On 30 September 2026 Judge Amit P. Mehta of the US District Court for the District of Columbia dismissed every count in the two private antitrust lawsuits publishers had brought against Google over AI Overviews. The opinion is 41 pages, it is free to read, and almost nothing being said about it this week survives contact with the document.
The ruling did not find that AI Overviews leave publisher traffic alone. It did not weigh any measurement of click loss. It assumed the plaintiffs’ facts were true and dismissed anyway, on the ground that an antitrust claim needs an agreement and no agreement was pleaded. The order that accompanies it is final and appealable rather than an invitation to try again, which means the next move is the D.C. Circuit, not a third complaint.
What happened
Judge Mehta granted Google’s motions to dismiss the amended complaints in two consolidated-for-decision cases: Chegg, Inc. v. Google LLC, No. 1:25-cv-00543 (D.D.C.), filed 24 February 2025, and Penske Media Corporation et al. v. Google LLC et al., No. 1:25-cv-03192 (D.D.C.), filed 12 September 2025. Chegg’s amended complaint carried six numbered counts; the Penske Media complaint carried seven. All thirteen were dismissed.
The Memorandum Opinion is published in full by the US Government Publishing Office as package USCOURTS-dcd-1_25-cv-03192, document -0. Every quotation below was read from that PDF on 5 October 2026. The filing date printed on it is 30 September 2026 — worth noting, because most coverage dates the ruling to 1 October, which is the day the trade press found it. Engadget’s report of 1 October 2026 gives the ruling date as 1 October; Forbes, the same day, correctly dates the opinion to 30 September.
Judge Mehta is the same judge who found Google liable for monopolisation in United States v. Google LLC and who wrote the remedies decision in that case. He opens this one by placing it in a sequence: “Before the court are two further entries in a series of private antitrust lawsuits brought against Defendant Google LLC arising from its dominance in the market for general search services.”
What the opinion actually says
The core of the dismissal is the collapse of what the plaintiffs called the fundamental bargain of the web: publishers let Google crawl for free, Google sends traffic back. Judge Mehta accepted that publishers hold that expectation and refused to treat it as a contract.
“But an expectation is not an agreement. It is simply how a general search engine works.”
He goes further on the mechanics, and this is the sentence SEO practitioners should read twice: “Google’s crawlers index publicly available websites through an automated process. Publishers, for their part, keep their content open to crawling because doing so serves their own independent interest in search traffic, not because they struck any bargain with Google.” In the court’s reading, leaving robots.txt open is a unilateral commercial choice, not consideration in an exchange.
The tying count — Penske Media’s only claim with no Chegg counterpart — failed on product definition: “It is thus hard to imagine consumer demand for one type of Google’s search results as separate from consumer demand for Google Search as a whole; demand for Google Search is inclusive of the demand for AI Overviews and any other item on Google’s SERPs.” AI Overviews, in other words, are not a second product being bolted onto search.
The monopolisation counts failed on antitrust standing: “Plaintiffs offer no allegations that Google’s conduct was directed at them specifically or that Google even viewed itself as competing with them.” The seven markets the complaints proposed — general search services, Online Publishing, Online Educational Publishing, Search Referral Traffic, Republishing Content, GAI Training Content and RAG Content — were rejected as pleaded, with Online Publishing called “implausibly overbroad” and Online Educational Publishing “hopelessly vague”. The California unjust-enrichment counts went with them: “the court declines to exercise supplemental jurisdiction over Plaintiffs’ unjust enrichment claims.”
And the judge said plainly what he thought of the outcome: “The court does not treat Plaintiffs’ alleged harms lightly. Nor is it unsympathetic to the situation publishers now find themselves in, and the knock-on consequences to journalists, educators, and other online creators whose content Google takes and repurposes without compensation.” He then quotes the standard that he “is bound to apply the law only as it is written, not how the [c]ourt or any party believes it ought to be”, and points elsewhere: “To the extent that this case highlights any deficiencies in the reach of the antitrust laws, that is an issue for Congress [or regulators] to consider.”
What the opinion does not contain
This is the part worth checking yourself, because the gap between the document and the headlines is the story. Four absences, each verifiable in the same PDF:
| Claim circulating this week | What the 30 Sept 2026 opinion contains |
|---|---|
| The court rejected the publishers’ traffic-loss evidence | No evidence was weighed. The standard applied is that “the court takes Plaintiffs’ well-pleaded factual allegations as true”. |
| The court found AI Overviews do not cut click-through | No percentage or figure for click-through loss appears anywhere in the document. It records the allegation of “significant declines in the click-through rate” and does not quantify or contest it. |
| The publishers’ numbers were thrown out | The only share figure in the opinion is quoted from the court’s own earlier remedies decision: “by 2009, 80% of all general search queries flowed through Google; by 2020, that percentage had increased to 89.2%”. |
| Publishers will file an amended complaint | The opinion grants no leave to amend and uses neither “with prejudice” nor “without prejudice”. Its closing line reads: “A final, appealable order accompanies this Memorandum Opinion.” |
That last row sets the only live date in the story. Under Federal Rule of Appellate Procedure 4(a)(1)(A), a notice of appeal in a civil case where the United States is not a party “must be filed with the district clerk within 30 days after entry of the judgment or order appealed from” — so roughly 30 October 2026 for a ruling entered on 30 September.
What the community said
Jason Kint, chief executive of the trade body Digital Content Next, gave Press Gazette the sharpest published reaction on 1 October 2026: “Publishers raised this harm in US vs Google, and when it wasn’t addressed there, they brought it directly to the courts. Now that path has been dismissed, too.” He added that “at some point, someone with authority has to address what is happening in the real world”. Neither Press Gazette nor Engadget obtained a statement from Google, Penske Media or Chegg.
