Spotify is building an AI product that lets fans make covers and remixes of real artists’ music, and it just got a lot bigger. During the company’s second-quarter earnings call on Tuesday, Spotify announced that Merlin, a licensing partner for independent labels and distributors, has joined the effort alongside Universal Music Group. According to TechCrunch AI, the Merlin deal brings more than 30,000 labels into the product, which will only feature artists who agree to take part.
What stands out here is the framing. Spotify isn’t chasing the fully synthetic “AI artist” trend that’s swamping streaming platforms. Co-CEO Gustav Söderström told investors the product is about “real artists, not fake artists.” That’s a deliberate line in the sand, and it matters given where the market is heading.
What Spotify is actually building
- A consent-based remix and covers engine. Fans get to “play around with covers and remixes” of an artist’s work, but only if that artist opts in. Co-CEO Alex Norström framed it around three commitments: consent, credit, and compensation.
- A new pay structure for artists. The tool launches as a paid add-on, which Spotify says creates an additional revenue stream for the artists who participate. Norström called it “the first legal way to partake in this AI tailwind” for interactive music.
- A growing rights base. UMG signed on first. Merlin’s 30,000-plus labels now widen the pool to include a large chunk of the independent world, not just the majors.
- A low barrier to launch. Spotify told investors it won’t need a full music catalog to get started, so the product can ship with a partial library and expand from there.
Why the timing is sharp
AI music is flooding streaming services, and the numbers are getting hard to ignore. TechCrunch AI reports that rival streamer Deezer recently found more than 50% of its daily track uploads were AI-generated, up from 10% in January 2025. That’s a fivefold jump in roughly a year and a half.
Spotify’s pitch is that it can channel that wave instead of drowning in it. Söderström put it plainly: “Normal generative music will happen with or without us. This product will not happen without us, and it needs to exist so that existing artists can participate in this.” In other words, the company is betting that the value isn’t in generating songs from nothing. It’s in giving the artists people already love a legal, paid way to join in.
How it compares
Most AI music startups generate fully artificial songs with no artist involvement and murky rights. Spotify’s approach flips that. Every track in the catalog traces back to a consenting, credited, and compensated artist. That’s the differentiator the company keeps hammering, and it’s a defensible one. Licensing deals with UMG and Merlin are hard to replicate, and they give Spotify something the pure-generation tools can’t easily match: permission at scale.
The catches
There are real caveats. Spotify still hasn’t given a launch date. The company only said a research preview would go out to a subset of users first, with no timeline attached. It’s also a paid add-on, so fans will pay extra to use it. And the whole thing depends on artists choosing to opt in, which means the catalog’s usefulness hinges on how many actually say yes.
This is significant because it’s one of the first serious attempts by a major platform to make AI music remixing a licensed, revenue-generating product rather than a legal headache. If the consent-and-compensation model works, it could become the template other streamers follow. If artists stay away, it stays a niche add-on.
Watch for the research preview and any signal on when it goes wide. Full details are available at the original TechCrunch AI report.