Sony Music and UMG Launch Second Lawsuit Against AI Music Generator Suno Over 60,202 Songs

Sony Music and UMG File Second Copyright Infringement Lawsuit Against Suno

Sony Music Entertainment and Universal Music Group have escalated their legal battle against AI music generator Suno by filing a second copyright infringement lawsuit in the U.S. District Court in the District of Massachusetts1

1

3

3

. The new complaint asserts 60,202 sound recordings that were not part of the original 2024 case, which is still ongoing3

3

. The labels initially attempted to amend their original complaint to include these additional works, but a judge denied the motion in August and directed them toward parallel proceedings instead3

3

. This Suno lawsuit marks a critical moment in the legal battle between AI and music industry stakeholders over intellectual property rights.

AI Model Trained on Unauthorized Recordings Despite Licensing Deals

The core of Sony Music and Universal Music Group's complaint centers on Suno's new v6 model, which the company launched in September 2025 with licensing deals with music labels including Warner Music Group, BMG, and Believe1

1

4

4

. While Suno claims this AI music generator was trained exclusively on licensed material, UMG and Sony Music Entertainment argue otherwise. The labels specifically take issue with Suno's public acknowledgment that the models were trained on "users' interactions" with Suno's service, which they interpret as "the outputs of, and preference signals derived from, prior models that Suno built by copying Plaintiffs' recordings without authorization"1

1

4

4

. The complaint alleges that Suno developed v6 using knowledge distillation and related model-transfer techniques, whereby a new "student" model is trained to reproduce the learned behavior of a predecessor "teacher" model built on unlicensed material4

4

.

Training a New Model on Infringing Outputs Doesn't Eliminate Copyright Infringement

"Training a 'new' model on the outputs of an infringing model does not eliminate the infringement; it launders it," the labels wrote in their complaint2

2

. "V6 is not a fresh start; it is the fruit of the same poisoned tree." This argument strikes at the heart of AI innovation and copyright law debates. The labels commissioned a forensic analysis after Suno refused to name the recordings used in training, and that analysis reportedly found millions of their works in the training corpus3

3

. A July 2026 hack of Suno's data revealed the company scraped millions of songs and lyrics from platforms like YouTube Music, Deezer, and Genius to train its earlier models1

1

. The breach exposed more than two million music clips scraped from YouTube Music, more than 17,000 hours worth of songs from Genius, and thousands of hours of material from other platforms2

2

.

Potential Damages Could Reach $9 Billion for Unlicensed Material

Under US copyright law, Suno could be liable for up to $9 billion in damages for the alleged infringement of 60,202 sound recordings1

1

. The complaint also seeks up to $2,500 for each instance Suno circumvented YouTube's anti-downloading technology to scrape songs through stream ripping1

1

3

3

. The lawsuit includes three counts: two for direct infringement split between post-1972 recordings and pre-1972 ones protected under the Music Modernization Act, and a third for circumvention3

3

. The labels argue the infringement is "knowing and willful" because Suno's own terms of service since January 2024 have required every user to grant the company a "worldwide, non-exclusive, fully paid-up, sublicensable" and "irrevocable" license over anything they upload3

3

. This demonstrates Suno understood that training on someone's audio requires permission, yet the company took commercial recordings for free while demanding licenses from its users.

Licensing Deals With Music Labels Prove a Functioning Market Exists

The three licensing deals Suno signed over the past year have become the labels' sharpest weapon in this copyright infringement case. Warner Music Group signed in November 2025, BMG followed on August 12, and Believe signed on September 8 to supply training data3

3

. Nine days after the Believe deal, Suno launched v6 on licensed catalogue3

3

. UMG and Sony Music Entertainment argue these three agreements in under a year "confirm that a functioning market exists for licensing sound recordings to train generative AI models"3

3

. Notably, Warner Music Group was originally a plaintiff in the first lawsuit but has since settled and partnered with Suno3

3

4

4

. Suno's chief product officer Jack Brody has tried to reframe the deals, saying the revenue share is "not in exchange for training" and that the deals are "not really about the data"3

3

. However, the complaint quotes Suno CEO Mikey Shulman saying the company "will be licensing works," calling the deals "a functioning licensing market, whatever nomenclature Suno adopts for litigation purposes"3

3

.

Suno Defends Its Position and Removed Artist-Name Blocking Feature

Suno firmly rejects the allegations, calling the claims "fundamentally flawed on both the facts and the law"1

1

2

2

. The company maintains that "Suno exists so that more people can make new music" and that v6 was "trained on content licensed from our partners, interactions including creations and preference signals from our community, and the accumulated learnings from our team"1

1

2

2

. A Suno spokesperson explained that when someone requests a particular artist, the system translates that into musical qualities behind the request "not so that a user can imitate someone else's work, but so that people learn the language of music and make something of their own"2

2

. Interestingly, Suno once blocked users from prompting it with the names of specific recording artists, with Shulman citing that policy as proof the models generated "completely new outputs"3

3

. The complaint notes this design feature has been switched off, and users can now prompt with artist names through prompt augmentation3

3

.

What This Means for AI Innovation and the Music Industry

This second Suno lawsuit highlights the ongoing tension between AI innovation and copyright law in the music sector. UMG and Sony argue that Suno is contributing to market over-saturation on streaming services with waves of AI-generated tracks, which "directly competes with, cheapens, and ultimately drowns out the genuine sound recordings on which the service is built"4

4

. The complaint also quotes an early Suno investor who told Rolling Stone: "If we had deals with labels when this company got started, I probably wouldn't have invested in it." The investor added that Suno "needed to make this product without the constraints"3

3

. As the case proceeds, watch for how courts interpret whether training a new AI model on outputs from an infringing model constitutes laundering copyright violations. The outcome will set precedents for how generative AI models can legally use copyrighted material and whether the fair use doctrine extends to AI training. Alongside the release of its v6 models, Suno ended support for its older models, suggesting the company believes offering a product trained on licensed material is more defensible1

1

. If a judge accepts the record companies' understanding of how Suno trained its models, that strategy might not be enough to shield the AI music generator from liability.