AI's already winning in music. We just haven't noticed yet

Everyone's busy dunking on Sora while synthetic quietly becomes the baseline

Jay Kim

Jay Kim

— 3 min read
AI's already winning in music. We just haven't noticed yet

Suno is the tell.

Suno Reportedly Set to Secure Funding at a $2 Billion Valuation
Suno is reportedly preparing to raise $100 million at a $2 billion valuation, which is four times larger than its value in May 2024.
Digital Music NewsDylan Smith
  • $150M ARR (~4x YoY)
  • reportedly raising at ~$2B
  • ~60% gross margin (vs. ~40% for AI coding agents)
audio models are smaller and cheaper to run than LLMs, which may help explain the gap.

Clearly not a ChatGPT-tier breakout yet, but way past the one-week gimmick.

The industry is already repositioning

UMG/WMG are wiring up AI licensing to mirror streaming economics (YouTube-style attribution + micropayments).

Spotify purged millions of spam tracks, yet keeps verified AI acts live in feed, now even co-launching AI products with the major labels.

for anyone wondering what this prompt actually sounds like

Exactly the kind of normalization you do before you industrialize.

Audience tolerance is sliding from "never" to "if it slaps, it plays," and platforms are triaging supply as they lay the rails for scale.

Spotify's incentives are straightforward

~70% of the pie flows to labels/artists today.

If a low-royalty or owned AI catalog can be deployed, payout pressure drops and margins expand.

Headliners will stay mostly human, but the long tail is easily machine-addressable.

Why?

Most of our listening is background anyway (driving, working, lifting), and AI "artists" neatly fit that dominant use case.

at least the royalty excuse won't fly when the time comes...

Now, obviously none of this is easy.

Spotify's walking a tightrope: wooing majors and human artists, policing slop, yet still leaning into AI where it improves retention, catalog depth, and margins.

Labels, in turn, are angling for a cut not just on plays but on training + generation.

Somewhere in that push and pull, whoever cracks attribution ends up setting the tolls.

Which is where the economics spill into culture

We're nearing the point where you won't be able to tell, so creators will start tagging work "No AI used."

The artistic analogue to a FAIR TRADE label.

Probably a small slice of listeners will pay to signal that choice, to feel virtuous backing human-only pipelines.

But if the label exists only because our ears can't tell, how many will still care once the vibe lands and the price drops?

Truth is, that's how synthetic becomes the default, with our full consent while we call it progress.

Read more