On the measurement question the court never reached, the best independent US figure remains Pew Research Center’s browsing-panel study of March 2025: 900 US adults, 68,879 searches, an 8% click rate on searches that returned an AI Overview against 15% on searches that did not. Pew measured behaviour, not causation, and it is not the litigation record.
What to do this week
- Read the opinion, not the recaps. Search govinfo for
USCOURTS-dcd-1_25-cv-03192and open document-0. It is 41 pages and the dispositive reasoning sits in about six of them. - Stop briefing the “content for traffic bargain” as a right. A US federal court has now held it is an expectation. If a client deck or a licensing pitch rests on that framing, rewrite it.
- Separate your AI-surface measurement from your legal hopes. Export AI-feature impressions and clicks from Search Console monthly and keep the series, because no court is going to produce that number for you. The analytics and measurement level of the course covers the exports worth keeping.
- Re-examine what your
robots.txtis actually buying. The court treats open crawling as your own commercial choice. Decide it as one, per crawler, and document the decision. - Diarise 30 October 2026. If no notice of appeal is filed by then, the private antitrust route to AI Overviews is closed at district level and the argument moves to regulators.
Where the industry genuinely disagrees
Three questions this ruling raises are open, and this site does not hold a position on any of them.
Whether a dismissal tells you anything about AI Overviews’ effect on traffic. One camp reads the ruling as vindication: a judge looked at the publishers’ case and rejected it, so the harm narrative is weaker than claimed. The other says a Rule 12 dismissal decides legal sufficiency only, with the facts assumed true, so it is evidence about antitrust doctrine and about nothing else. The document supports the second camp’s description of the procedure and is silent on the magnitude. Nobody has published a measurement that the court considered, because the court considered none.
Whether a complaint’s figure is a measurement. The numbers attached to these cases in trade coverage — a 34.5% click-through reduction, a $30 billion licensing market by 2034 — come from the complaints as reported, not from the opinion, which contains neither. One camp treats a pleaded figure as reportable because the pleading is a sworn document; the other says an allegation with no exhibit, no sample and no methodology is a claim, and that a court taking it as true for procedural purposes is not a court confirming it.
How alarmist to be about click loss at all. The ruling changes none of the evidence on that, and the evidence still splits: large zero-click shares and a measured Pew gap on one side, AI referral traffic at a fraction of a percent of sessions on the other. The GEO and AIO level presents both without settling it.
The author’s opinion — Txema Hermoso
What follows is my view, not reporting, and it is deliberately narrow.
The defensible claim here is documentary: the opinion is dated 30 September 2026, it is published free by the US government, it resolved thirteen counts on the pleadings, and it contains no number for click loss and no leave to amend. Anyone can confirm all four in one PDF. I think that is the most useful thing to say about this story, and it is the thing almost nobody said.
I will not tell you the ruling means AI Overviews are costing you less traffic than you thought. The opinion gives no basis for that, and I would be laundering a procedural outcome into an empirical one. Nor will I tell you publishers were obviously right and the judge obviously wrong; the court’s reading of how crawling actually works — unilateral, self-interested, not bargained — matches how I have always described it when explaining how Google works, and that is uncomfortable for the bargain framing rather than for the judge.
What I would say to a site owner is narrower still: your leverage was never the courtroom. It is what your pages are worth to a model that has to cite something.
What is still unknown
Whether Penske Media or Chegg appeals, and on which count. Whether the D.C. Circuit would treat the crawl-for-traffic relationship differently from the district court. Whether a differently pleaded market definition — narrower than Online Publishing, sharper than RAG Content — could survive, since the court rejected these markets as pleaded rather than the idea of such markets. What the publishers’ actual traffic data shows, which never entered the record. And whether Judge Mehta’s pointer to Congress has any addressee: no US bill on AI answer compensation has passed either chamber.
Sources
- Memorandum Opinion, Judge Amit P. Mehta, US District Court for the District of Columbia, filed 30 September 2026, 41 pages — Chegg, Inc. v. Google LLC, No. 1:25-cv-00543, and Penske Media Corporation et al. v. Google LLC et al., No. 1:25-cv-03192. Published by the US Government Publishing Office, package
USCOURTS-dcd-1_25-cv-03192, document-0. Read in full 5 October 2026. - Docket, Penske Media Corporation et al v. Google LLC et al, No. 1:2025cv03192 (D.D.C.), filed 12 September 2025, cause 15 U.S.C. § 1 — Justia Dockets, read 5 October 2026.
- Complaint, Chegg, Inc. v. Google LLC and Alphabet Inc., No. 1:25-cv-00543 (D.D.C.), filed 24 February 2025 — CourtListener RECAP archive.
- Rick Ellis, “Google Wins Dismissal Of Penske Media, Chegg AI Lawsuits”, Forbes, 1 October 2026.
- “Judge dismisses lawsuits claiming Google’s AI Overviews siphon web traffic”, Engadget, 1 October 2026.
- “Penske AI Overviews lawsuit dismissed because ‘no formal bargain’ struck with Google”, Press Gazette, 1 October 2026 (updated 2 October 2026) — source of the Jason Kint quotations.
- “Judge Mehta dismisses Penske’s Google AI Overviews antitrust suit”, PPC Land, 2 October 2026.
- “Judge Dismisses Penske, Chegg AI Overviews Suits”, Relevant Audience, October 2026.
- Federal Rule of Appellate Procedure 4(a)(1)(A) — Legal Information Institute, Cornell Law School, read 5 October 2026.
- Pew Research Center, browsing-panel study, March 2025: 900 US adults, 68,879 searches; 8% click rate with an AI Overview against 15% without.
- Google Search Status Dashboard,
incidents.json, read 5 October 2026: no incident with a begin timestamp in October 2026; the September 2026 spam update still carries no end timestamp.