Safety cartel faces California court as major model subscribers sue four giants for antitrust collusion

Four of the world’s biggest artificial-intelligence companies, Anthropic, OpenAI, SpaceX AI, and Google, face a class-action lawsuit filed in the Northern District of California, accusing them of violating antitrust law by entering a unlawful agreement to coordinate a slowdown of their technology development. The suit, brought by lawyers representing four subscribers who pay regular fees for ChatGPT, Cloud, Grok and Gemini services, says that collusion among these competing parties to curb the pace of innovation limits free competition in the market and deprives subscribers of the full value they pay for those subscriptions.

The case papers rely on a chain of events, most notably the article published by Dario Amodei, CEO of Anthropic, on September 12, urging the entire sector to cooperate in slowing model development on safety grounds. On the same day, Sam Altman, CEO of OpenAI, Elon Musk, CEO of SpaceX AI, and Demis Hassabis, chair of Google DeepMind, publicly endorsed Amodei’s proposal. The lawsuit notes that this convergence began to materialise in practice in July 2026, when senior staff at those labs signed a statement acknowledging intense competitive pressures that prevent any party from slowing down unilaterally, and called for government intervention to support international efforts to rein in autonomous systems.

The plaintiffs do not object to any company independently choosing to reduce the pace of its research for safety reasons, but they argue that U.S. antitrust laws prohibit replacing individual responsibility with collective agreements that produce slower results than natural competition yields.Nick Riley, the lead counsel for the plaintiffs, warns that leaving safety protocols to private agreements serving the interests of giant for-profit companies risks letting AI escape human control.Amodei had acknowledged the possibility of facing antitrust action, urging the U.S. government to issue a limited exemption that would permit joint discussions on safety, while Altman remarked that his company sees no need to wait for legislation or a judicial exemption before beginning to coordinate trust-building measures.

This coordination runs into a hard political wall in Washington, where President Donald Trump rejected calls to rein in the industry, describing attempts to curb the pace of development as a conspiracy that would drive companies toward bankruptcy, announcing the formation of a task force and appointing an AI czar to ensure U.S. labs stay ahead of their Chinese counterparts. Republican leaders in Congress also took a hard line, with Senator Josh Hawley stating that granting an antitrust exemption to the most powerful commercial entities in history would enable collusion and choke the market.The legal and political clash reveals the fragility of major firms’ attempts to impose uniform safety standards independent of federal commercial law.

These judicial developments impose new calculations on institutions and firms in the Gulf, Egypt and the Arab Levant, which rely in their digital infrastructure and advanced projects on the APIs and enterprise subscriptions of those four companies. A freeze in the pace of improvements or a coordinated slowdown among commercial providers, coinciding with the ongoing lawsuit, means that regional CTOs face paying full-subscription costs for a restricted, legally uncertain upgrade path. If you lead a technical team or a project that depends on these models, the direct professional takeaway is to break the single-source reliance on the major Western providers and redirect part of the budget toward open-weight models and independent local infrastructure, anticipating any disruption that antitrust disputes in U.S. courts might cause to the software roadmaps they control.