Deezer has disclosed that AI-generated music now makes up roughly 44% of all new tracks uploaded to the platform each day — about 75,000 fully AI-generated songs in every 24-hour cycle. The figure marks a step-change in the composition of the music supply itself, not just a tail-end curiosity.
The listening picture is more nuanced. Despite dominating uploads, AI tracks account for only 1–3% of total streams on the platform. Listeners are not actively seeking out these songs in meaningful volume. That ratio — flooding the library but capturing a sliver of attention — points to a supply-side problem rather than a demand shift.
The fraud numbers are the more alarming signal. Deezer reports that 85% of the streams that AI-generated tracks do receive were flagged as fraudulent in 2025. The pattern is well-understood at this point: automated services upload mass-generated tracks, then use bot farms to stream them and siphon royalty pools that would otherwise flow to human artists.
For streaming platforms, the policy question is sharpening quickly. Deezer has moved toward explicit tagging of AI-generated content and exclusion from royalty-sharing algorithms, but the detection side is a moving target — generation quality is improving faster than discrimination tooling. The 44% number is likely a floor, not a ceiling, and the industry's revenue-distribution model was not designed for a world where most of the catalog is synthetic.