There is a specific kind of AI-generated music that has been quietly making money on Spotify for the last two years — generic lo-fi beats, ambient-piano backgrounds, "focus music," "sleep music" — music that nobody really listens to, in the sense of sitting down and paying attention to it. It is designed to play behind something else. And because Spotify's pro rata royalty model treats a stream as a stream regardless of intent, every play of an AI-generated ambient track counts exactly the same as a play of a working musician's single.
For the last two years the music press has been focused on the copyright-infringement angle — did a generative model train on my song? — and that's a genuinely important fight. But the bigger, quieter story is on the payouts side. Even AI music that was trained legally and doesn't infringe on anyone's copyright is still, at scale, redistributing real money.
The pool problem
I've written before about how the pro rata pool actually works. A quick recap: Spotify doesn't pay per stream. It pays a share of the royalty pool proportional to your share of total streams. If the total number of streams on the service goes up 30% in a year, but most of that growth comes from AI catalogs and playlists nobody is paying active attention to, then every artist's share of the pool gets smaller, even if their own streams stayed flat.
This is exactly what has been happening. Spotify's total streaming volume has grown roughly in line with subscriber growth for the last decade. In 2024-2025, total streams grew significantly faster than subscriber revenue, and the gap is almost entirely attributable to AI catalogs that lean on background-music playlists. The pool grew slowly. The denominator grew fast. The quotient — per-stream payout — dropped.
Who benefits
Three groups:
- Companies that manufacture AI catalogs at scale. A handful of outfits — most notably the AI generators with corporate customers — have built catalogs of tens of thousands of tracks optimized for background-use playlists. Individually, each track earns very little. In aggregate, they earn real money.
- Playlist-seeding operators. If you can get your catalog onto algorithmic or editorial playlists that auto-play in cafes, fitness studios, and corporate environments, you don't need fans at all. You need placement.
- Spotify, indirectly. The company has consistently argued that its interest is in paying more money to more artists, but the economic reality is that AI-filled playlists have lower licensing costs (sometimes zero) and comparable ad-revenue generation to licensed background-music alternatives. Every minute of AI-filled focus-music playlist is, by definition, cheaper than a licensed equivalent.
Who loses
Everyone else. In particular: working musicians who depend on streaming as a meaningful percentage of their income. The math doesn't work equally for everyone — major-catalog artists with a vast back catalog and a wide listener base are relatively insulated because their share of the pool is driven by engaged listening, which doesn't slide as easily. The pain concentrates at the middle, where artists are getting 100,000 to 5 million streams a year and watching their effective per-stream rate drop quarter over quarter.
What Spotify has said, and not said
The official position is that AI is an "opportunity" and that the platform is working to ensure "quality" content is surfaced first. The policy responses have been partial and contradictory. The 1,000-stream demonetization threshold was framed as anti-fraud, but a lot of AI catalogs clear that threshold easily, while many legitimate new artists don't. Watermarking standards are being discussed but not adopted. Attribution requirements are patchy. None of it addresses the fundamental issue, which is that the pro rata model was designed in an era when all streams came from humans.
What to do about it
If you're an indie artist watching your streaming income drift downward without your release cadence changing, two honest reactions are available. One is to ride it out — Spotify is still a meaningful distribution channel, and nobody has a better one. The other is to diversify away. Bandcamp, direct-to-fan email, sync, merch, and live all pay on different math. The streaming economy is increasingly a single, concentrated risk in a modern artist's portfolio. Acting like it's the only one is a choice.
The services that claim they can fix streaming for you can't. The underlying model is what it is. Build around it.