Suno Unveils V6 AI Music Models Trained on Licensed Songs

Suno turns text prompts into songs but faces copyright actions from Universal Music Group, Sony Music Group and Germany’s music rights society GEMA. The US litigation, coordinated by the Recording Industry Association of America, centers on alleged use of protected recordings. The stakes extend beyond model training: AI tracks can be uploaded at scale and paired with artificial streams to divert royalty revenue, making reliable provenance a priority for music platforms.

On Aug. 6, Suno co-founder and CEO Mikey Shulman said the company would introduce tamper-resistant audio watermarking and fingerprinting for platform-generated songs. Suno also signed an agreement to use Musixmatch’s Sentinel copyright-detection system, plans to restrict mass downloads and updated its rules against unauthorized voice impersonation. The company raised $400 million in a Series D round in June. A German court ruled for GEMA in late July, with damages yet to be determined.

Suno generates music from text prompts, making the provenance of its training data central to a broader copyright fight over generative AI. The Recording Industry Association of America and major record labels have accused the company of using protected recordings without authorization. If authenticated, the leaked material could bolster those claims and influence how courts assess fair use, while raising separate questions about Suno’s safeguards for the personal data collected from its users.

As of July 22, 2026, reports said a hacker breach exposed information tied to about 55.3 million Suno accounts along with records that appeared to describe its training-data collection. The files allegedly show Suno scraping hundreds of thousands of hours of audio from services including YouTube Music and Deezer, equivalent to decades of continuous listening. The disclosure has turned the incident into a dual crisis involving both alleged copyright infringement and the security of user information.