Suno scraped millions of tracks from YouTube and Deezer for AI training
Leaked data reveals AI music generator Suno trained its models on over two million clips from major streaming platforms, escalating the industry's legal battle over copyright and fair compensation.
AI music generator Suno scraped over two million music clips from platforms including YouTube, Deezer, and Genius to train its generative models. The scale of this data harvesting was exposed after a hacker using the pseudonym Ellie.191 accessed the company’s 2023 and 2024 source code alongside records of hundreds of thousands of users.
The scraped material extends well beyond mainstream music streaming. According to the exposed data, Suno also ingested podcasts and vast catalogs from stock music libraries such as Pond5, Jamendo, and Freesound, as well as the International Music Score Library Project.
Suno has previously acknowledged its aggressive data collection methods. The company stated it trains its models using "all music files of reasonable quality that are accessible on the open internet," a practice it defends as "fair use."
This utilitarian approach was summarized by Suno CEO Mikey Shulman. "I think the majority of people don’t enjoy the majority of time they spend making music," he told 20VC. "It takes a lot of time, it takes a lot of practice, you have to get really good at an instrument or really good at a piece of production software."
Corporate partnerships under fire
The revelation lands amid a fragile restructuring of the music industry's relationship with artificial intelligence. Suno and rival platform Udio recently struck a licensing agreement with Warner Music Group.
However, this corporate detente has sparked its own backlash. This week, WMG asked a US court to dismiss a lawsuit brought by a musicians' union, which accused the label of striking the AI deal "without compensation or credit" for the artists whose work underpins the training data.
Implications for European markets
For Europe, the leaked data presents a direct challenge to established streaming infrastructure and copyright frameworks. Deezer, a major European streaming platform, appears to have been a direct target of Suno’s web scrapers. This raises significant questions about the vulnerability of European digital businesses to automated data extraction.
The economic stakes are high for the continent's creative sectors. If AI models can legally ingest millions of copyrighted tracks to produce competing compositions at near-zero marginal cost, traditional revenue pipelines for European songwriters and performers could face severe disruption. The ongoing US legal battles will likely set a precedent that European regulators and rights holders will be forced to address as they enforce the new AI Act.
