- Publisher Lawsuits: Google won dismissal of education firm Chegg and publisher Penske’s lawsuits over alleged traffic losses from its AI search summaries.
- Content Exchange: Judge Amit Mehta found no plausible agreement to trade publishers’ content for search referrals, defeating their central antitrust argument.
- Legal Options: Both dismissals can be appealed; Penske’s order preserves possible refiling, while Chegg’s does not expressly address it.
- Click Behavior: Pew Research Center observed traditional-result clicks on 8% of visits to AI-summary pages versus 15% without summaries in March 2025 browsing data.
Google has won the dismissal of lawsuits by education company Chegg and publisher Penske Media alleging that AI-generated search answers use their content while depriving their websites of visitors and revenue. In a September 30, 2026 ruling, U.S. District Judge Amit P. Mehta found that the companies had not adequately pleaded the antitrust violations on which their cases depended.
The ruling dismisses both amended complaints against Google and its parent, Alphabet, in the U.S. District Court for the District of Columbia. At this stage, Mehta treated properly pleaded factual allegations as true and assessed whether they supported viable legal claims.
Why Content Access Was Not an Agreed Traffic Exchange
AI Overviews generate answers within Google Search, with links to supporting websites. Google introduced the summaries in the United States in May 2024, adding an on-page way to obtain information alongside traditional links that take users elsewhere.
Penske alleged that a reader satisfied by a generated answer has less reason to visit the website supplying the information. Its sites depend on those visits for advertising revenue, affiliate commissions and some subscriptions. Chegg, which sells learning support and access to educational question-and-answer resources, likewise alleged lost subscription revenue as Google answered users’ questions itself.
Both companies described allowing Google to index their content as an exchange: they supplied information, and Google supplied visitors. Their reciprocal-dealing claims under the Sherman Act, the federal competition law, alleged that Google conditioned those referrals on free access to content for snippets, AI training and generated responses.
Mehta found the alleged exchange lacked a plausibly pleaded agreement. Such an agreement could be unwritten or implied. But the complaints identified no offered terms that the publishers accepted, negotiations over quantity or duration, commitments by either side, or communications showing mutual assent.
Mehta wrote, “But an expectation is not an agreement.” In his account, Google’s automated indexing and the publishers’ decision to allow crawling each served independent commercial interests.
That missing relationship defeated the reciprocal-dealing claims under both relevant sections of the Sherman Act. Although Section 2 can address a monopolist’s independent conduct, the particular claim of reciprocal dealing still requires a two-sided dealing relationship.
Google denied wrongdoing and argued that it was not obliged to index publishers on their preferred terms. Chegg had filed its lawsuit in February 2025; Penske’s case followed in September that year. Both companies amended their complaints before Mehta considered the dismissal motions together.
Penske’s Traffic Argument Undermined Its Tying Claim
Penske also alleged unlawful tying: forcing people who wanted Google Search to take AI Overviews with it. That claim required a plausible basis for treating the two as separate products with separate consumer demand.
The publisher argued that ordinary search sends people to other websites, while AI Overviews keep them on Google’s results page. Mehta turned to Penske’s own allegation that users satisfied by an AI answer have little reason to click onward. He reasoned that this suggested demand for information, whichever format supplied it, rather than distinct demand for two products.
The tying claim therefore failed on its separate-products premise.
Search-Market Injury and Publishing-Market Definitions
The companies’ claims that Google unlawfully maintained its search monopoly faced a separate problem: antitrust standing, which requires a sufficiently connected injury to pursue the competition claim.
Mehta found that the alleged subscription and licensing losses arose in publishing and content markets. The companies also argued that obtaining their content for free gave Google an advantage over search rivals that paid for it. But their complaints did not adequately explain how those costs excluded competitors or made the publishers’ losses an indispensable part of maintaining Google’s search monopoly.
Their claims that Google was extending its power into publishing failed on market definition. A relevant market needs a plausible account of which products can substitute for one another, so that competitive power can be assessed.
Penske’s proposed market covered essentially all text published online. Mehta found that too broad: the complaint did not explain why a legal brief, fictional story and news article belonged in one market, or quantify Google’s share of it.
Chegg proposed a narrower educational market, but its boundaries remained unclear. Chegg emphasized checked, carefully selected educational material, but those features also appear in journalism and other nonfiction. The court found that Chegg had neither adequately distinguished materials intended for learning from ordinary information nor plausibly placed Google’s answers in the curated, verified educational market it described. Those defects undermined both companies’ attempted-monopolization and monopoly-leveraging claims.
The Traffic Concern Behind the Cases
Independent click measurements help explain the publishers’ concern. In Pew Research Center’s March 2025 web browsing analysis, traditional-result clicks occurred on 8% of visits to pages with AI summaries, compared with 15% without them. Only 1% of summary-page visits led to a click on one of the first three source links within a summary.
The study used tracked browsing from 900 U.S. adults and reconstructed their search results in April. Results could have changed between those dates, and different kinds of queries were more likely to trigger summaries.
Google has offered a different account of referral value. In an August 2025 response, Search chief Liz Reid argued that AI answers encouraged longer, more complex questions and new opportunities to discover websites. Google described visits that did not immediately return to Search as higher-quality clicks. That argument concerns what visitors do after clicking, whereas Pew measured how often clicks happened.
The publishers also alleged that preventing AI reuse would sacrifice their visibility in Search. Google’s documented Search controls let site owners manage crawler access and limit snippets or exclude pages from indexing. Google-Extended, its control for certain other AI training and response uses, has a separate scope. Blocking Search’s crawler would remove the content access on which indexing and referrals depend.
Mehta acknowledged the alleged disruption to journalists, educators and other online creators. He said any shortcomings in antitrust law’s reach were for Congress or regulators to consider.
What the Dismissal Orders Leave Open
Both cases also included California unjust-enrichment claims, alleging that Google benefited from their content without compensating them. After dismissing the federal claims, Mehta declined to exercise the court’s authority to hear the accompanying state claims. Those claims were dismissed on that jurisdictional basis.
Penske’s order expressly dismisses its complaint “without prejudice,” preserving the possibility of a later filing. Chegg’s order grants dismissal without specifying prejudice. Both orders are final and appealable.
