The 75 Million Removal Statistic
In September 2025, Spotify announced the removal of over 75 million spammy tracks over the preceding 12 months.
| Metric | Value |
|---|---|
| Tracks removed | 75+ million |
| Timeframe | September 2024 - September 2025 |
| Total Spotify catalog | ~100 million tracks |
| Removal as % of catalog | ~75% equivalent |
Key context: Some tracks were blocked before upload via filtering, while others were removed after being identified post-publication.
What Was Actually Removed
The 75 million figure includes multiple categories of problematic content:
Spam Tactics Targeted
| Tactic | Description |
|---|---|
| Mass uploads | Bulk content flooding the platform |
| Duplicate tracks | Same content uploaded multiple times |
| SEO manipulation | Metadata gaming for discovery |
| Artificially short tracks | Under 30-second loops to maximize per-stream payments |
| Stream farming content | Tracks designed purely for fake streams |
Note The removals were not exclusively AI-generated music. Spam, fraud, and low-quality content existed before AI tools, and the removal statistics reflect a broader enforcement action against royalty manipulation.
What This Was Not
The 75 million removal was not:
- A targeted action against all AI music
- Removal of legitimate AI-generated tracks
- Based solely on AI detection
Spotify's enforcement focused on behavior patterns (bulk uploads, manipulation) rather than the source of creation.
Why This Action Was Necessary
Spotify's payouts have grown dramatically, attracting exploitation:
| Year | Total Payouts |
|---|---|
| 2014 | $1 billion |
| 2024 | $10 billion |
"Big payouts entice bad actors," Spotify stated. The growth of generative AI made spam tactics "easier to exploit" at scale.
The problem: Every stream over 30 seconds generates royalties. Spam tracks dilute the royalty pool, reducing payments to legitimate artists.
Comparison to Other Platforms
Deezer Data
| Metric | Value |
|---|---|
| Daily AI uploads | 20,000 tracks |
| AI as % of uploads | 28% |
| AI as % of streams | 0.5% |
Deezer's data shows the gap between upload volume and actual engagement. AI accounts for 28% of uploads but only 0.5% of streams, suggesting much AI content is low-quality or spam.
Spotify's Three-Part Response
Following the mass removal, Spotify announced a full policy:
1. New Spam Filter
A system to identify and stop recommending spammy uploaders and tracks. Rolling out gradually to avoid penalizing legitimate creators.
2. Impersonation Policy
Prohibition of:
- Unauthorized AI voice clones
- Deepfakes
- Vocal replicas
- Artist impersonation
3. DDEX AI Disclosures
Support for industry-standard metadata indicating AI involvement in vocals, instrumentation, or post-production.
What This Means for AI Music Creators
If You Create Legitimate AI Music
Good news: Quality AI music with proper licensing remains welcome on Spotify.
Requirements:
- Commercial rights from your AI tool
- Proper metadata and disclosure
- No artist impersonation
- No spam behavior
Warning Signs to Avoid
| Behavior | Risk Level |
|---|---|
| Uploading 100+ tracks at once | High |
| Identical or near-identical tracks | High |
| Tracks under 30 seconds | High |
| Metadata gaming (fake genre tags, etc.) | High |
| Single tracks with proper metadata | Low |
| Curated album releases | Low |
How to Stay Safe
Best Practices
- Quality over quantity: Release curated catalogs, not mass uploads
- Proper metadata: Accurate genre, mood, and track information
- No manipulation: No fake streams, no gaming algorithms
- Follow policies: Stay current on Spotify's guidelines
- Disclose appropriately: Use DDEX AI disclosure when available
The 1,000 Stream Minimum
As of 2025, Spotify requires tracks to have at least 1,000 streams in the past year to generate royalties. This policy particularly affects:
- Low-engagement spam content
- Tracks without legitimate audience
Legitimate AI music with genuine listeners is unaffected.
Detection and Enforcement
Spotify's approach combines:
Pre-upload filtering:
- Pattern recognition for spam behavior
- Detection of known manipulation tactics
Post-upload monitoring:
- Streaming pattern analysis
- User reporting
- AI detection technology (developing)
Ongoing enforcement:
- Tracks can be removed at any time
- Uploader accounts can be flagged
- Distribution partners held accountable
Industry Impact
Label Support
Universal Music Group "welcome[s] Spotify's new AI protections as important steps forward consistent with our longstanding Artist Centric principles."
Distributor Response
Distributors are implementing:
- Stricter upload review
- AI disclosure requirements
- Better filtering before submission
The Bigger Picture
The 75 million removal reflects the streaming industry's growing pains with AI:
| Year | Situation |
|---|---|
| Pre-2024 | Limited AI music, manageable spam |
| 2024 | AI tools democratize music creation |
| 2025 | 7M songs/day on Suno alone; spam explodes |
| 2025-2026 | Platform enforcement catches up |
AI music is not being rejected, but low-effort exploitation is being filtered. The platform distinction is between:
- Creators making quality AI music (welcome)
- Operators flooding the system with garbage (removed)
For legitimate AI music creators, these enforcement actions protect your royalty share by removing content that dilutes the pool without contributing genuine value to listeners.