Chief Judge Yvonne Gonzalez Rogers ruled on July 20 that Google's anti-bot system doesn't protect copyrighted work the way the DMCA requires, dismissing the core of its case against a small API company that built a business reselling Google's own search results.
Google built itself into the world's most valuable company by crawling and indexing the public web without asking anyone's permission first. It's now trying to use copyright law to stop a 40-person company in Texas from doing something functionally similar to what made Google rich. Last week, a federal judge said no. Frankly, it was hard to see how it was ever going to go any other way.
Judge Yvonne Gonzalez Rogers, Chief Judge of the U.S. District Court for the Northern District of California, dismissed Google's DMCA claims against SerpApi on July 20, 2026. The core ruling is straightforward: the Digital Millennium Copyright Act's anti-circumvention provisions only apply when a technical barrier protects actual copyrighted material. Standard search results, URLs, text snippets, factual index data: these are publicly accessible facts. They're not copyrighted works. You can't use the DMCA as a weapon against someone scraping information that copyright law doesn't protect in the first place. That portion of Google's case was dismissed without leave to amend. It's dead.
Google filed the lawsuit in December 2025 after launching "SearchGuard," an anti-bot system designed to block automated scraping. SerpApi bypassed it. According to Google's complaint, SerpApi's query volume had grown by as much as 25,000% over two years, with the company using IP rotation and browser spoofing to impersonate real users at massive scale. Google estimated hundreds of millions of artificial requests per day. SerpApi scrapes those results and resells them via API, charging developers between $75 and $275 per month depending on volume, at roughly $15 per 1,000 requests.
The judge's ruling doesn't give SerpApi a total clean bill of health. On one narrow set of claims, specifically search results that include licensed images where a third-party copyright holder might have standing, Google was given 21 days to amend and refile. But the central theory, that bypassing an anti-bot system is itself a copyright violation, is gone. As Search Engine Journal reported, the court made clear the DMCA requires an actual copyrighted work behind the gate. A traffic management system protecting factual data doesn't qualify.
SerpApi's argument throughout the case was that Google is "the largest scraper in the world," one that built its entire search index by crawling copyrighted material on third-party sites without explicit permission. That framing landed. When Google called SerpApi's conduct "parasitic" in its original complaint, it was an odd choice of words from a company whose Googlebot has been doing something structurally similar to the open web since 1998. Techdirt put it bluntly in a headline last December: "Google Built Its Empire Scraping The Web. Now It's Suing To Stop Others From Scraping Google."
Why every AI company is paying attention
The ruling matters well beyond the two parties. A huge portion of the AI industry runs on scraped web data, for training, for retrieval-augmented generation, for real-time search products. Startups and infrastructure companies alike have built pipelines that pull from public web sources at scale. Had Google won a broad DMCA theory that bypassing anti-bot systems equals copyright infringement regardless of what's behind the gate, those pipelines would have faced serious legal exposure overnight.
That theory is now off the table, at least in the Northern District of California, which happens to be where most of the relevant companies are incorporated and litigating. The ruling won't stop platforms from deploying access controls, and it won't prevent future lawsuits on different legal theories. But the specific DMCA hook Google tried to use, the one that worked so well against DVD-ripping software in the late 1990s, doesn't extend to factual search result data. Courts have been narrowing that interpretation for years, and this ruling fits the pattern.
For AI companies specifically, the practical impact is real. Services that aggregate search data, build retrieval layers from web content, or train models on publicly accessible information now have clearer footing. The hiQ v. LinkedIn rulings from the Ninth Circuit already established that scraping publicly available data doesn't violate the Computer Fraud and Abuse Act. This ruling adds a parallel boundary on the DMCA side. Neither decision is a blank check. The EU AI Act, entering full enforcement in August 2026, is imposing its own data provenance requirements on AI developers regardless of what U.S. courts say. But in the American legal picture, the tools Google hoped to use against scrapers just got significantly blunter.
Google can refile on the narrow image-licensing theory. Whether it does tells you something about what this lawsuit was actually for. If the goal was genuinely to protect licensed images, Google will amend and push forward. If the goal was to establish a broad legal deterrent against anyone scraping its search results, the judge closed that door. The DMCA is not a general-purpose anti-scraping tool. Google's own history of building on the open web makes it a particularly awkward plaintiff for that argument, and a federal court has now said as much, in writing.
Also read: OpenAI quietly slashed GPT-5.6 Sol's reasoning power by 87% four days after launch • X Money goes live nationwide with 6% APY and a Visa card but Elon Musk still doesn't have a bank • The Pentagon cleared Starlink from America's router blacklist while every other foreign-made device gets shut out
