On 30 September 2026, a federal judge dismissed the antitrust lawsuits that Penske Media and Chegg brought against Google over AI Overviews. The core theory, that Google owed publishers something because they expected referral traffic in exchange for their content, did not survive a motion to dismiss. If you were waiting to see whether a court would force Google to compensate the open web for AI Overviews before deciding how to respond, that question now has an answer in this venue. It changes what's worth planning around and what isn't.
Judge Amit Mehta of the U.S. District Court for D.C. dismissed the AI Overviews antitrust suits brought by Penske Media and Chegg on 30 September 2026, in a 41-page opinion. He rejected the publishers' central claim that an expectation of search traffic amounts to an implied agreement with Google, along with their tying and market-definition theories; once the federal claims fell, the court declined to keep hearing the related state-law unjust enrichment claims too. The ruling doesn't touch copyright or licensing disputes, which run on a separate legal track, and a parallel EU competition complaint about Google's use of publisher content is unaffected. Penske's dismissal was reported as without prejudice, so an amended complaint remains possible, and no appeal had been confirmed as of early October.
What the judge actually decided
Strip away the commentary and the ruling does three specific things and leaves several others untouched. That distinction matters more than the headline, because most of the coverage treats "Google wins" as a single fact when it's really a set of narrow legal conclusions sitting next to a longer list of open questions.
| Claim or question | Status after 30 September 2026 |
|---|---|
| Antitrust claim based on an implied agreement from expected referral traffic | Dismissed — the court held that expecting traffic isn't a legal agreement |
| Tying claim (bundling AI Overviews with search results) | Dismissed — didn't meet pleading requirements for a distinct product market |
| Market-definition claims | Dismissed — same pleading failure |
| State-law unjust enrichment claims (California) | Dismissed — court declined jurisdiction once the federal claims fell |
| Copyright and content-licensing claims | Not decided — a separate legal track this ruling doesn't touch |
| EU competition complaint over Google's use of publisher content | Not decided — separate jurisdiction, unaffected by this ruling |
| Direct commercial licensing negotiations with Google | Not decided — still open regardless of this outcome |
| Penske's ability to refile | Open — dismissal reported as without prejudice |
The pattern is worth noticing: everything dismissed was an antitrust theory resting on the idea that Google and publishers have an implicit deal. Everything still open runs on a different legal basis entirely — copyright, direct contract, or a different regulator's jurisdiction. The ruling closes one door. It doesn't touch the others.
Why "expecting traffic" isn't a contract
The reasoning behind the dismissal is the part worth understanding, because it's the piece that will keep resurfacing in every future attempt to bring a similar claim. Penske's theory was a form of reciprocal dealing: publishers make content available expecting search traffic in return, Google takes the content for AI Overviews without the traffic flowing back at the same rate, and that imbalance should count as an antitrust violation.
The court rejected the premise before it reached the imbalance. An expectation isn't an agreement, and accepting that it was would mean Google holds an implied contract with every one of the billions of websites it crawls, each with its own unstated terms a court would somehow have to infer. That reasoning doesn't just defeat this case; it forecloses the entire category of claim that depends on "we expected clicks and didn't get enough of them" as a legal theory. Any future publisher bringing a similar suit on the same foundation is working against this precedent directly, in this court.
This isn't the Google antitrust case you might be thinking of
Google is fighting more than one antitrust battle at once, and it's easy to conflate them. They involve different judges, different conduct and different remedies, so treating them as one continuous story leads to wrong conclusions about what's actually been decided.
| Case | Judge | Decided | What it's actually about |
|---|---|---|---|
| US v. Google (search distribution monopoly) | Amit Mehta, D.D.C. | Liability, August 2024 | Exclusive default-search payments to device makers and browsers |
| US v. Google (ad tech) | Leonie Brinkema, E.D. Va. | Liability April 2025; remedies September 2026 | Tying the publisher ad server to the ad exchange — see our ad tech ruling breakdown |
| Penske Media & Chegg v. Google (AI Overviews) | Amit Mehta, D.D.C. | Dismissed, 30 September 2026 | Whether using publisher content in AI Overviews without a new agreement violates antitrust law |
The AI Overviews dismissal and the 2024 search monopoly finding share a judge, which is a coincidence worth flagging precisely because it invites confusion: the same judge who found Google an illegal monopolist in search distribution found, in a separate case on separate facts, that publishers hadn't pleaded an antitrust claim over AI Overviews. One ruling doesn't soften or contradict the other; they're answering different legal questions about different conduct.
