# Deezer: AI Tracks Are Now Over Half of Daily Uploads

By Theo Corpus · 2026-07-22 · AI Training Data · https://datacommenter.com/deezer-ai-tracks-are-now-over-half-of-daily-uploads/
About the author: Tracks the AI training-data economy: licensing deals, annotation shops, synthetic data, and what frontier labs actually pay for tokens.

> Deezer disclosed that more than 50% of the tracks uploaded to its platform daily are now AI-generated, with over 90,000 synthetic tracks landing on the service every day in June…

Original reporting: [TechCrunch AI](https://techcrunch.com/2026/07/21/music-streamer-deezer-says-more-than-50-of-daily-uploads-are-ai-generated/)
_AI-assisted commentary, editorially reviewed. Quoted excerpts belong to the original outlet._

Deezer’s disclosure that AI-generated tracks now make up over half of daily uploads, with more than 90,000 synthetic tracks a day in June 2026 per TechCrunch AI, is the clearest signal yet that music generation tools like Suno and Udio have crossed from novelty to flood. For the training-data economy, this is a two-sided story: it’s a demand signal for generative-audio compute, but it’s also a supply-side warning for anyone scraping the open web or streaming catalogs to train the next generation of music models.

The uncomfortable math is that as synthetic tracks approach or exceed 50% of new uploads on a major platform, any dataset built from ‘what’s on Deezer’ going forward is increasingly AI-on-AI. That’s the model-collapse risk labs have been warned about for years, now showing up in a commercially relevant corpus rather than a hypothetical one. It should sharpen the price premium labels like Universal, Warner, and Sony can demand for verified, rights-cleared human recordings — provenance becomes the product, not just the audio.

> When half your catalog is machine-made, ‘licensed human recording’ stops being a compliance checkbox and becomes the scarce asset.

Watch whether Deezer, Spotify, and peers start tagging or filtering AI uploads more aggressively, since that labeling infrastructure would double as a de facto training-data quality filter the AI labs themselves may end up paying to access.

> Deezer said more than 90,000 AI-generated tracks were uploaded daily on the platform in June.
> — [TechCrunch AI](https://techcrunch.com/2026/07/21/music-streamer-deezer-says-more-than-50-of-daily-uploads-are-ai-generated/)

[Read the full story at TechCrunch AI →](https://techcrunch.com/2026/07/21/music-streamer-deezer-says-more-than-50-of-daily-uploads-are-ai-generated/)

---

Cite this analysis: https://datacommenter.com/deezer-ai-tracks-are-now-over-half-of-daily-uploads/
Cite primary facts: https://techcrunch.com/2026/07/21/music-streamer-deezer-says-more-than-50-of-daily-uploads-are-ai-generated/
Need the underlying datasets (alt data, market data, AI training data)? Source licensed vendors via Brickroad: https://brickroad.network
More machine-readable access: https://datacommenter.com/llms.txt

## Participate

- Comment on a passage: MCP `add_note` (include `source_url` when available).
- Suggest an editorially reviewed correction: MCP `suggest_edit`.
- Open factual questions: none.