Chief Judge Yvonne Gonzalez Rogers of the US District Court for the Northern District of California on November 9, 2026 dismissed Google's amended complaint against SerpApi, the Austin company that sells Google search results to developers through an API, finding that none of the four sources of copyright authority Google put forward could carry its anti-circumvention claims. The dismissal is with leave to amend, but only to cure the defects the court listed, and Google has 21 days to file again or see the case dismissed with prejudice.

In Short

A federal judge has thrown out Google's revised lawsuit against SerpApi, which accuses the company of getting around SearchGuard, the system Google launched in January 2025 to stop bots from copying its search results. The ruling matters to the SEO platforms, AI products and data vendors that depend on automated access to Google results, because Google's copyright route against scrapers has now failed twice on the same missing link: facts showing that the owners of copyrighted material in those results authorized SearchGuard to protect it. Google has until the end of November to try a third time, and the order makes clear that a missed deadline ends the case for good.

What the order decided

The 19-page order, entered as Docket No. 62 in Google LLC v. SerpApi, LLC, Case No. 4:25-cv-10826-YGR, grants SerpApi's motion to dismiss in full. It was decided on the papers. According to a footnote in the order, the court found "in its discretion" that the motion was appropriate for resolution without oral argument, so the hearing that had been on the calendar never took place. A hearing set for October 13 had been vacated on October 5, with the judge saying at the time that it would be reset or decided on the papers.

Google brought two claims under Section 1201 of the Digital Millennium Copyright Act. The first, under 17 U.S.C. § 1201(a)(1)(A), targets the act of circumvention itself: the statute says that "no person shall circumvent a technological measure that effectively controls access to a work protected under this title." The second, under § 1201(a)(2), targets anyone who offers a product or service primarily designed to circumvent such a measure. Google says SerpApi's paid Google Search API is that service.

Both claims rest on the same foundation. In its first ruling, on July 20, 2026, the court held that Google had to plead two things about SearchGuard: that it effectively controls access to a work protected by copyright, and that it was implemented with the authority of the copyright owner, as § 1201(a)(3)(B) requires. Google had pleaded neither to the court's satisfaction. That July order dismissed Google's original complaint without leave to amend for search results that contain no copyrighted content, and with leave to amend for results that do.

The November ruling deals only with the second category. And it opens its analysis with a sentence that frames everything after it. Describing the four types of licensed content Google relied on, the court wrote: "Google does not allege that the Google Search results that SerpApi has allegedly scraped by circumventing SearchGuard contained any of these types of content."

How SearchGuard works, according to Google's complaint

The order recites the facts as Google pleaded them, which the court must accept as true at this stage. Google Search results "often" include a Knowledge Panel drawn from Google's Knowledge Graph, a database built from "licensed and open sources." Some of what appears in a panel is copyrighted and licensed from third parties. Results "may" also include copyrighted music lyrics, business images and descriptions, and user-generated content.

According to the amended complaint, SerpApi's business is built on web scraping the output of services such as Google Search and selling it to third parties, without permission from or payment to the copyright holders. Google's terms of service prohibit automated access. Google says the automated queries impose a "deadweight loss" on it, because it earns no revenue from them and so cannot offset the cost of processing them. That framing is worth noting for advertisers: the harm Google pleaded is partly commercial, describing queries that cost money to serve but, by Google's account, produce no revenue.

SearchGuard, launched in January 2025, sends a JavaScript "challenge" to queries from unrecognized sources. A human visitor's browser runs the code and returns information about the browser and the user without any visible interruption. Systems submitting automated queries at large scale "typically cannot solve" the challenge, and are denied the results.

The complaint says SearchGuard initially blocked SerpApi's service but that SerpApi "quickly developed methods to circumvent" it. Those methods, as the order summarizes them, involve masking automated queries to look human, and solving the challenge "with a legitimate request and then syndicating the resulting authorization to other unauthorized browsers that have not solved the challenge." Google alleges harm from the cost of handling SerpApi's "billions of automated queries," from the undermining of its investment in licensed content, from the threat to its licensor relationships, and from the cost of building better defenses. It seeks damages and an injunction. When the case was filed, PPC Land reported that Google was seeking statutory damages of $200 to $2,500 for each act of circumvention.

Four sources of authority, examined one by one

Google's amended complaint, filed on August 10, was an attempt to fill the gap the July order identified. It named four sources from which Google said it drew authority to put SearchGuard in front of licensed content. SerpApi moved on August 24 to dismiss again, arguing that Google had built its case on three contracts it had never produced. The court took each source in turn and tested it against three questions: is the content plausibly copyrighted, did it appear in the search results SerpApi allegedly scraped, and did the copyright owner authorize a measure like SearchGuard?