What to actually do now that the legal route is closed
Almost no business reading this has standing to bring an antitrust claim against Google, and now neither do Penske Media or Chegg on this theory. So the useful question was never "will a lawsuit fix this." It's what you control regardless of how any case comes out. That comes down to how exposed each piece of your content actually is to being fully answered without a click.
| Content tier | What it looks like | Why it's exposed | What to do about it |
|---|---|---|---|
| Fully answerable | Single-fact definitions, generic how-to steps, FAQ language every competitor states the same way | An AI Overview can complete the whole query without a click | Keep it for topical authority, not for traffic; don't budget click-through revenue against it |
| Partially answerable | Comparison queries, pricing ranges, "which option should I pick" decisions with real trade-offs | AI can summarize the options but can't make the final call for a specific reader's budget or situation | Build the decision framework, table or calculator an answer engine would need to cite rather than restate |
| Not answerable from the public web | Your own data, outcomes specific to a client's situation, a named process you actually run | No public corpus describes it, so there's nothing for an answer engine to summarize | This is where case studies and service pages belong; protect and expand it deliberately |
- Audit your top pages against these three tiers. Most sites discover the majority of their content sits in the first tier, competing for a click an AI Overview already satisfied.
- Move investment toward tier two and three content — the material a search engine's AI Overview and a true AI answer engine both have to cite rather than paraphrase, because paraphrasing it loses the specific number, comparison or outcome that made it useful.
- Decide your crawler policy deliberately rather than by default, since blocking or allowing AI crawlers now also decides your eligibility for whatever compensation mechanisms exist today, including Google's own invitation-only payment pilot for a small set of sites.
- Stop waiting on a legal outcome to set your content strategy. This ruling, like the one before it, confirms that litigation isn't going to be the lever. The lever was always which pages contain something an answer engine can't write without you.
Frequently asked questions
What did the court rule in the Google AI Overviews antitrust case?
On 30 September 2026, Judge Amit Mehta of the U.S. District Court for D.C. dismissed the antitrust lawsuits that Penske Media and Chegg brought against Google over AI Overviews, in a 41-page opinion. He held that an expectation of referral traffic is not a legal agreement, which sank the publishers' central reciprocal-dealing theory, along with their tying and market-definition claims.
Does this mean Google can use my content in AI Overviews without paying me?
It means no U.S. court has found an antitrust violation in how Google uses publisher content in AI Overviews, not that compensation is ruled out everywhere. Copyright and direct-licensing claims run on a separate legal track this ruling doesn't touch, and Google's own invitation-only payment pilot for a small set of sites continues alongside it.
Can publishers still sue Google over AI Overviews?
Penske's dismissal was reported as without prejudice, meaning an amended complaint is possible, and no appeal had been confirmed in reporting as of early October 2026. A separate EU competition complaint about Google's use of publisher content is unaffected by this U.S. ruling and continues on its own track.
Is this the same case as the Google ad tech antitrust ruling?
No. This case is specifically about AI Overviews and was decided on 30 September 2026 by Judge Amit Mehta, the same judge who separately found Google held an illegal search distribution monopoly in August 2024. The ad tech ruling is a distinct case, decided by Judge Leonie Brinkema, concerning Google's publisher ad server and ad exchange business.
What should a business website owner do differently because of this ruling?
Treat the legal route as closed for now and put effort into what you still control: which pages contain information an AI answer can't write without you, whether your highest-intent commercial content survives being summarized, and whether you're blocking or allowing the crawlers that would cite you. None of that depended on the lawsuit's outcome, and all of it matters more now that the lawsuit is over.
The takeaway
A dismissal on these grounds is a narrower event than the headlines suggest: it closes one legal theory, in one court, for claims resting on an implied agreement that traffic would flow a certain way. It doesn't resolve copyright, doesn't bind the EU's separate complaint, and doesn't stop Penske from trying again with a different pleading. What it does do is confirm, again, that the content strategy question was never going to be settled in a courtroom. Spend the time you'd have spent watching the appeal on the content audit instead, or on briefing whoever runs your content production on which tier each page actually sits in.