Deezer: Over Half of Daily Music Uploads Are Now AI-Generated
Deezer reported that more than 90,000 AI-generated tracks were uploaded to the platform daily in June, representing over 50% of all daily uploads. The milestone signals a structural shift in how music content is created and distributed at scale.
Original sourceMusic streaming platform Deezer has disclosed that AI-generated tracks now account for more than half of all daily uploads to its platform, with over 90,000 such tracks arriving every day as of June 2026. The company has been among the more vocal streaming services on the issue of AI content, having previously implemented detection systems to identify and manage non-human-generated music.
The scale of the number is significant: 90,000 tracks per day is not a rounding error or an emerging trend — it is the dominant mode of music production on the platform by volume. For context, the entire recorded history of music took decades to accumulate what AI pipelines can now generate in weeks. The sheer throughput raises immediate questions about discovery, royalty distribution, and whether streaming platforms' existing infrastructure was built to handle content at this density.
Deezer has previously advocated for an 'artist-centric' royalty model that would deprioritize low-engagement, high-volume content — a policy framework that takes on new urgency when AI-generated noise tracks and filler content can be uploaded programmatically at scale. The company has pushed for industry-wide standards to distinguish human-created music from AI output, though no cross-platform consensus has emerged.
The disclosure puts pressure on other streaming platforms — Spotify, Apple Music, and Amazon Music — to either publish similar data or explain why they won't. It also sharpens the policy debate around AI content labeling, minimum engagement thresholds for royalty eligibility, and whether platforms have an obligation to gate or throttle automated uploads that threaten to drown out human artists.
Panel Takes
The Skeptic
Reality Check
“The real story here isn't that AI generates music — it's that Deezer's upload pipeline has effectively become a spam intake system, and 'detection tools' are clearly not keeping pace. Every streaming platform almost certainly has similar numbers; Deezer is just the one admitting it. The prediction: Spotify ships an upload gate with mandatory AI disclosure within 12 months, and the platforms that don't will become the dumping ground for every programmatic royalty-farming scheme on the internet.”
The Founder
Business & Market
“The royalty pool math is the actual crisis here — if 50% of uploads are AI-generated and capturing even a small percentage of streams, human artists are getting diluted out of real money by content that cost nothing to make. Deezer's artist-centric model is a direct response to this unit economics problem, but it only works if every major DSP adopts it simultaneously, which is a coordination problem disguised as a policy debate. The platform that solves authenticated human-creator identity first has a genuine moat; the one that doesn't becomes a content landfill with a discovery problem.”
The Creator
Content & Design
“90,000 tracks a day means the discovery surface on Deezer is now majority AI output, and no recommendation algorithm was trained to filter at that ratio — so the editing problem for human musicians just became 'how do I get found in a flood.' The volume isn't the issue; the issue is that tasteless, functional content (sleep noise, study beats, 30-second loops designed to game stream counts) doesn't need to be good to win on engagement metrics. Human artists aren't competing with AI creativity; they're competing with AI quantity, and that's a fight designed for the platform to lose.”
The Futurist
Big Picture
“The falsifiable thesis here is: within 3 years, streaming platforms will split into two tiers — authenticated human-creator platforms with premium royalty pools, and open-upload platforms where AI and human content coexist at commodity rates. Deezer's disclosure is the canary, but the structural dependency is whether rights holders and labels force DSPs to implement provenance standards before the royalty pool is too diluted to matter. The second-order effect nobody is talking about: if AI content reaches 80% of uploads, the advertising and licensing value of 'music' as a category collapses, because brands can just generate their own.”