Spotify Just Deleted 75 Million Tracks — Here's What It Means for Emerging Artists

In the twelve months leading up to September 2025, Spotify removed more than 75 million tracks from its platform — music it classifies as spam. That's not a typo, and it's not a one-off purge. It's the start of a permanent shift in how the biggest streaming service decides what gets heard.

For an emerging electronic artist, this is the most consequential change on the platform in years. It rewrites who gets recommended, whose uploads survive, and how AI shows up next to your name. Here's what actually happened — and what to do about it.

Three Things Changed at Once

Spotify's September 2025 policy update moved on three fronts simultaneously.

A spam filter. Rolling out from late 2025, a new system hunts the tactics people use to game the platform: mass uploads, duplicate tracks with slightly altered metadata, SEO manipulation in track and artist names, and “artificially short” tracks — songs uploaded at just over 30 seconds, the minimum length that earns a royalty, purely to farm plays. When the filter catches an uploader, it tags the tracks and stops recommending them. No algorithmic push, no editorial consideration.

AI labelling. Spotify has adopted DDEX, an industry-standard disclosure format that lets you declare where AI was used in a track — generated vocals, instrumentation, or post-production. Distributors pass it through, and it appears in a song's credits. The in-app beta went live on 16 April 2026. This isn't a ban on AI. It's a requirement to be honest about it.

Impersonation rules. Vocals that clearly imitate another artist now require that artist's documented consent. Unauthorised AI voice clones are removed on sight, and Spotify is testing tools with distributors to stop fraudulent uploads landing on the wrong artist's profile.

The reasoning behind all of it came from Spotify's Sam Duboff:

“When music doesn't take much effort to create, it tends to be low quality.”

And low-quality music, the platform is now betting, doesn't deserve a recommendation.

Why This Hits Electronic Music Hardest

Electronic music is the genre most exposed to every one of these changes — for reasons that have nothing to do with your talent.

The tools are cheap and fast. A generative model can produce a passable loop in seconds, which is exactly why the spam wave landed here first. The same accessibility that lets a bedroom producer finish a track also lets a spam farm upload ten thousand of them. When Spotify tightens the net, honest producers working in the same sonic territory can get caught by association.

Short edits are native to the genre. Sped-up versions, VIP edits, intro tools, DJ-friendly cuts — electronic artists legitimately release short tracks. The filter doesn't read intent; it reads length and behaviour. If your catalogue is a wall of sub-90-second uploads, you look like the problem.

AI is already in the workflow. A generated pad, a cleaned-up vocal, a texture layer. Not deception — just production. Under DDEX that's completely fine, but only if it's declared.

What to Actually Do

This is a good change for anyone making real music, because the noise you were competing with is being cleared out. Five moves put you on the right side of it:

1. Slow your release schedule down. Volume for its own sake is now a liability, not a growth hack. A steady cadence of finished tracks beats a flood of half-ideas — and always did.

2. Declare your AI use. If you used a generative tool for any part of a track, disclose it in your distributor's DDEX fields. It costs nothing and protects you from being flagged as deceptive later.

3. Clean your metadata. No keyword-stuffed titles, no fake features, no near-duplicate uploads of the same song. Every version you put up should be a real, distinct release.

4. Be careful with short tracks. Legitimate edits are fine — but don't build a catalogue out of 31-second uploads. Keep tool tracks and intros as a genuine part of a real body of work.

5. Never touch another artist's voice without permission. An AI vocal that sounds like someone recognisable isn't a bold sample. It's an instant removal, and now a fraud flag.

The Bigger Picture

Strip away the mechanics and this is a platform deciding that effort and authenticity are worth protecting. For a few years, the game rewarded whoever uploaded fastest. That game is closing.

For emerging artists who care about their craft, that's the best news in a long time. The barrier to being heard was never really talent — it was noise. Spotify just deleted 75 million pieces of it.

At ANGVIS, this is the world we've always built for: real artists, real releases, and real people deciding what moves forward. The platforms are finally catching up.