Reddit

Google's most specific allegation concerned Reddit. According to the amended complaint, Reddit licenses Google to use the copyrighted content of Reddit and its users in "Search Services," and the agreement directs Google "not to enable third parties to extract and independently commercialize the licensed content." Google also pleaded that when Reddit "suspected that scrapers like SerpApi were accessing, taking, and reselling the content that Reddit had licensed to Google, it specifically asked Google to employ technical measures to prevent such unauthorized appropriation."

That was not enough, for three reasons.

First, Google said nothing about what the Reddit content actually is. The bar for copyright protection is low. Quoting the Supreme Court's Feist decision, the order notes that "the requisite level of creativity is extremely low; even a slight amount will suffice." But a complaint still has to allege some facts from which creativity can be inferred, and here "the AC is devoid of any facts indicating that the content subject to Google's licensing agreement with Reddit was creative."

Second, Google did not connect the Reddit content to the results SerpApi is accused of scraping. The court quoted SerpApi's motion, which said the amended complaint "does not identify a single Reddit post that appears in any search result, or any portion of a search result scraped by SerpApi," and added: "The Court agrees." Google never defined "Search Services," so the court could not infer that the phrase covers the Google Search results at issue.

Third, Reddit does not own most of what appears on Reddit. Its users do. Google alleged that "Reddit has the power" to protect both its own works and those its users licensed to it. The court called that allegation "conclusory." Without facts showing how Reddit obtained the right to authorize protection of its users' content, the claim failed for user-owned material.

The two unnamed licensing partners

The second and third sources are licensing agreements whose counterparties Google did not name. PPC Land's analysis in August noted that the amended complaint pleaded confidential contracts as the basis for authorization, a choice that put those deals at the center of the case.

One is described as "a prominent licensing partner that holds copyrights to millions of works that it licenses Google to use in its Search results." That agreement allegedly obligates Google "to use commercially reasonable efforts to safeguard the licensed content against unauthorized third-party access." The court found these allegations "bare and conclusory." They do not identify the parties, do not describe the content, and do not describe the terms closely enough to show that "unauthorized third-party access" covers automated scraping of the kind at issue.

The other is a "major content provider" whose agreement "requires that Google ensure the content [the major content provider] licenses will not be available for download by third parties." This one failed on every count. Google did not allege the content was copyrighted, only that it was licensed. It did not allege the content appeared in the scraped results. It did not describe terms showing that a duty to prevent "downloads" permits measures against scraping. And it did not allege that the provider owned the copyrights in the first place.

Google's Privacy Policy

The fourth source was Google's own Privacy Policy. Google pleaded that the policy covers user information "including the user-uploaded and licensed content they supply, such as photos or reviews," and that Google says in it that it will protect such content from "unauthorized access."

SerpApi filed the policy itself with its motion, and the court treated it as part of the complaint because Google's claims rely on it. Google did not argue otherwise. The text did not match the pleading. "As SerpApi points out, and Google does not dispute, the Google Privacy Policy contains no terms that refer to 'licensed' content or to Google's protection of the same," the order says. The court concluded that the policy "neither grants Google a license to copyrighted content nor contains terms indicating that Google has the authority to protect licensed copyrighted content in any way." The claim also failed because Google did not allege that photos and reviews are creative enough to be copyrighted, or that "Google services," another undefined term, include the search results in dispute.

Lyrics, images and user content

Google's complaint also mentions copyrighted Knowledge Panel images, music lyrics, business images and descriptions, and user-generated content. None of it was tied to any of the four sources of authority, and Google alleged no other facts showing that the owners of that material authorized SearchGuard. Those claims were dismissed as well.

Where Google won the argument

The order is not one-sided on every point, and SerpApi lost several of its arguments along the way.

SerpApi had contended that authorization under § 1201(a)(3)(B) has to come from the written licensing agreement, and that Reddit's request to Google amounted to an "ex-contractual oral authorization" that could not count. The court rejected that. SerpApi cited no authority for the proposition that authorization must come through a written contract, and "in the absence of any authority showing that authorization under 17 U.S.C. § 1201(a)(3)(B) cannot be provided by way of a non-contractual request," the court found Google's allegations about Reddit's request sufficient to infer authorization for Reddit's own content. The Reddit claims failed on the other grounds, not on this one.

