A group of four consumers filed a federal antitrust lawsuit on Friday against Anthropic, OpenAI, SpaceXAI and Google, accusing the AI industry's biggest players of illegally conspiring to slow the pace of artificial intelligence development — and, in the process, of shortchanging the paying customers who rely on their products.
The complaint, filed in the U.S. District Court for the Northern District of California, alleges that the companies' coordination violates the Sherman Act, the 1890 law that bars competitors from agreeing to restrain trade. The plaintiffs are seeking class-action status on behalf of paid subscribers who, they argue, will be harmed if the companies slow improvements to the products they sell, Politico reported. For more context on this story, see our ongoing AI news.
What the Lawsuit Alleges
The suit names four of the most recognizable names in artificial intelligence: Anthropic, the San Francisco-based developer of the Claude chatbot; OpenAI, the maker of ChatGPT; SpaceXAI, Elon Musk's AI venture, formerly known as xAI and now folded into SpaceX; and Google.
At the heart of the complaint is a simple theory: when direct competitors agree among themselves to hold back product improvements, that is not safety policy — it is a restraint of trade. The plaintiffs contend that subscription customers of the four companies stand to lose concrete value if agreed-upon "pacing" means fewer capability upgrades for the same monthly fee.
"Humanity deserves ironclad safeguards when it comes to extinction event threats," said Nick Rowley, an attorney for the plaintiffs, in a statement. But those safeguards, he argued, should come from governments rather than from private coordination among rivals.
The 'Pace the Frontier' Essay at the Center
The lawsuit is a direct response to one of the most consequential AI policy interventions of the year. Earlier this month, Anthropic CEO Dario Amodei published an essay urging AI companies to coordinate industrywide to slow the rate at which they improve model capabilities while safety testing catches up — a position the complaint characterizes as the seed of an illegal agreement.
According to reporting cited in the lawsuit's aftermath, the essay quickly drew public support from across the industry. OpenAI CEO Sam Altman and Demis Hassabis, co-founder of Google's DeepMind AI lab, each agreed with the proposal within hours of its publication, and Musk also publicly endorsed it.
A footnote to Amodei's proposal acknowledged a legal obstacle: for joint industry safety talks to happen at all, the government would need to mediate or enable them — including by issuing a narrow antitrust waiver. That waiver idea has drawn sharp opposition in Washington, and it is now also the backdrop for a private antitrust suit seeking to stop the coordination from happening at all.
Washington Pushes Back on a Waiver
The antitrust waiver proposal has met resistance on Capitol Hill. Sen. Josh Hawley, R-Mo., told FBI Director Kash Patel at a Senate Judiciary Committee hearing on Tuesday that he would not give "the most powerful companies in the history of the world" an antitrust exemption.
That skepticism spans the political spectrum. Progressive critics of the big AI labs worry that coordinated pacing entrenches incumbent giants and raises prices; conservative critics have framed it as collusion by a handful of California companies. The newly filed lawsuit gives that bipartisan unease a courtroom vehicle.
'Self-Serving Agreements,' Plaintiffs Argue
Rowley, one of the attorneys who filed the lawsuit on behalf of the four plaintiffs, said in a statement to Politico that the case was filed to ensure that "private self-serving agreements between the world's most powerful 'for-profit' technology companies" do not result in AI "quickly spin[ning] out of human control."
The statement captures the plaintiffs' delicate position. They are not arguing that AI safety is unimportant — their attorney explicitly endorses "ironclad safeguards." Their claim is procedural and structural: decisions about how fast a transformative technology should move belong to elected governments and regulators, not to a handful of executives coordinating among themselves.
What Happens Next
The complaint is at the very beginning of its procedural path. To survive, the plaintiffs will need to convince a judge that their alleged harm — paying for products whose improvement has been deliberately slowed — is a recognized injury, and that the defendants' public endorsements of pacing amount to, or foreshadow, an actual agreement rather than parallel public posturing.
For the AI industry, the suit marks an uncomfortable new phase in the slowdown debate. What began as an argument over essays and op-eds has now moved into court, where the same coordination Amodei proposed as a safety measure will be examined through the lens of the Sherman Act. However the case unfolds, it raises the stakes of every handshake, joint statement and safety pact among frontier labs — and it ensures that the question of who decides the pace of AI progress will be litigated, not just debated.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
