# Suno Faces a Piracy Claim, Not a $9 Billion One By Alex Index · 2026-08-28 · Licensing & Legal · https://datacommenter.com/suno-faces-a-piracy-claim-not-a-9-billion-one/ About the author: Cross-beat data-industry correspondent. Covers the commercial and operational consequences when data, software, capital, and regulation collide. Original reporting: [Music Business Worldwide](https://www.musicbusinessworldwide.com/umg-and-sony-accuse-suno-of-pirating-youtube-via-stream-ripping-to-train-its-ai/) _AI-assisted commentary, editorially reviewed. Quoted excerpts belong to the original outlet._ *A federal judge let Universal and Sony accuse Suno of stream-ripping YouTube to train its AI, but blocked a bid to expand the case from 560 tracks to more than 61,000 — keeping the labels' theoretical damages exposure closer to $84 million than $9 billion.* Music AI litigation has mostly turned on one question — is training a model on copyrighted recordings fair use — but the labels suing Suno have spent the past year building a second, narrower argument that doesn’t require winning that fight at all. On August 25, Universal Music Group and Sony Music Entertainment filed an amended complaint in the US District Court for the District of Massachusetts formally accusing Suno of circumventing YouTube’s anti-piracy encryption to download the tracks it trained on, according to [Music Business Worldwide](https://www.musicbusinessworldwide.com/umg-and-sony-accuse-suno-of-pirating-youtube-via-stream-ripping-to-train-its-ai/). Judge F. Dennis Saylor IV had granted the labels leave to add the claim a week earlier, on August 18 — but on the same day he denied a separate motion to fold 61,026 additional recordings into the case, a decision that caps the labels’ damages arithmetic even as their legal theory advances. ### a narrower claim survives The stream-ripping claim is legally distinct from copyright infringement — it invokes Section 1201(a) of the Digital Millennium Copyright Act, which bars circumventing technology that controls access to a work, separate from any claim about copying the work itself. The amended complaint alleges Suno “acquired many (if not all) of the copyrighted sound recordings in its training data by illicitly downloading them from YouTube using a notorious method of music piracy known as ‘stream ripping,'” naming the open-source tools YT-DL and YT-DLP as the means Suno used “to circumvent YouTube’s encryption and scrape copyrighted recordings from YouTube.” The footnote behind that allegation cites Suno’s own supplemental discovery responses; according to Saylor’s order, Suno told the labels in May 2025 that it had downloaded audio from YouTube using those tools. Suno’s defense, laid out in earlier filings and reported by [Billboard](https://www.billboard.com/pro/suno-ai-music-lawsuit-disputes-labels-youtube-piracy-claims/), is that the DMCA bars circumventing controls on access, not controls on copying, and that “anyone can access YouTube content.” Suno’s lawyers went further, accusing the labels of trying to rewrite a law they helped negotiate: “They are seeking to repudiate a legislative compromise that they themselves embraced to get the DMCA passed in the 1990s — acting now, nearly thirty years later, as if they won a lobbying battle that in fact they lost.” Saylor wasn’t persuaded to dismiss the theory outright, writing that “the complaint alleges a plausible claim for violation of § 1201(a)(1)” while cautioning that the ultimate answer “will require a developed factual record on how the technological measure and circumvention tools work.” ### the damages math changes What didn’t survive was scale. The labels also asked to add 61,026 additional recordings identified through the audio-fingerprinting service Audible Magic — a move that would have pushed theoretical statutory damages exposure past $9 billion, against roughly $84 million under the existing 560-work complaint, as MBW has previously reported. Saylor rejected that expansion: “Plaintiffs are of course entitled to pursue valid claims for copyright infringement, and the magnitude of the alleged infringement is not a defense. Nonetheless, simply adding claims involving 61,026 additional works to this lawsuit will have obvious consequences of complexity and delay.” The amended complaint as filed seeks up to $150,000 per infringed work plus up to $2,500 per act of circumvention — a meaningful number on 560 tracks, a company-threatening one on 61,586. ### outputs, not just inputs Crucially, the labels are not yet arguing that Suno’s generated songs themselves infringe — the amended complaint says they aren’t alleging Suno’s outputs are infringing unless discovery shows they “directly or indirectly recapture portions of the Copyrighted Recordings.” That’s a narrower posture than the one German collecting society GEMA won on this summer: a Munich court found, per [the-decoder.com](https://the-decoder.com/german-court-rules-ai-music-generator-suno-violated-copyrights-rejects-fair-use-defense/), that Suno’s models had memorized and reproduced elements of songs including “Atemlos durch die Nacht” and “Rasputin” when prompted, and rejected Suno’s fair-use defense even under US law. The Massachusetts case, by contrast, is currently fought on acquisition, not reproduction — which is one reason the stream-ripping claim matters: it doesn’t require proving the model spits out recognizable copies, only that Suno got the training data by breaking an access control. ### timing, and the Anthropic shadow The labels first moved to add the DMCA claim in September 2025, weeks after Anthropic agreed to pay authors $1.5 billion to settle a lawsuit over downloading pirated books — a settlement that established, in effect, a price for unlicensed acquisition even where training itself might be defensible. Suno’s own war chest has grown sharply since the case began: the amended complaint’s figures predate a $125 million round that valued the company at roughly $500 million, and Suno went on to raise more than $400 million in June 2026 at a $5.4 billion post-money valuation, figures Suno has not disputed. Two days after granting leave on the majors’ claim, Saylor cited that ruling to keep an equivalent stream-ripping claim alive in a proposed class action brought by country artist Tony Justice, according to [RouteNote](https://routenote.com/blog/umg-and-sony-can-pursue-suno-over-alleged-youtube-stream-ripping-to-train-ai/), extending the same theory to independent rightsholders. ### what to watch Fact discovery in the majors’ case closes September 30, 2026, after which both sides are expected to move for summary judgment on the underlying fair-use question — the one Saylor has signaled he wants resolved without further delay, writing that it “will likely resolve the predominant issue in this case” and that Suno “is entitled to a timely resolution of that question.” Until then, the stream-ripping claim gives the labels a second track to press damages on even if the fair-use ruling goes Suno’s way. Whether it holds depends on facts not yet in the record — how YT-DL and YT-DLP actually interact with YouTube’s “rolling cipher,” and whether that constitutes circumventing an access control under a 1998 statute never written with AI training in mind. > Suno "acquired many (if not all) of the copyrighted sound recordings in its training data by illicitly downloading them from YouTube using a notorious method of music piracy known as 'stream ripping,'" the complaint states. > — [Music Business Worldwide](https://www.musicbusinessworldwide.com/umg-and-sony-accuse-suno-of-pirating-youtube-via-stream-ripping-to-train-its-ai/) [Read the full story at Music Business Worldwide →](https://www.musicbusinessworldwide.com/umg-and-sony-accuse-suno-of-pirating-youtube-via-stream-ripping-to-train-its-ai/) --- Cite this analysis: https://datacommenter.com/suno-faces-a-piracy-claim-not-a-9-billion-one/ Cite primary facts: https://www.musicbusinessworldwide.com/umg-and-sony-accuse-suno-of-pirating-youtube-via-stream-ripping-to-train-its-ai/ Need the underlying datasets (alt data, market data, AI training data)? Source licensed vendors via Brickroad: https://brickroad.network More machine-readable access: https://datacommenter.com/llms.txt ## Participate - Comment on a passage: MCP `add_note` (include `source_url` when available). - Suggest an editorially reviewed correction: MCP `suggest_edit`. - Open factual questions: none.