Ten Thousand Versions of Your Song
On Spotify's derivatives announcement, who controls the infrastructure, and what 'new revenue stream' means from the artist's end
Spotify has announced that its technology for enabling fan-created AI remixes, covers, and derivative versions of existing songs is ready. The only barrier, the company says, is licensing. This is the sentence that should make every artist on the platform read the fine print very carefully — not because it is necessarily a bad development, but because 'the technology is ready and licensing is the only barrier' is what a platform CEO says when it has decided where it wants to go and is now negotiating the terms on which it gets there. And there’s no telling, in our current volatile environment, whether the terms will be fair.
Spotify has announced that its technology for enabling fan-created AI remixes, covers, and derivative versions of existing songs is ready. The only barrier, the company says, is licensing. This is the sentence that should make every artist on the platform read the fine print very carefully — not because it is necessarily a bad development, but because 'the technology is ready and licensing is the only barrier' is what a platform CEO says when it has decided where it wants to go and is now negotiating the terms on which it gets there. And there’s no telling, in our current volatile environment, whether the terms will be fair.
Well isn’t this narrative all too familiar? A new feature. A new revenue stream for artists. Fans empowered to engage with music they love in new and creative ways. Streaming already saved the industry from piracy. This will save streaming from stagnation. The question you are not being asked while this pitch is delivered is - who controls the tools, the data generated by the derivative interactions, the distribution of the derivatives, and who sets the royalty logic? Not the artist. Obviously.
Spotify's technology, as described, would allow fans to generate AI remixes and covers of existing songs inside the platform.(1) Unlike fully AI-generated music — which uses synthetic material trained on other people's work — derivatives would use actual recordings. The masters and publishing rights of real artists, real recordings made by real musicians, becoming raw material for an infinite number of fan-generated variations, all hosted on, tracked by, and monetised through Spotify's infrastructure.
Consider the arithmetic from the artist's position. You make one song. That song generates some streams, some royalties, some listener engagement. Under the derivatives model, that song potentially generates ten thousand AI variations inside the platform. Each variation is a separate piece of content competing for listener attention, recommendation algorithm placement, and share of the royalty pool. The original track now has to fight its own derivatives for discoverability.
Does the artist earn proportionally from the derivatives? Possibly. Does each derivative variation generate a royalty payment to the original rights holder? If so, at what rate, and compared to what the original stream would have generated? Does the platform take a larger percentage from derivative content on the grounds that it provided the generation infrastructure? Does the existence of ten thousand derivative versions change how the algorithm recommends the original? Does it increase streams or distribute them across the variants?
These are all unknowns. Supposedly they will be answered in the licensing negotiations, which will be conducted behind closed doors, between Spotify and the major labels who are its major shareholders, whose interests in this arrangement are certainly not congruent with the interests of independent artists.
The wider strategic context makes the derivatives announcement look even less like a fun new feature and even more like a strategic direction. Spotify's Q4 2025 results showed 751 million monthly active users, 290 million premium subscribers, a 33% gross margin, and €2.9 billion in free cash flow.(2) Management is explicitly repositioning the company as an 'agentic media platform' — a phrase that means, in practice, a platform that uses AI to mediate between listeners and content, learning from every interaction, building increasingly personalised audio environments. The AI DJ feature already has 90 million users. Prompted Playlists let listeners describe what they want in natural language and the algorithm builds it. The company's best developers reportedly have not written a line of code since December 2025 because AI is doing it for them.
The picture that emerges is of a platform building a 'language-to-music' data layer of extraordinary commercial value: a map of what hundreds of millions of people want to hear, in what contexts, in what moods, at what times, under what conditions, and how those preferences change when they are exposed to variations of familiar material. The derivatives feature generates exactly this data, at massive scale, using artists' existing recordings as the substrate.
One possible future might look like this. Human-created music becomes the 'prestige layer' and AI music the 'scalable utility layer,' with the platform monetising both while sitting in the middle. Human musicians provide the cultural legitimacy that makes the platform worth using. AI generates the infinite personalised (or should we say depersonalised) content that fills the space between the prestige moments. The platform captures the value created by both.
Of course, this is not necessarily the only possible version of how derivatives might work. Artists should be asking questions and demanding answers before any licensing framework is agreed. Questions like, can we opt in rather than having derivative use as the default? Are the royalty splits transparent and auditable? Do we own the data generated by derivative interactions with our work? Can we take down derivative versions that we find damaging or substandard? Who decides what qualifies as official?
A derivatives model with genuine artist control, clear royalty structures, and opt-in consent would be a different thing from a derivatives model that expands the platform's AI data layer using artists' recordings as raw material. Both are technically 'derivatives features.' The licensing framework is where the nuance lives.
The rights framework question is not abstract. Remixing inside Spotify would use actual recordings. Masters, publishing rights, neighbouring rights — real intellectual property created by real people.(3) The question of whether artist consent is required, and what form it takes, and what the royalty flow looks like, and what the opt-out mechanism is, will determine whether this is a new revenue stream for independent artists or a new extraction mechanism wearing a new revenue stream's clothes.
Technology moves fast. Rights frameworks move slow. Spotify knows this. The announcement that the technology is ready is a negotiating position: the thing exists, it will happen, the only question is the terms. The terms will be set by the parties with the most leverage in the negotiation, which are, not surprisingly, not independent artists.
The appropriate response is not opposition to fan creativity or to AI tools. Fans remixing music they love has a long and culturally productive history that predates Spotify by several decades and will likely postdate it, pending human created music surviving the robot invasion. The appropriate response is to ask what the terms are before they are set, to insist on opt-in rather than opt-out, to demand transparent royalty accounting, and to recognise that 'new revenue stream for artists' is a phrase that has appeared at every stage of the streaming economy's development, and that its track record as a description of what actually arrives is mixed at best.
Ten thousand versions of your song could mean ten thousand new listeners finding their way to the original. It could mean ten thousand variations competing with the original in the same recommendation algorithm, diluting your discoverability and your royalties while generating data for the platform's personalisation engine. The architecture is being decided right now, in rooms you are not in, by people whose interests are not yours.
Notes
Music Business Worldwide — Spotify eyes AI derivatives as new revenue stream. https://www.musicbusinessworldwide.com/spotify-eyes-ai-derivatives-as-new-revenue-stream-for-artists-and-says-its-tech-to-let-fans-make-remixes-covers-is-ready/
App Economy Insights / Bertrand Seguin LinkedIn — Spotify Q4 2025 analysis. https://www.appeconomyinsights.com/p/spotify-the-wrapped-effect
If you've made it this far, you probably care about where music is headed.
So do we — that's why we built something different. The Pack Music Co-operative is Australia's first musician-owned streaming platform: cooperative-governed, human-curated, and built on the radical premise that the people who make the music should own the infrastructure that distributes it.
Join the Pack — become an early adopter member, support our crowdfunding campaign, or lend your voice as an Ambassador: 👉 packmusic.au/join-the-pack
Back the campaign — every dollar goes directly toward getting us to launch: 👉 crowdfunding.startsomegood.com/thepackmusiccoop
Read our story — where we came from, why we built it, and what we believe: 👉 packmusic.au/who-we-are
Say hello — we genuinely want to hear from you: 👉 packmusic.au/contact