Back to News
Market Impact: 0.25

Deezer says AI song uploads have nearly overtaken human music

SPOT
Artificial IntelligenceTechnology & InnovationMedia & EntertainmentCompany FundamentalsProduct Launches
Deezer says AI song uploads have nearly overtaken human music

Deezer says AI-generated songs now account for about 44% of its daily uploads, up from 10,000 to 75,000 per day since launching its detection tool in January 2025. However, AI tracks still make up only 1% to 3% of total streams, and Deezer is continuing to demonetize, tag, and remove them from recommendations while licensing its detection tool to other companies. The update highlights growing operational pressure from AI music, but the direct market impact appears limited.

Analysis

The important second-order issue is not monetization leakage from AI tracks today; it is catalog inflation. When low-friction generation overwhelms ingest, the platform’s core economic asset shifts from being a music marketplace to being a spam-filtering and trust layer, which raises moderation cost while diluting recommendation quality. That dynamic is more damaging to engagement than the headline stream share suggests, because recommendation systems tend to compound small ranking errors over time. For SPOT, the near-term P&L impact is probably modest, but the strategic risk is larger: if users increasingly perceive the catalog as noisier or less curated, session depth and retention can erode before revenue metrics show it. The company is also signaling a willingness to harden curation, which can help brand trust but may reduce long-tail discovery and create friction for legitimate creators using AI as a tool. That tradeoff matters over a 6-18 month horizon as platform differentiation shifts from breadth to verified quality. The competitive winner is likely whichever platform can authenticate provenance at lowest friction. If Deezer’s tool becomes licensable infrastructure, the real upside may accrue to detection vendors and rights-management middle layers, not the streaming services themselves. A more contrarian angle: the market may be underestimating how quickly AI music could become a legal and contractual problem for labels and distributors, forcing industry-wide standards that advantage platforms with better metadata pipelines and compliance tooling.

AllMind AI Terminal