautiously.
What they object to is the alleged substitution of joint restraint for individual accountability.
A firm that decelerates while rivals keep racing, they argue, risks losing customers, talent, and technical lead.
An understanding that everyone will ease off removes that penalty and, in their view, functions as an output-limiting agreement.
The complaint asks the court for an injunction, a declaration that the alleged pact is unlawful, and damages.
Counsel for the plaintiffs have described private, self-serving arrangements among for-profit labs as an inadequate way to govern technology that could escape human control.
Company representatives had not issued detailed public responses at the time the suit was reported.
The case sits at the intersection of two live debates: how society should manage the risks of rapidly advancing AI, and whether competitors may lawfully coordinate the speed of that advance.
Courts will now have to decide whether public statements of agreement, prior joint statements, and reported working-group contacts amount to an illegal combination or remain protected speech and advocacy.
