What the Lawsuit Actually Claims
Four paid subscribers to ChatGPT, Claude, Grok, or Gemini filed a proposed class-action antitrust suit against Anthropic, OpenAI, SpaceXAI, and Google on Friday, September 19, 2026, in the US District Court for the Northern District of California. Their claim: the four companies illegally agreed to slow the pace of AI development, and that coordinated restraint reduced the value of what subscribers pay for every month. Lead attorney Nick Rowley is seeking nationwide class-action status on behalf of paying subscribers across all four services.
The plaintiffs are explicit that they are not objecting to any single company choosing to prioritize safety over speed. Their claim is narrower: that an agreement among direct rivals to move slower together, rather than compete on speed individually, is itself an anticompetitive act, because it substitutes collective restraint for the individual accountability that competition is supposed to enforce.
The Essay That Triggered It Already Asked for Legal Cover
The complaint's central evidence is a September 12, 2026 essay by Anthropic CEO Dario Amodei, urging industrywide cooperation on decelerating AI development in favor of stronger safety measures. The same day, OpenAI's Sam Altman, SpaceXAI's Elon Musk, and Google DeepMind's Demis Hassabis each responded publicly in agreement. The complaint also points to a July 2026 statement signed by employees across several labs, acknowledging what it called intense competitive pressure not to unilaterally slow down, and calling on governments to support a coordinated global slowdown.
Amodei's own essay anticipated exactly this legal exposure. It flagged the antitrust risk directly and asked for government mediation or, in his words, a narrow waiver for certain kinds of safety conversations. No such waiver exists. The essay that plaintiffs now cite as evidence of an illegal agreement is also the industry's own advance warning that this conversation carried legal risk, which undercuts any later argument that the coordination was unwitting.
Ten Days Later, Two Defendants Went to War on Price
The theory faces a factual problem that arrived within days of the filing. On September 22, 2026, both Anthropic and OpenAI cut the price of their newest models by as much as 58 percent, with OpenAI pricing its new GPT-6 Sol model at exactly half of Anthropic's newly discounted Claude Opus 5.5 rate, on the day Anthropic's own price changed. That is not restraint. It is two of the four defendants openly undercutting each other's prices on a specific rival product, ten days after the essay the complaint treats as evidence of a slowdown pact.
A plaintiff does not need every fact to fit for a case to survive a motion to dismiss, but a pricing war between two defendants, timed this close to the alleged agreement, is the kind of contrary evidence a defense team builds an entire argument around.
What Happens Next
None of the four companies had responded to requests for comment as of the weekend the suit was filed. Antitrust law generally treats public statements followed by parallel conduct with real skepticism unless a plaintiff can show something more than coincidence, a rule that exists precisely because rivals commenting on each other's public statements is normal, lawful behavior in most industries. What makes this case unusual is that the alleged agreement was announced in public, by name, by four chief executives, rather than inferred from private conduct, which is both what makes it easy to plead and what will make it hard to prove without more. Whether a public safety essay can be read as the kind of concerted action antitrust law is built to catch is now a live question for a federal court, and the answer will shape how openly AI labs can discuss safety cooperation with each other going forward.
Read next: Microsoft Writes Down What Everyone Else Endorsed | The Best AI Hacking Tools Are Now Invitation-Only
