The Ruling
Judge Amit P. Mehta dismissed antitrust lawsuits brought by Penske Media — publisher of Rolling Stone, Variety, The Hollywood Reporter, Billboard and Deadline — and by Chegg, over Google's AI Overviews.
Both argued Google used its search dominance to compel publishers to supply content for AI systems without compensation: material republished in search, training data, and grounding for AI answers.
Four Grounds, and They Are Worth Reading Separately
1. There was no agreement
The reciprocal dealing theory required a contract. Mehta found none: publishers keep their sites open because it serves their own interests, not because they agreed to anything.
"An expectation is not an agreement."
That sentence is the one that will be quoted for years, and it is the load-bearing one. The implicit bargain publishers believed they had with search — we let you crawl, you send us traffic — was never written down. Mehta is saying it therefore does not exist as a legal matter.
2. AI Overviews is not a separate product
The tying claim failed because Mehta treated AI Overviews as an integrated search feature, not a distinct product. Treating each result type as its own product would fragment search implausibly.
3. The injury was in the wrong market
Plaintiffs were harmed in publishing markets, but the alleged monopoly is in general search. Mehta also noted Google reached 80% search share by 2009 — fifteen years before AI Overviews launched in 2024.
4. The markets were badly defined
Penske's "Online Publishing" was implausibly broad; Chegg's "online educational publishing" too vague to survive.
The Court Was Sympathetic and Ruled Against Them Anyway
Mehta described the court as "not unsympathetic" to the publishers, and said it is bound to apply the law as written rather than as parties believe it should be. He suggested Congress address the economic dislocation caused by new innovation.
That is a judge saying: you may well be right about the harm, and antitrust is not the tool.
Chegg had argued it invested hundreds of millions of dollars building its 135-million-question database. The ruling does not dispute that. It holds that spending money building something a competitor then benefits from is not, by itself, an antitrust injury.
What This Closes Off for Other Publishers
The reasoning generalises, which is why this matters beyond two plaintiffs:
- Reciprocal dealing theories now face a high pleading bar — you need an actual agreement, not an industry norm
- AI features built into existing products are insulated from tying claims
- Injury must be established in the search market, not the downstream market where publishers actually lose money
- Broad market definitions will not survive a motion to dismiss
Available reporting does not state whether dismissal was with or without prejudice, so whether these specific suits can be refiled is unclear.
What We Think This Means, as a Publisher
We will declare the interest: this site's content is pulled into AI answers. Bing's own report shows 9,621 Copilot citations to our pages across 313 queries in a single period. So we are describing our own position, not an abstract one.
The honest reading is that the legal route to compensation just got substantially narrower, and the practical conclusion follows from it: stop waiting for a remedy and optimise for being the source that gets cited.
That is not a satisfying answer for anyone whose business model assumed search traffic. It is the accurate one. A citation in an AI answer is worth something — brand, authority, the queries where a reader does click through — and it is currently the only compensation on offer. Google is separately reported to be paying around 100 publishers for AI Overviews content, which suggests commercial deals, not litigation, are where this gets resolved for the publishers big enough to negotiate.
For everyone else, the lesson of this ruling is that the implicit bargain was never a bargain. Plan accordingly.
FAQ
Who sued Google and why?
Penske Media and Chegg, alleging Google used search dominance to obtain content for AI Overviews without compensation.
Why were the cases dismissed?
Four grounds: no plausible agreement, AI Overviews being an integrated feature rather than a separate product, injury occurring outside the search market, and inadequate market definitions.
What did the judge actually say?
"An expectation is not an agreement," and that the court is "not unsympathetic" but bound to apply the law as written. He suggested Congress address the issue.
Can they appeal or refile?
Final appealable orders were issued. Whether dismissal was with prejudice is not stated in available reporting.
Does this affect other publisher lawsuits?
The reasoning narrows antitrust routes generally — particularly reciprocal dealing and tying theories against AI features integrated into existing products.