A case filed in the Northern District of California targets four major companies active in the field of artificial intelligence. The complaint alleges that Anthropic, OpenAI, xAI and Google agreed to coordinate their approach to slowing the pace of AI development, and that this action could violate U.S. antitrust laws.
The plaintiffs, who identify themselves as paying subscribers to the ChatGPT, Claude, Grok and Gemini services, claim that such coordination reduces competition among companies and, as a result, reduces the value users receive for paid subscriptions. However, the allegation’s appearance in the complaint does not mean it has been proven, and judicial proceedings on these charges will continue.
The matter began with Anthropic CEO’s call
One of the main focuses of the case dates back to September 12, 2026, when Dario Amodei, CEO of Anthropic, in an article titled “We should slow the frontiers of development at a controlled pace” called for slowing the advancement of advanced AI models.
Amodei argued that AI models’ capabilities are increasing at a pace that outstrips researchers’ ability to understand and control them. He did not call for a complete halt to AI development and emphasized creating more time to develop safety mechanisms.
The proposal was immediately met with responses from other company executives. Sam Altman, CEO of OpenAI, said he agreed with Amodei on the need to control the pace of advanced model development. Elon Musk also briefly expressed support for Amodei’s viewpoint. Demis Hassabis of Google DeepMind likewise supported the safety approaches raised.
What do the plaintiffs call collusion?
In the complaint, these public responses are presented as part of the evidence that, according to the plaintiffs, indicates coordination among industry rivals. They also point to a July 2026 statement in which some senior employees of AI labs discussed intense competitive pressure and the difficulty of unilaterally slowing development.
The plaintiffs’ argument is that companies can independently decide to slow development of some technologies for safety reasons, but should not agree with each other to limit competition. From their perspective, if rival companies agree on the pace of technological progress, that action can have different consequences than the independent decision of each company.
Amodei himself had pointed out the antitrust problem
Notably, Amodei had previously warned about the legal challenges of such cooperation. He suggested that the U.S. government, if necessary, could act as a mediator among companies or create a limited antitrust exemption for specific collaborations related to safety.
Altman also responded to the debate by saying OpenAI welcomes creating a federal framework to set uniform safety requirements, but does not believe companies must necessarily wait for an antitrust exemption to begin safety collaborations.
The issue has become more sensitive as some industry leaders warned about the emergence of capabilities like AI assistance in building the next generation of models. Amodei believes this trend could accelerate progress and reduce sufficient time for safety testing and evaluation. He also suggested independent evaluators with access levels similar to internal staff oversee the performance of leading labs; a proposal that Anthropic and OpenAI have said they intend to pursue.
Disagreement over the role of government in controlling AI
The case has arisen amid serious disagreement in the U.S. about how to regulate the AI industry. Donald Trump, the U.S. president, has opposed broad restrictions on the industry and his administration emphasizes maintaining the pace of U.S. companies’ progress, especially in competition with China. Trump has also recently announced the formation of a “special AI task force” and the appointment of an “AI czar.”
On the other hand, some lawmakers have warned against granting antitrust exemptions to large AI companies. For example, Senator Josh Hawley has opposed such an exemption, arguing that allowing large companies to cooperate outside normal antitrust frameworks could weaken competition.
Ultimately, the current case poses an important question to the AI industry: can rival companies collaborate to increase safety and slow the pace of development, or should such cooperation only occur within a framework set by government and antitrust laws? The court’s answer in this case could affect how leading companies cooperate on AI safety as well as the relationship between commercial competition and regulation in this industry.