Four consumers have filed an antitrust lawsuit against Anthropic, OpenAI, SpaceXAI and Google, alleging that the four companies agreed to slow the pace of development of their competing AI products. The class-action complaint, filed on September 18 in the US District Court for the Northern District of California, alleges that the companies coordinated to limit AI development, restricting competition between their products.

The plaintiffs are seeking treble damages and an injunction against the alleged agreement, while also asking the court to certify a nationwide consumer class. The proposed class covers US consumers who purchased paid subscriptions to ChatGPT, Claude, Grok or Gemini.

Allegations in detail: The complaint alleges that Anthropic, OpenAI, Google and SpaceXAI coordinated to slow frontier AI development, rather than making independent decisions about their products. According to the plaintiffs, the coordination began before the public proposal: Anthropic, OpenAI and Google representatives allegedly formed a working group in July to develop an industry standards body, while senior executives had already acknowledged that no company could slow development alone without losing customers and technological ground.

On September 12, Anthropic CEO Dario Amodei proposed “industry-wide coordination” to set limits on AI progress, including restrictions on training compute, training runs and the use of AI to improve AI, alongside capability checkpoints and independent evaluators to verify compliance. The complaint says Musk, Altman and Hassabis publicly endorsed the proposal that day. It further alleges that OpenAI later confirmed weeks of collaboration with Anthropic and Google, while Altman said AI progress “should be slower than it otherwise could be”.

What Amodei proposed: The Anthropic CEO proposed to slow, rather than stop, frontier AI development by creating safeguards that keep pace with increasingly capable models. First, Anthropic will give independent third-party evaluators permanent, employee-level access to its systems so they can assess development and safety practices, report incidents and evaluate model alignment. Second, Amodei wants frontier AI companies in democratic countries to establish common safety standards and limits on unchecked AI progress, including potential restrictions on training compute, training runs and AI systems used to improve AI. Finally, he proposed broader coordination between the US, other democracies and countries including China.

At the same time, Amodei argued that the US must retain its AI lead through tighter controls on advanced chips and semiconductor equipment, action against unauthorised model distillation and stronger protection against model-weight theft. He said this lead could create more time to develop AI safely as capabilities accelerate.

Also read: