Judge dismisses Chegg and Penske antitrust suits over Google AI search

US District Judge Amit Mehta — the same judge who presided over the DOJ’s search-monopoly case — dismissed the Chegg and Penske Media antitrust suits against Google’s AI search on October 1. The dismissal is on the merits: the plaintiffs could not show the kind of arrangement antitrust law reaches.

The facts

• What plaintiffs claimed: Chegg said Google scraped its educational content to train Gemini and diverted traffic; Penske (Rolling Stone, Variety) alleged its indexed content was harvested for AI answers with no opt-out, costing traffic and revenue.
• Why it failed: Mehta found no formal arrangement between Google and publishers — plaintiffs claimed only an “expectation” of search traffic, and “an expectation is not an agreement. It is simply how a general search engine works.”
• The judge’s posture: Mehta wrote that he is “not unsympathetic” to publishers, but antitrust law cannot substitute for legislation.
• The wider picture: the UK has already ordered AI-search opt-outs, the European Commission is weighing similar rules, and Google’s pilot paying about 100 publishers drew a poor reaction.

Where publishers go from here

The ruling pushes remedies onto three non-litigation tracks: legislation (opt-out regimes advancing in the UK and EU), licensing (per-use payments or revenue shares), and technical enforcement (robots rules and paywalls). Google’s paying-publisher pilot flopped commercially, which suggests the licensing market still lacks a price benchmark.

What it means for content sites

For sites living on search traffic, no legal rescue is coming soon: AI Overviews diversion continues, and courts have now confirmed that being indexed carries no promise of traffic. The practical shift is measurement — from rankings and clicks to citations and conversions — plus an explicit stance on robots rules and licensing terms.

Editorial take

The same judge who found Google illegally monopolistic in search dismissed these suits — the difference is between exclusionary agreements and product evolution, and that distinction will be cited in every AI-content case that follows. Reddit killing RSS and its public API is the supply-side response to the same pressure, and every point of quality models like Argon gain makes diversion heavier. The decisive node next is legislative: the UK is done, the EU is deliberating.

The precedent being set

Mehta’s ruling matters mostly for the sentence it turns on: “an expectation is not an agreement.” Every publisher suit against an AI answer engine will now be filtered through that line — the claim has to show exclusionary dealing, not lost visitors. Chegg’s narrower claims (its stock collapsed after students moved to chatbots) failed for the same reason: the harm came from competition, not conspiracy.

The losing strategy

Both plaintiffs built their cases on scrapes and rankings rather than on any dealing Google offered. That choice looks strategic in hindsight: framing AI answers as “search” pulled the case into antitrust, where Google wins; framing it as “copyright” or “contract” — the route the New York Times took against OpenAI — keeps it alive. Expect the next wave of publisher litigation to change legal theory, not just defendants.

The legislative track decides this

The UK’s opt-out regime and the EU’s deliberations are now the load-bearing path for publishers. Australia has floated similar rules, and US proposals surface every session without moving. Sites that want payment for AI use should be watching those bills, not dockets.

One more consequence worth naming: the dismissal removes the discovery process that plaintiffs were using to document Google’s content practices. Private licensing talks now happen without a courtroom paper trail, which weakens publishers’ negotiating position further and makes the legislative record — hearings, filings, responses — the only public archive of this dispute.

The docket closes, the policy question does not.

For now, the archive of what Google’s AI search actually did to these publishers stays sealed inside the companies that lived it — and the next case will start from a worse information position than this one did.