Judge dismisses Chegg and Penske antitrust suits over Google AI Overviews
A federal judge found Chegg and Penske Media had not plausibly pleaded a traffic-for-content bargain or the product markets needed to support their antitrust claims against Google.
On September 30, U.S. District Judge Amit P. Mehta granted Google’s motions to dismiss Chegg’s and Penske Media’s amended antitrust complaints. He found that neither publisher had plausibly pleaded the agreement, product markets or competitive injury its theories required.
The decision removes the complaints as currently pleaded. It is not a verdict after trial. At this stage, the court accepted well-pleaded factual allegations as true and asked whether they stated plausible legal claims; it did not weigh evidence or determine that Google’s conduct is lawful under every potentially applicable law.
An expectation of traffic was not an agreement
Chegg and Penske argued that publishers made their content available to Google’s crawler in exchange for referral traffic. They alleged that Google disrupted that reciprocal arrangement by using publisher material in AI Overviews, which can answer queries above traditional search links and reduce the need to click through. That alleged arrangement is distinct from Google’s separate publisher-payment pilot for AI answers.
Mehta found that the complaints described an expectation about how search worked, not an agreement between Google and the publishers. They did not plausibly allege offered and accepted terms, negotiations over quantity, duration or exclusivity, commitments by either side, or communications showing mutual assent. Google’s encouragement of high-quality online content also did not amount to a promise that it would deliver traffic or provide any particular volume of referrals.
That conclusion defeated the publishers’ reciprocal-dealing claims. Their related Section 2 theory failed on the same foundation: reciprocal dealing presupposes a dealing relationship, and the complaints did not establish a meeting of minds or an actual reciprocal arrangement that could qualify as exclusionary conduct.
The proposed product theories also fell short
Penske’s tying claim treated Google Search and AI Overviews as separate products. The court found no plausible allegation of distinct consumer demand for them. Instead, the complaint supported the inference that both perform the same basic function—answering search queries—and acknowledged that another search engine, DuckDuckGo, also places a generative summary above conventional results.
Claims tied to a general-search market failed for lack of antitrust standing. The publishers alleged lost subscriptions and uncompensated use of content in publishing markets, while the asserted restraint on competition occurred in general search. The court found that mismatch insufficient to connect their alleged injuries to the market in which competition was supposedly reduced.
The narrower publishing-market theories ran into different problems. Penske’s proposed Online Publishing market grouped material as varied as news, blog posts, legal briefs, fiction and Chegg’s question-and-answer repository without a workable explanation of why consumers would treat it all as substitutable. Penske also did not quantify Google’s share of that proposed market or plead facts showing a dangerous probability that Google would monopolize it.
Chegg’s proposed Online Educational Publishing market was too vague for the court to assess substitution. Features such as pedagogical purpose, curation, verification and authority did not clearly separate the proposed market from other nonfiction material. The complaint also did not plausibly establish that Google participated in the market as Chegg defined it, because it alleged no comparable curation or verification in the AI Overview process.
After dismissing the federal antitrust claims, the court declined to exercise supplemental jurisdiction over the publishers’ California unjust-enrichment claims. That disposition did not decide whether Google was unjustly enriched under state law.
More news

MIT and Sakana AI propose SIFT to prioritize coding-agent benchmark runs

Google outlines verifiable private federated learning system for Gboard

AWS details multi-turn RL workflow for SageMaker search agents
