The generative AI music platform Suno has suffered a major data breach. Leaked internal source code and logs reveal the company systematically scraped over two million audio clips from YouTube Music, totaling approximately 113,879 hours of data, to train its AI models.
The breach, which occurred in November 2025 via a supply-chain attack, exposed harvesting operations across multiple platforms. This includes more than 12,000 hours from Deezer, 17,000 hours from Genius, and over 62,000 hours from stock media site Pond5.
The documents show Suno used commercial proxy networks to mask its scraping activity. This new evidence significantly strengthens the copyright infringement cases already being pursued by major record labels against the company.
Suno recently completed a $400 million funding round, achieving a valuation of $5.4 billion. The company claims the leaked code is outdated and that no customer payment data was compromised.
For the crypto and blockchain sector, this incident highlights the value of projects building decentralized intellectual property and data rights infrastructure. A legal crackdown on unauthorized AI training data could create major market demand for auditable, blockchain-based alternatives.