SerpApi also argued that Google had to attach the licensing agreements or quote their terms. The court agreed with Google that it did not, noting that the cases SerpApi relied on concern breach of contract, not anti-circumvention claims. What Google must do is describe the terms with enough specificity to make authorization plausible.

A third SerpApi argument, that the "prominent licensing partner" was described only as a copyright "holder" rather than an "owner," also failed. Google cited Ninth Circuit decisions, including UMG Recordings v. Shelter Capital Partners and VMG Salsoul v. Ciccone, that use the two words interchangeably, and SerpApi did not distinguish them.

The court also declined to consider arguments that SerpApi tried to fold in by reference from earlier briefs, because local rules require all arguments supporting a motion to fit within a single 25-page brief. And SerpApi did not get everything it asked for. Its August motion sought dismissal with prejudice. The court instead granted leave to amend once more.

SerpApi's account

SerpApi published its reading of the ruling on November 10, in a blog post by its general counsel, Chad Anson. "On November 9, Chief Judge Yvonne Gonzalez Rogers of the U.S. District Court for the Northern District of California granted SerpApi's motion to dismiss Google's amended complaint," the post says. "It is the second time the court has dismissed Google's claims in this case."

Anson's summary of the four sources tracks the order closely. The post describes Google's amended complaint as pointing to "four supposed sources of authority: a licensing agreement with Reddit, agreements with two unnamed licensing partners, and Google's own Privacy Policy," and lists the court's reasons for rejecting each. It quotes the order's limit on the leave to amend, "but only for the purpose of curing the deficiencies," and notes the 21-day window and the threat of dismissal with prejudice.

The post closes on the company's position. "Whatever Google chooses to do, we view this as another win for our customers and for everyone who depends on an open internet," Anson wrote. "If Google does file again, we are ready to continue the defense of our company, our customers, and our principles."

The post does not mention the points on which the court sided with Google: that a non-contractual request from a licensor can count as authorization, that licensing agreements need not be attached or quoted, and that "holder" and "owner" are interchangeable for these purposes. Those holdings matter for what Google might do next. SerpApi first set out its public response to the lawsuit in January, and in its first motion to dismiss, filed on February 20, it argued that statutory damages calculated per act of circumvention would exceed $7 trillion.

The 21-day clock

The conclusion of the order sets the terms. Google may, "within twenty-one (21) days of the date of this order, file an amended complaint consistent with the rulings herein." By PPC Land's count, that window runs to Monday, November 30, 2026. The permission comes with a condition: Google may amend only "to the extent that it may do so without contradicting the allegations in the AC." It cannot rewrite the facts it has already pleaded.

Google must also comply with the judge's standing order on amended complaints, which requires a redline comparing the new complaint with the previous one. "A failure to file an amended complaint within 21 days will result in the dismissal of this action with prejudice," the order says.

If Google does file, SerpApi's response is due 21 days later, and any new motion may not raise "any new argument that could have been raised previously." Discovery remains stayed until a new complaint is filed and any motion against it is resolved. That stay has held since July. In September, the court also denied SerpApi's motion to compel production of Google's licensing agreements.

What would a third complaint need? The order is close to a checklist. For each source of authority, Google would have to describe the licensed content well enough to show it is copyrightable, allege that the content appeared in the results SerpApi is accused of scraping, and describe the terms of authorization in enough detail to show they reach automated scraping. For the two unnamed agreements, the court specifically faulted the failure to identify the parties. For Reddit's user content, Google would need facts about how Reddit obtained authority from its users. Whether Google is prepared to name licensing partners in a public filing is a commercial question as much as a legal one.

Reddit's New York case reads differently

The order relies heavily on another case in which SerpApi is a defendant. Reddit, Inc. v. SerpApi LLC, decided in 2026 in the Southern District of New York and cited as 840 F. Supp. 3d 308, is the benchmark the court uses again and again to show what Google's complaint lacks. According to the order, the New York court denied a motion to dismiss Reddit's claims under § 1201(a)(1)(A) and § 1201(a)(2), because Reddit alleged that the defendants' circumvention gave them access to content "authored by Reddit and its users."

The contrast is pointed. In New York, Reddit pleaded that the scraped snippets consisted of "creative contributions from users," enough to suggest minimal creativity. It also pleaded that its user agreement grants Reddit a license to user content and "expressly inform[ed] users that Reddit has the authority to protect users' content." Google's complaint, built in part on Reddit's content, contained neither allegation. The same material, in the same kind of claim against the same defendant, is moving forward in one court and has been dismissed in another, because of what each plaintiff wrote down.

