Invest1 publisher3 min readPublished
Four subscribers turn one day of AI safety agreement into an antitrust class action
The complaint, filed six days after Dario Amodei's Sept. 12 essay, builds its conspiracy out of four public statements made on a single day. It asks a court to value capability that was never shipped.
The Investor · Invest desk
What happened
- A lawsuit filed Friday in the Northern District of California claims Anthropic, OpenAI, SpaceXAI and Google made an illegal deal to slow their AI development, in violation of antitrust laws.
- The alleged coordination is dated Sept. 12, when Dario Amodei published an essay urging the industry to cooperate on decelerating advances in favour of enhanced safety measures.
- Sam Altman, Elon Musk and Demis Hassabis each publicly responded to Amodei's proposals in agreement that same day, according to the lawsuit's account of events.
- Four named plaintiffs who pay for ChatGPT, Claude, Grok or Gemini are suing on behalf of a proposed nationwide class of other paid subscribers to those services.
Compiled by The InvestorSomething wrong?How this is made
Why it matters
- exposure The pleaded agreement is made of public statements, so a CEO's next post about pacing is pre-collected evidence, and the four labs' subscription revenue is the pool a certified class would claim against.
- constraint The narrow safety waiver that would clear these conversations needs a sponsor in Washington, and both Hawley's refusal and Trump's treatment of limits as a conspiracy leave the labs coordinating in the zone the complaint targets.
- decision Each lab now chooses between endorsing slower pacing on the record and routing the discussion through a government-mediated channel, and OpenAI's stated refusal to wait for an exemption is the position the plaintiffs are using.
- precedent If a court lets same-day public endorsement of an essay stand in for concerted action, any industry that coordinates safety in public inherits the same pleading.
The alleged injury runs through product quality. Plaintiffs say the coordinated slowdown would reduce the value consumers get for paid AI subscriptions [2]. Damages would measure the gap between what a Claude or ChatGPT subscription delivered after Sept. 12 and what it would have delivered had the four labs kept racing [3][4]. Nobody trained that faster model. Fortune did not report a damages figure or a size for the proposed class [6].
As described, the conspiracy is an essay and three replies. Amodei published on Sept. 12 urging industrywide cooperation on decelerating advances in favour of stronger safety measures, and Altman, Musk and Hassabis each responded publicly in agreement the same day [3][4]. The plaintiffs argue that an agreement among the chief rivals that their progress "should be slower than competition would otherwise produce has an anticompetitive effect on consumers" [5]. Nick Rowley, the lead attorney for the plaintiffs, said "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" [7].
Amodei acknowledged the antitrust problem in the essay itself, writing that it would be helpful for the U.S. 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" [9]. Altman said 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" [10]. The complaint arrived six days after the essay [16]. Representatives for Anthropic, OpenAI, Google and SpaceXAI did not immediately respond to Fortune's request for comment [8].
The waiver Amodei described requires someone in Washington to grant it. Sen. Josh Hawley said in a recent Senate hearing that "there is no world" in which he would agree to give "the most powerful companies in the history of the world" an exemption from antitrust laws to collaborate [14]. Trump rejected the calls for regulation on social media, calling efforts to limit the technology part of a "conspiracy" and questioning why industry leaders would seek rules that "if strongly implemented, will drive them into oblivion and bankruptcy" [12]. He said Saturday that he is forming an AI task force and will appoint an "AI czar", with scant detail [13].
The case can go three ways. A court reads four same-day public statements as parallel conduct and dismisses, in which case the labs have bought a legal bill. Or the complaint survives to discovery, and the price is internal pacing documents at four companies. Or Washington issues the narrow waiver, which Hawley's stated position makes the least likely of the three [14]; the plaintiffs say they do not oppose the companies asking for one [11].
I would expect the second path to be the expensive one, whatever the verdict. Four labs now have to treat every public statement about pacing as a document in a class action, and a published essay with three replies was the cheapest coordination channel they had [4]. The evidence against that would be in the docket. Fortune reports that the lawsuit places the coordination "largely" on Sept. 12 [3]; if the complaint also pleads private cross-lab communications, the dismissal odds change.
What to watch
- Whether the companies' first filing is a motion to dismiss arguing parallel public conduct, and whether the docket shows any private cross-lab communications.
- Whether Trump's AI task force and the promised AI czar touch the narrow safety waiver Amodei asked for, given Hawley's refusal to grant an exemption.
- Whether any of the four CEOs publicly endorses coordinated pacing again now that the statements are pleaded as an agreement.