Suno Training Data and Leak Details Exposed
404 Media · rss · 2026-07-15
404 Media reports that Suno suffered a data breach, leaking its training library and internal data. The leaked materials reveal that Suno's music generation system scraped massive amounts of content from YouTube Music, Deezer, Genius, as well as Pond5, Jamendo, Freesound, IMSLP, and podcast RSS feeds. This includes scripts and instructions allegedly used to scrape songs directly from YouTube.
The article notes that the leaked code contains multiple dataset directories and scale metrics, such as youtubemusic, geniushq, deezer, and pond5music, totaling the equivalent of "decades of music." Hackers reportedly compromised the company via supply chain attacks and employee accounts, briefly accessing customer info and Stripe-related payment data. Suno responded that the incident occurred in November 2025 and was quickly contained, mainly affecting outdated source code without exposing sensitive personal information.
The report also brings the issue of training data sources for generative music products back into the spotlight. Suno previously admitted in lawsuits that its training data covers "almost all high-quality music files available on the internet." While Suno attempts to mitigate the risk of generated outputs resembling existing songs, controversies regarding copyright, scraping, and data compliance persist.
More from Multimodal
- LanSE: using sparse encoders to evaluate multimodal model outputs — xuanalogue · 2026-10-26
- Kling 4.0 demo: Tesla Roadster takes an impossible route around the world — SimplyAnnisa · 2026-10-26
- 4K AI music video "TENDER" made with Kling, DomoAI and Seedance — Superb_Vegetable_684 · 2026-10-26
- Moss character LoRA training finally clicks in Fizgig, beating Krea results for a few dollars — derTommygun · 2026-10-26
- Claude Code rebuilt its own ComfyUI pipeline to fix prop continuity in a 2-min AI film — Odd_Definition_9724 · 2026-10-26
- Gingerbread watercolor storybook clipart prompt shared with polished sample outputs — azed_ai · 2026-10-26