Reddit sued SerpApi, Oxylabs, AWMProxy and Perplexity on October 22, 2025, alleging that SerpApi accessed 784 million Google results pages containing Reddit content in the first six days of July 2025. SerpApi filed a renewed motion to dismiss Reddit's amended complaint in March 2026.

The order uses a second recent decision in the same way. Ted Entertainment, Inc. v. ByteDance Inc., decided on October 2, 2026 by a judge in the same district, denied a motion to dismiss a § 1201(a)(1)(A) claim because the plaintiff alleged that circumvention gave access to "volumes of [copyrighted] YouTube videos." Google itself had drawn the court's attention to that decision in a statement of recent decision filed on October 12. The court cited it as a contrast to Google's own pleading.

Why search marketers are watching

SerpApi sits in the data supply chain of search marketing. Its Google Search API returns each SERP as structured data, SerpApi's own blog publishes guides to building a rank tracker on top of it, and when the company released its Extract API in October, its chief executive pointed to AI agents and research workflows as uses. A ruling for Google under the DMCA would have brought statutory damages per act of circumvention to what is otherwise a breach of Google's terms of service, and could have offered other operators of bot-detection systems a template.

For now, that template has not held. Two complaints have failed on the same element: the link between SearchGuard and specific copyrighted works whose owners authorized it. Anti-circumvention claims against automated access are not going away, though. In the lawsuit filed by The New York Times and three other newspapers against Brave on November 2, the publishers included a DMCA claim of their own over a paywall blocker. A newspaper suing over its own articles stands closer to Reddit's position in New York than to Google's in California, because it can plead ownership of the works behind the measure. Google, which displays material licensed from others, has so far not pleaded that chain.

The legal track has also run alongside a technical one. Google dropped the num=100 parameter in September 2025, multiplying the requests needed to collect 100 results. In August 2026 its goto redirect links pushed some rank trackers to 500 to 1,000 requests per query, and SerpApi restored direct URLs on September 5. Since September, vendors have reported Google returning unrelated results to automated collection: Nozzle found swapped results on 8% of pages, and DemandSphere put the rate at 11% of its daily rank checks. None of those measures needs a court's approval.

There is also the matter of consistency. Google built its search business by crawling other people's content, a point PPC Land raised when it examined how Google sues a scraper while being the internet's biggest scraper itself. The November order does not engage with that argument. It turns on narrower questions of pleading, and on whether Google can show that someone with copyright in the scraped results asked for, or agreed to, the lock.

Will Google file a third time? The order leaves the door open for three more weeks. SerpApi says it is ready either way.

Timeline

Summary

Who: Google LLC, the plaintiff; SerpApi, LLC, the Austin-based company that sells scraped search engine results through APIs; Chief Judge Yvonne Gonzalez Rogers of the US District Court for the Northern District of California; and SerpApi general counsel Chad Anson, who published the company's account. Reddit features as one of Google's claimed sources of authority and as the plaintiff in a separate New York case against SerpApi.

What: The court granted SerpApi's motion to dismiss Google's amended complaint under Sections 1201(a)(1)(A) and 1201(a)(2) of the DMCA. None of the four sources of authority Google cited - a Reddit licensing agreement, two agreements with unnamed licensing partners, and Google's Privacy Policy - plausibly showed that copyright owners authorized SearchGuard to protect content in the search results SerpApi allegedly scraped. The court rejected some SerpApi arguments, including that authorization must come from a written contract. Dismissal is with leave to amend, only to cure the identified deficiencies.

When: The order is dated November 9, 2026, and SerpApi published its post on November 10, 2026. Google has 21 days from the order, which by PPC Land's count runs to November 30, 2026, to file a further amended complaint. The case was filed on December 19, 2025, and the original complaint was dismissed on July 20, 2026.

Where: The US District Court for the Northern District of California, in Case No. 4:25-cv-10826-YGR. The order draws on a 2026 decision in Reddit's case against SerpApi in the Southern District of New York.

Why: Section 1201 claims require a technological measure that controls access to copyrighted works with the authority of the copyright owner. Google's search results mix public information with licensed material owned by others, and Google has twice failed to plead facts linking SearchGuard to specific copyrighted works and owners' authorization. The outcome matters to the SEO tools, rank trackers and AI products that rely on automated access to Google results, and to any company considering anti-circumvention claims against scrapers.