Leaked internal data obtained through a hack reportedly shows that AI music company Suno trained its music generator by scraping millions of songs and lyrics from platforms including YouTube Music, Deezer, and Genius. The report matters because it offers rare evidence about Suno’s training data practices at a time when the company is already facing lawsuits over alleged copyright infringement and the legality of using protected material under fair use. It may also strengthen claims that Suno went beyond using openly available content by bypassing platform protections to extract audio.

The leaked files, said to include Suno source code from 2023 and 2024, reportedly contained instructions for scraping content from YouTube Music, Deezer, Genius, Pond5, Jamendo, Freesound, IMSLP, and other sources. One file said Suno had taken 2,013,545 YouTube Music clips, while other datasets allegedly covered hundreds of thousands of hours from YouTube Music and thousands more from several other services; code also suggested it aimed to collect about one million hours of podcasts. The hacked material reportedly also exposed customer data such as email addresses, phone numbers, and Stripe payment details, and some users told 404 Media they were never informed; Suno said it became aware of a security incident in November 2025 and contained it quickly.