Sony and UMG logos on a legal document, symbolizing the lawsuit against Suno for AI model laundering.

Editorial illustration for Sony And UMG Sue Suno Over 'Model Laundering' Claims

Sony and UMG Sue Suno Over Model Laundering

By AI Daily Post Edited by Brian Petersen, Editor-in-Chief

September 27, 2026 • 4 min read

Sony Music and Universal Music Group filed a new copyright lawsuit against Suno on Thursday, the second time the labels have taken the AI music startup to court. This one centers on Suno's v6 model, which the labels say still runs on stolen material even though the company claims it was built from scratch. Their argument: v6 was trained partly on outputs generated by Suno's earlier models, and those earlier models were trained on unlicensed recordings pulled from YouTube and other sites.

So the infringement, in the labels' telling, never really went away. It just got a new layer of software between the theft and the product.

Sony and UMG are among the labels that never signed a licensing deal with Suno, unlike some competitors who settled into partnerships with the company. That holdout status is part of why this suit exists. When Suno rolled out v6, an executive named Jack Brody told The Verge the model was trained "from the ground up" using a new dataset that included user data. A company spokesperson, Rachel Racusen, later added detail about licensed partner content and user "creations" feeding the system, without saying whether that user data traced back to the outputs of Suno's earlier, unlicensed models.

Sony and Universal Music Group filed yet another suit against Suno. The labels claim its new v6 model still infringes on their copyrights because it’s trained on user outputs from previous models, which were themselves trained on unlicensed music ripped from YouTube and other sources.

Sony and UMG are suing Suno again - The Verge AI

Why this matters

The "model laundering" framing is the interesting part here, not the fact of another lawsuit. Sony and UMG aren't just arguing Suno's v6 model was trained on stolen recordings, they're arguing that even scrubbing the training set doesn't matter if the new model learned from outputs generated by an older, tainted model. That's a theory that could outlast this particular case.

If courts buy it, "we retrained on synthetic data" stops being a clean defense for any AI company that got caught training on unlicensed material the first time around. For builders in music, audio, or any generative space watching Suno's licensing deals with Warner and others, this suit is a reminder that a signed agreement with some labels doesn't inoculate you against the ones who didn't sign. Sony and UMG staying out as holdouts, then suing anyway, tells you litigation and licensing aren't sequential steps, they're parallel tracks labels are willing to run at once.

Worth watching whether "model laundering" becomes the go-to legal vocabulary for the next wave of AI copyright fights.

Common Questions Answered

What is the 'model laundering' claim that Sony and UMG are making against Suno's v6 model?

Sony and UMG argue that Suno's v6 model was trained on outputs generated by Suno's earlier models, which themselves were trained on unlicensed recordings pulled from YouTube and other sites. The labels contend that even though Suno claims v6 was built from scratch, the model still infringes on their copyrights because it learned from tainted data indirectly through the previous models' outputs. This 'model laundering' theory suggests that retraining on synthetic data doesn't eliminate copyright infringement if the original training data was stolen.

Is this the first lawsuit Sony Music and Universal Music Group have filed against Suno?

No, this is the second lawsuit the labels have filed against the AI music startup. Sony Music and Universal Music Group previously took Suno to court, and this new lawsuit specifically centers on the company's v6 model and the model laundering claims.

What defense strategy could be undermined if courts accept the 'model laundering' theory in this case?

If courts accept the model laundering argument, the common AI industry defense of 'we retrained on synthetic data' would no longer be a clean legal defense for companies that generated that synthetic data from earlier models trained on unlicensed material. This could have significant implications for how AI companies approach model development and training practices going forward.

Where did Sony and UMG claim Suno sourced the unlicensed recordings used to train its earlier models?

According to the lawsuit, Suno sourced unlicensed recordings from YouTube and other sites to train its earlier models. These tainted models then generated outputs that were subsequently used to train the v6 model, creating the chain of copyright infringement that Sony and UMG are challenging.

Ship an AI product this weekend — no engineers required.

Structured, in-depth lessons on the exact no-code tools — not scattered tutorials.

The exact platforms, taught in depth
Build real, working projects
Our honest review + a reader discount

Read the review →