Skip to content
Markets & Trends

Sued For Agreeing To Slow Down: The Antitrust Case Against Anthropic, OpenAI, Google And SpaceXAI, And Why It Matters To Anyone Who Buys AI

A week after Dario Amodei's essay calling for industry-wide pacing of AI capabilities - and the same-day agreement from Sam Altman, Elon Musk and Demis Hassabis that sent chip stocks down 5.8% - four paying subscribers sued. Their proposed class action in the Northern District of California argues that Anthropic, OpenAI, Google and SpaceXAI violated antitrust law by coordinating to restrain AI development, cutting the value of the subscriptions they pay for. The complaint leans on a July statement signed by senior staff at several labs acknowledging 'intense competitive pressure not to unilaterally slow'. It does not challenge any company choosing to slow on its own. Whatever the outcome, it has put a question in front of every business that depends on these models: who decides the pace, and what happens to your roadmap if they agree?

AlchmAI Editorial11 min read

4

Named plaintiffs - paying subscribers to ChatGPT, Claude, Grok or Gemini - seeking to represent a nationwide class

4

Defendants: Anthropic, OpenAI, Google and SpaceXAI, sued in the US District Court for the Northern District of California

12 Sept

The day Amodei published his pacing essay and Altman, Musk and Hassabis publicly agreed - the core of the complaint

-5.8%

Philadelphia Semiconductor Index on the following Monday, the market's own verdict on what 'pacing' might mean

It took a week. On 12 September Anthropic's Dario Amodei published an essay calling for industry-wide cooperation to slow the pace at which frontier AI capabilities improve, in favour of safety. Sam Altman, Elon Musk - now running SpaceXAI - and Google DeepMind's Demis Hassabis each publicly agreed the same day. Chip stocks fell 5.8% on the Monday. And by the end of the following week four paying subscribers had filed a proposed class action in the Northern District of California arguing that the agreement itself was illegal.

The legal theory is straightforward antitrust. Competitors, the plaintiffs argue, are not permitted to agree among themselves to restrain the product they compete on - and slowing capability improvements, by agreement, reduces the value of the subscriptions customers pay for. The complaint points to a July 2026 statement signed by senior employees at several labs acknowledging 'intense competitive pressure not to unilaterally slow' development and calling on governments to back a worldwide deceleration. The plaintiffs' line is blunt: the antitrust laws do not permit competitors to decide among themselves that competition is too dangerous. Importantly, the suit does not challenge any company choosing to slow its own development independently.

Why This Is Everyone's Problem, Not Just The Labs'

It would be easy to treat this as a Silicon Valley legal story. For businesses that have built products and operations on these models - and financial services has done so at scale - it surfaces three questions that deserve an answer whatever the court decides.

  1. 01Your roadmap depends on a pace you do not control. Many firms' AI plans assume the next model generation arrives on schedule, cheaper and more capable. The labs have now said in public that they might coordinate to slow it. That assumption belongs in a risk register, not a slide.
  2. 02Concentration is now visibly correlated. Four providers, one essay, same-day agreement. If your resilience plan is 'switch to another frontier provider', the week of 12 September shows that the providers can move together. The Bank of England has already named AI-provider concentration as a stability concern.
  3. 03Regulation may come from courts and not from Parliament. The UK has chosen not to legislate for AI; the US is largely not regulating either. A federal judge ruling on whether safety coordination is lawful could shape the industry's behaviour more than any statute in 2026.

“The case asks an odd question - can competitors agree to be careful? - but it forces a useful one for every customer: what is my plan if they do?”

What Sensible Firms Are Doing

  • Building to the models they have, not the ones they are promised. Systems designed around today's capability, with tested value now, are insulated from any change in pace.
  • Keeping a provider-agnostic layer. A thin abstraction between applications and model APIs - with evaluations run against more than one provider - turns 'what if they slow down or change terms' into a configuration question.
  • Considering open and regional models for suitable workloads. The Cohere and Aleph Alpha combination and the UK's sovereign AI programme exist partly because concentration in a handful of US labs is a strategic risk.
  • Putting AI supply risk on the board agenda. The same governance applied to any critical supplier - concentration, substitutability, exit plan - applies to model providers.

The Bottom Line

Four subscribers have sued Anthropic, OpenAI, Google and SpaceXAI for allegedly agreeing to slow AI development, pointing to the public 12 September essay and endorsements and a July statement admitting no lab would slow alone. The case has a strong factual base and a hard question of harm, and it lands with no federal framework yet in place to authorise coordination. For the businesses built on these models, the lesson does not depend on the verdict: the pace of AI is set by a few companies that can move together, so plans should rest on today's capability, a provider-agnostic architecture, and the same supplier governance you would apply anywhere else. That is how we design AI systems as a fintech AI agency in London - for the models that exist, with an exit from any one of them.

References & Further Reading

AI antitrustAI slowdownAI Agency UKFintech AI Agency LondonAI AutomationAI regulationvendor risk
Share Email
AI

AlchmAI Editorial

Research and analysis, London

The AlchmAI team writes about the markets, technology and regulation we work with every day. We build trading platforms, real-time charts and AI analysis tools for brokers, prop firms and fintech teams from our office in Mayfair, London. Every article lists its sources. Nothing we publish is investment advice.

This article is general information and commentary. It is not investment advice or a recommendation to buy or sell any investment. Important information

Keep reading

All articles