Every rank tracker, visibility platform, and SERP dataset in this industry rests on the same foundation: somebody is collecting search results at scale. That has always carried a legal question mark, and in December 2025 Google put a case behind the question mark. In July a judge threw most of it out.
We are not lawyers and nothing here is legal advice. But the reasoning in this ruling is unusually legible for a technical audience, and it changes the risk picture for a category of tool that a lot of marketing budgets quietly depend on. Worth understanding properly rather than through a headline.
What the Google SerpApi ruling actually held
The case ran in the Northern District of California before Chief Judge Yvonne Gonzalez Rogers. Google's claims were built on DMCA Section 1201, the anti-circumvention provision, arguing that SerpApi got around SearchGuard, Google's technical measure for restricting automated access to search results.
The court split the claims and treated them very differently, which is the part most summaries skip.
| CLAIM | OUTCOME | REASON GIVEN |
|---|---|---|
| Search results containing no copyrighted content | Dismissed with prejudice | Section 1201 protects access controls guarding a copyrighted work. No work, nothing to protect. Decided as a matter of law. |
| Licensed images in Knowledge Panels | Dismissed with leave to amend | Section 1201(a)(3)(B) requires the measure to operate with the copyright owner's authority. Google does not own third-party licensed images and did not plead that those owners authorized it. |
With prejudice is the phrase to notice. That theory cannot be refiled. The court did not say Google failed to plead it well enough; it said that where search results contain no copyrighted content, there is no protected work for an access control to guard, and that is a legal conclusion rather than a factual gap.
The second bucket is alive. Google was given 21 days to amend and plead facts showing the copyright owners of those licensed images authorized Google to deploy SearchGuard on their behalf. That is a narrower, more technical claim than the one it started with, and it turns on a licensing relationship rather than on scraping itself.
It is worth pausing on how narrow the surviving claim really is compared to what Google filed. The original complaint went after the act of collecting search results. What is left is a question about whether Google had authority from specific third-party rights holders to enforce a technical measure on their behalf over specific licensed images appearing in Knowledge Panels. Those are different cases wearing the same caption. If you were worried in December that a Google win would make SERP data collection categorically unlawful, the version of the case still standing does not do that even if Google prevails on it, and the docket itself is the better guide here than the commentary around it.
Why the reasoning matters more than the outcome
The line worth reading twice is the court's characterization of what SearchGuard does. The judge observed that it functions to regulate automated access rather than to protect copyrighted works, operating as a general-purpose gate over a mixed bag of content rather than protecting any specific work.
That framing is what gives the ruling reach beyond this one defendant. Search results are largely factual output: URLs, titles, snippets drawn from other people's pages, positions computed by a ranking system. Facts are not copyrightable. A blanket technical control sitting over factual output does not become a copyright measure because the operator would prefer it were one.
“The court did not decide that scraping search results is fine. It decided that copyright anti-circumvention law is the wrong instrument for objecting to it.”
Take that as the actual holding. It narrows one specific legal theory, forcefully, on grounds that generalize. It does not bless data collection in the round, and reading it as a green light would be a mistake that costs someone a lot of money eventually.
There is a second-order reason this framing should interest SEO teams beyond the tooling question. The same argument gets made in the other direction about your own site. Publishers deploying bot walls over pages that are mostly factual product data, pricing, or specifications are in structurally the same position Google was: they have a legitimate interest in controlling automated access, and a weaker claim that the thing being controlled is a copyrighted work. That does not make blocking wrong or unwise. It does mean the legal foundation under aggressive anti-bot posture is thinner than it is often assumed to be, and any strategy built on the assumption that copyright law will back up your access controls deserves a second look.
What this does not change
Three things survive completely intact, and any team building on SERP data should hold them clearly.
That third point deserves the most weight operationally, because it is the one that actually shows up in your tooling. The realistic threat to your rank data has never mainly been litigation. It is access hardening: heavier bot detection, more aggressive challenges, more surfaces that render in ways collectors cannot cheaply parse. That pressure is orthogonal to the lawsuit and has been building regardless, in the same way access control decisions elsewhere have caused abrupt visibility incidents with no legal process involved at all.
The fourth point is the one most likely to be misreported over the next few months, so it is worth stating flatly. A motion to dismiss tests whether the claims as written can proceed, not whether the underlying conduct was acceptable. No evidence was weighed. No finding was made about what SerpApi did or how it did it. Coverage that describes this as a court declaring scraping legal is describing something that did not happen. What happened is that one plaintiff picked a statute the court found inapplicable to most of what it was complaining about, which is a real and consequential outcome, and also a much narrower one.
None of which means the ruling is unimportant. Anti-circumvention claims are attractive to plaintiffs precisely because they route around the difficult questions, whether the underlying material is copyrightable at all, and instead attach to the act of getting past a technical gate. A court holding that the gate must actually guard a copyrighted work to qualify removes a shortcut. That matters for every data provider in this industry, and it matters most for the small ones, who settle not because they are wrong but because defending a federal case is more expensive than folding.
The practical read for data-dependent teams
If your reporting, client deliverables, or product depend on third-party SERP data, this ruling reduces one tail risk and leaves the operational ones where they were. Handle it accordingly.
The first-party point is the durable one and it long predates this case. We made a version of this argument when mapping the four datasets that make up real AI visibility measurement, and the ruling does not change the conclusion: data you generate is data nobody can take away in a term sheet, a policy update, or a court filing. Third-party SERP data is genuinely valuable and worth paying for. It should not be load-bearing.
There is a broader pattern here that goes past rank tracking. The same question, who may collect what from whom, is being litigated and negotiated across crawling, AI training, and retrieval simultaneously. The distinction the court drew between a general-purpose bot gate and a copyright protection measure is the same distinction that runs underneath separating training crawlers from retrieval crawlers. Access control and content rights keep getting conflated because the same infrastructure enforces both, and courts are starting to pull them apart.
What to watch next
The near-term question is whether Google amends the Knowledge Panel image claim and whether that narrower theory holds up. It is a much smaller case than the one filed in December, and it would turn on licensing authority rather than on scraping. Even a Google win there would not revive the theory dismissed with prejudice. The likelier long-run response is not a better lawsuit anyway. It is product and infrastructure: more of the result page rendered in ways that resist cheap collection, more challenges in front of automated access, and more of the interesting data living inside interfaces that were never designed to be parsed in the first place.
The more consequential thing to watch is not legal at all. Watch coverage quality in your tools. If your provider's data starts thinning on particular result types, or refresh intervals stretch, that is access hardening showing up in your dashboard, and it will affect your reporting long before any appellate decision does. That signal lives in your own quality checks, not in a docket.
Do this next: list every metric in your standard client or executive report and mark each one first-party or third-party. Most teams find the split more lopsided than expected, usually toward data they do not control and cannot reproduce. Rebalancing that is ordinary analytics hygiene, and it is the version of this story you can actually act on. For teams in regulated categories where data provenance gets asked about directly, that inventory is worth having on paper before someone requests it.
See where you are cited today
A free snapshot audit of your rankings and AI citations before we ever talk.
Tyler leads work at the intersection of SEO and generative engines at Something Inc., helping B2B brands get ranked and cited across every major AI engine.