A California federal court gave Bright Data summary judgment against Meta in January 2024, but that ruling decided a contract dispute over logged-off public data. The court found no evidence that Bright Data scraped Facebook or Instagram while logged into an account or reached material hidden behind a privacy wall. It then held Meta's terms did not bar the logged-off scraping at issue.
That victory matters, but not in the broad way its headline suggests. Meta did not ask the court to decide whether copying entire copyrighted songs for AI training infringes copyright, and the order did not grant scrapers blanket immunity whenever material can be reached without a login. A public-data scraping victory is therefore not the same thing as a copyright license.
The X case is narrower in a different way. In May 2024, Judge William Alsup dismissed X's contract and tort theories based on scraping and selling public data, reasoning that X could not use state law to create copyright-like control over material it did not own. The opinion left federal copyright law to govern whether copying protected expression is lawful, including whether fair use applies.
Those are allegations, not findings. But they target a category of material unlike public profile fields, usernames, likes, prices, or other facts that dominated earlier scraping disputes. Songs, sound recordings, and lyrics can embody protected expression, so public accessibility does not by itself erase the copyright owner's exclusive reproduction rights.
The X opinion actually cuts against using X as a universal shield. Its logic was that X could not privately replace the Copyright Act with sweeping terms that let the platform control public user content. That leaves an actual copyright owner free to invoke the federal rights that X itself lacked.
That standard helps Bright Data in one obvious respect. Its proxy networks, browser automation, scraping tools, and data services have substantial lawful uses across research, price monitoring, search, e-commerce, and other public-data tasks. Under Cox, knowledge that a customer may misuse a broadly useful service is not enough by itself.
Gerencia's complaint appears drafted with that hurdle in mind. It does not stop at saying Bright Data knew Suno was scraping music. It alleges Bright Data offered functionality for extracting separated audio tracks for audio-language and multimodal training, along with products that solve CAPTCHAs, rotate proxies, and defeat anti-bot measures.
The hard issue is how narrowly a court defines the relevant service. Bright Data can point to a general platform with many legitimate uses, while Gerencia will want the court to focus on the specific tools, configurations, customer work, and marketing allegedly used for mass music acquisition. Cox makes that framing battle central rather than incidental.
The same features become more legally sensitive if evidence shows they were specifically sold, configured, or promoted to obtain protected works without authorization. Cox distinguishes a service that happens to facilitate infringement from one that is affirmatively promoted for infringement or effectively built around it. Gerencia therefore needs evidence of purpose, not just capability.
That is why the old Meta and X wins may still be useful without deciding this case. They give Bright Data strong language around logged-off public access, contractual overreach, and the legitimacy of general-purpose scraping infrastructure. They do not answer whether a copyright owner can prove that a particular scraping service was intentionally adapted or encouraged for unauthorized copying of protected music.
The result could turn on facts that the 2024 cases barely had to confront. Customer communications, product configurations, training-data workflows, marketing aimed at AI music, and the actual balance of lawful versus infringing uses could matter far more than the simple label "public web data." A court applying Cox will likely care less about whether scraping sounds familiar and more about what Bright Data intended the relevant service to do.
That victory matters, but not in the broad way its headline suggests. Meta did not ask the court to decide whether copying entire copyrighted songs for AI training infringes copyright, and the order did not grant scrapers blanket immunity whenever material can be reached without a login. A public-data scraping victory is therefore not the same thing as a copyright license.
The X case is narrower in a different way. In May 2024, Judge William Alsup dismissed X's contract and tort theories based on scraping and selling public data, reasoning that X could not use state law to create copyright-like control over material it did not own. The opinion left federal copyright law to govern whether copying protected expression is lawful, including whether fair use applies.
Meta and X did not bless every form of scraping
That distinction now sits underneath Gerencia 360’s claims against Suno and Bright Data, filed in the Northern District of California on August 31, 2026. The complaint alleges Bright Data supplied proxy and scraping services that Suno used to obtain songs and lyrics from services including YouTube Music, Deezer, and Genius. It also alleges Bright Data's tools were used to bypass anti-bot controls and prepare audio for AI training.Those are allegations, not findings. But they target a category of material unlike public profile fields, usernames, likes, prices, or other facts that dominated earlier scraping disputes. Songs, sound recordings, and lyrics can embody protected expression, so public accessibility does not by itself erase the copyright owner's exclusive reproduction rights.
The X opinion actually cuts against using X as a universal shield. Its logic was that X could not privately replace the Copyright Act with sweeping terms that let the platform control public user content. That leaves an actual copyright owner free to invoke the federal rights that X itself lacked.
Cox makes intent the center of the Bright Data fight
A newer Supreme Court ruling changes the contributory-infringement analysis too. In March 2026, Cox Communications v. Sony Music held that a service provider is not liable merely because it knows some customers will infringe and continues supplying a general-purpose service. A plaintiff must instead show inducement or that the service was tailored to infringement.That standard helps Bright Data in one obvious respect. Its proxy networks, browser automation, scraping tools, and data services have substantial lawful uses across research, price monitoring, search, e-commerce, and other public-data tasks. Under Cox, knowledge that a customer may misuse a broadly useful service is not enough by itself.
Gerencia's complaint appears drafted with that hurdle in mind. It does not stop at saying Bright Data knew Suno was scraping music. It alleges Bright Data offered functionality for extracting separated audio tracks for audio-language and multimodal training, along with products that solve CAPTCHAs, rotate proxies, and defeat anti-bot measures.
The hard issue is how narrowly a court defines the relevant service. Bright Data can point to a general platform with many legitimate uses, while Gerencia will want the court to focus on the specific tools, configurations, customer work, and marketing allegedly used for mass music acquisition. Cox makes that framing battle central rather than incidental.
Bright Data’s lawful uses may still matter
Current Bright Data materials strengthen both sides of that argument. The company markets Web Unlocker as a system that handles proxy rotation, CAPTCHA solving, fingerprinting, and other blocking techniques, while its video-data offering advertises synchronized video, audio, captions, and separated audio tracks for model training. Those capabilities can support lawful collection, licensed training, research, and many other uses.The same features become more legally sensitive if evidence shows they were specifically sold, configured, or promoted to obtain protected works without authorization. Cox distinguishes a service that happens to facilitate infringement from one that is affirmatively promoted for infringement or effectively built around it. Gerencia therefore needs evidence of purpose, not just capability.
That is why the old Meta and X wins may still be useful without deciding this case. They give Bright Data strong language around logged-off public access, contractual overreach, and the legitimacy of general-purpose scraping infrastructure. They do not answer whether a copyright owner can prove that a particular scraping service was intentionally adapted or encouraged for unauthorized copying of protected music.
The result could turn on facts that the 2024 cases barely had to confront. Customer communications, product configurations, training-data workflows, marketing aimed at AI music, and the actual balance of lawful versus infringing uses could matter far more than the simple label "public web data." A court applying Cox will likely care less about whether scraping sounds familiar and more about what Bright Data intended the relevant service to do.
