When more than half of daily music uploads are AI-generated, the distributionplatform stops being a marketplace and starts being an infrastructure problem.
Deezer CEO Jeronimo Folha told TechCrunch this week that more than 50% of the tracks uploaded to Deezer daily are now AI-generated — a threshold the company confirmed publicly for the first time. That’s not a rounding error or a marginal trend; it represents a categorical shift in what a music streaming platform actually receives from the supply side of its marketplace.
The disclosure arrives in the context of an ongoing industry-wide fight over royalties and platform gatekeeping. Deezer has been among the most vocal DSPs (digital service providers) pushing for what it calls an “artist-centric” royalty model — a system that deprioritizes streams from functional and AI-generated content in favor of tracks with demonstrated human listener intent. The AI upload surge makes that policy choice no longer optional; it’s existential.
The mechanics are straightforward: generative audio tools like Suno, Udio, and a growing cohort of no-name API wrappers can produce thousands of tracks per day at near-zero marginal cost. When those tracks get uploaded to a DSP, they compete for algorithmic placement, playlist slots, and — critically — royalty pool share. Even micro-streams, when multiplied across millions of AI tracks, drain the pool that human artists depend on.
The key insight: Deezer’s 50% figure isn’t a content moderation problem — it’s a market structure problem. When the marginal cost of supply collapses to zero, every platform built on a pro-rata royalty pool becomes a target for systematic dilution. The platform’s incentive architecture was designed for scarcity; AI has abolished it.
HOW THE SUPPLY WALL COLLAPSED
2021–2022
Streaming platforms report ~60,000 new tracks/day. Upload velocity treated as a vanity metric, not a risk variable.
Early 2024
Universal Music Group pulls catalog from TikTok; DSPs begin acknowledging AI-generated “noise tracks” gaming royalty pools. Spotify quietly removes millions of Boomy AI tracks.
Late 2024
Suno and Udio face RIAA lawsuits. Deezer announces “artist-centric” royalty model; sets minimum stream thresholds to qualify for payout. Upload volumes continue accelerating.
July 2026
Deezer CEO confirms >50% of daily uploads are AI-generated — the first major DSP to put a majority figure on the record publicly.
The Structural Read
The framing most outlets will use is “content moderation.” That’s the wrong frame. Content moderation is about removing bad actors at the margin. What Deezer is describing is a supply-side inversion — where the majority of inputs to the platform now come from a category of producer with zero marginal cost, zero reputational skin in the game, and a rational incentive to maximize volume rather than quality.
This is a textbook case for the Map of AI framework. Deezer sits at the Distribution layer of the AI value stack — it doesn’t train models, it doesn’t build tools, it surfaces content to listeners. For years, that position was defensible because supply was constrained by human creative output. AI generators have effectively commoditized the input layer that Deezer’s curation and discovery engine depended on for differentiation.
The deeper structural problem: pro-rata royalty pools are a fixed-pie mechanism. Every AI-generated track that earns even a fraction of a cent dilutes the per-stream value of every human-made track. The platform didn’t change its payout rules — the denominator changed underneath it. And unlike content moderation, you cannot ban a category of tool without triggering the next generation of tools designed to evade detection.
Map of AI — Distribution Layer Thesis
“When the cost of generating a supply-side input collapses to zero, distribution platforms face a structural tax: their curation and discovery infrastructure — built for scarcity — must be rebuilt for abundance. Those that rebuild it become infrastructure. Those that don’t become a landfill with a search bar.”
The artist-centric model Deezer is piloting — requiring minimum listener thresholds before a track qualifies for royalties, and weighting payouts toward tracks with repeat, intentional listening — is the right directional move. But it’s a demand-side filter applied to a supply-side flood. The real game is whether Deezer can build detection, classification, and gating infrastructure fast enough to make its catalog signal-dense rather than noise-dense. Spotify has more resources to throw at this; Deezer has more urgency.
Three Implications
IMPLICATION 1 — ROYALTY POOL ARCHITECTURE IS BROKEN BY DESIGN
The pro-rata model was never designed to handle an infinite-supply input. Every DSP that hasn’t moved to a user-centric or intent-weighted royalty model is sitting on a structural subsidy for AI spam. The longer they wait, the more their human artist relationships erode — and human artists are the only supply-side asset that generates genuine listener loyalty and subscription retention.
IMPLICATION 2 — AI DETECTION IS THE NEW PLATFORM MOAT
The DSP that builds the most accurate, adversarially robust AI-content classification layer doesn’t just solve a moderation problem — it builds a defensible moat. Catalog signal quality becomes a product differentiator. Curators, labels, and artists will route to the platform they trust not to dilute them. Deezer’s public acknowledgment is a strategic signal that it intends to compete on this axis; the question is whether it has the engineering capacity to execute at scale.
IMPLICATION 3 — THE MUSIC INDUSTRY PREVIEWS EVERY CONTENT PLATFORM’S FUTURE
What is happening to Deezer today will happen to podcast directories, stock media libraries, e-book marketplaces, and eventually professional content networks within 18–36 months. The music industry’s compressed economics — tiny per-unit payouts, vast catalogs, thin margins — simply made it the first sector to hit the supply-side wall. Every platform built on user-generated or third-party content supply should be reading Deezer’s disclosure as a leading indicator, not a music-specific anomaly.
Business Engineer Framework
The Map of AI — Where Deezer Sits in the Stack
The Map of AI maps 200+ companies across 9 layers of the AI value chain — from infrastructure and models to applications and distribution. Understanding which layer a company occupies tells you which structural threats it faces when AI collapses the cost of adjacent layers. Deezer’s crisis is a Distribution layer crisis caused by a Foundation Model layer shock. The Map of AI framework gives you the language and the lens to see those dynamics before they become headlines.
Deezer crossing the 50% AI-upload threshold is not a moderation milestone — it’s a market structure inflection point, and the music industry’s compressed economics simply made it visible first. Every platform that monetizes third-party content supply through a fixed royalty pool or algorithmic feed is now on the same trajectory; the only variable is how long before they hit their own 50% moment and what infrastructure they’ve built in advance to handle it. The platforms that treat AI supply inflation as an engineering problem to solve — rather than a PR problem to manage — will own the catalog quality signal that defines the next decade of streaming economics.
Gennaro is the creator of FourWeekMBA, which reached about four million business people, comprising C-level executives, investors, analysts, product managers, and aspiring digital entrepreneurs in 2022 alone | He is also Director of Sales for a high-tech scaleup in the AI Industry | In 2012, Gennaro earned an International MBA with emphasis on Corporate Finance and Business Strategy.
Scroll to Top
Discover more from FourWeekMBA
Subscribe now to keep reading and get access to the full archive.
