Antitrust suits converge on AI labs as Google keeps its ad exchange

Four plaintiffs filed the proposed class action. Each of them pays for ChatGPT, Claude, Grok or Gemini. They claim the labs coordinated to slow AI development in a way that "reduce[s] the value consumers get for paid AI subscriptions." Tom's Hardware reported the suit on 21 September, citing the Associated Press. The plaintiffs trace the coordination to July 2026, when the leading labs signed a statement acknowledging "intense competitive pressure not to unilaterally slow" development.

The legal theory is unusual, and the case is at an early stage.

Lead counsel Nick Rowley framed the agreement as a substitute for regulation. "AI will quickly spin out of human control and could kill us all if we allow AI safety and protocol ... to be controlled by private self-serving agreements between the world's most powerful 'for profit' technology companies," he said, according to Tom's Hardware. The plaintiffs accept that development should slow for safety reasons. But they argue that a cooperation proposal from Anthropic founder Dario Amodei is a "shortcut" that "substitutes collective restraint for individual accountability."

The political reaction was split. Sam Altman wrote on X that OpenAI welcomes a federal framework for frontier AI safety but does not believe an antitrust exemption or new legislation is needed first. President Trump rejected the premise outright, calling the idea that AI is "taking over the World, destroying Humanity, and all other things bad" a "HOAX." Chinese state media also weighed in. China Daily described the proposed club as one "whose membership rules have been drafted before the guest list is announced," adding that "a global AI-safety framework that excludes China is not quite global."

A different antitrust outcome, same week

Google's position in the AI case sits alongside its separate ad tech fight, which ended differently. A judge ruled the company broke antitrust law but allowed it to keep its advertising exchange, avoiding a forced sale of AdX. The ruling, covered in the dossier's context headlines on 2 and 3 September, ended rigged auctions while leaving Google on both sides of the market, according to one headline summary.

In other words, the same competition law is producing two very different kinds of pressure: a damages claim over AI safety coordination, and a structural remedy that stopped short of a breakup.

That contrast matters for how the AI suit is read. Antitrust claims against coordination between competitors normally turn on prices and output. Here the alleged harm is the value of a subscription, which is harder to quantify and harder to prove. The plaintiffs' own filing acknowledges the safety rationale, which gives the defendants room to argue that any coordination served a legitimate purpose rather than suppressing competition.

The doomer backlash in Europe

Europe's tech industry is pushing back on the safety framing that underpins the argument. Politico reported on 24 September that executives and cybersecurity professionals are calling the US-generated panic about AI extinction a distraction. James Wise, a partner at Balderton Capital and chair of the UK government's sovereign AI fund, said of the current narrative: "Sadly, it has jumped up to the Terminator is coming. I don't think the Terminator is coming at all."

Marcus Hutchins, the researcher known for stopping WannaCry in 2017, argued that frontier labs neglect hacker behaviour in their doomsday theories. On the Hugging Face incident, he said OpenAI knew the model was good at exploiting software bugs. It still challenged it to do so with reduced guardrails, without containing it properly or monitoring it closely enough to intervene. OpenAI declined to comment and pointed to its report. Anthropic did not respond to Politico's request.

Alexandru Voica of UK firm Synthesia put the priority differently, telling Politico: "We should spend more time making the rules and standards we have work at scale, and less time focusing on fears about theoretical risks or news reports from the Silicon Valley AI doomer bubble."

Not everyone in the dossier's sources agrees on the severity. Politico notes that even sceptics accept frontier models are powerful cyber tools, and that Europe needs access to them. The disagreement is about emphasis and remedy, not about whether the risk exists.

Where the money and the rules are moving

The US government is funding grid technology rather than AI rules. On Thursday 24 September, Canary Media reported a $1.9 billion Department of Energy SPARK grant programme covering 31 projects across 26 states, with $3.35 billion in matching funds from participating utilities, for advanced conductors and dynamic line rating systems. Energy Secretary Chris Wright said the grants will "get more out of the infrastructure we already have."

Private capital, meanwhile, keeps moving. Tech.eu reported on 24 September that Swiss tech companies raised more than €1.1 billion in the first half of 2026, with semiconductors taking around €281 million, energy €223 million and healthtech €220 million, and the ten largest rounds accounting for roughly two-thirds of the total. That is a different picture from the litigation headlines: concentrated bets in capital-intensive sectors, not a market retreating from technology.

The antitrust question now sits in two places at once. One is a courtroom where four subscribers are testing whether safety coordination between AI labs is a competitive harm. The other is a ruling that let Google keep the ad exchange it was found to have used unlawfully. Both will shape how much room the largest technology companies have to act together, and neither is close to resolution.