Deezer is now receiving nearly 75,000 AI-generated tracks every day — roughly 44% of all new music uploaded to the platform. Before you panic, here’s the catch: those AI tracks account for only 1-3% of total streams, and 85% of those streams are fraudulent. This isn’t a demand revolution. It’s a supply flood, and streaming platforms are drowning in noise.

The numbers were released by Deezer on April 20, 2026, and they paint a picture that the music industry has been dreading.

The Flood by the Numbers

Let’s lay this out because the scale is bonkers:

The trajectory is even more telling. In January 2025, Deezer was seeing about 10,000 AI-generated tracks per day. That number has increased 7.5x in just over a year. Every single day, nearly as many AI tracks hit Deezer as the total number of songs the entire human race released in a year a decade ago.

But here’s the thing that cuts through the hype: almost nobody is listening to them. The 1-3% stream share tells you this is a supply-side problem, not a demand revolution. And when you strip out the 85% that are clearly fraudulent — bots streaming bots’ music to generate royalty payments — the real listening figure is vanishingly small.

The Economics Are About to Break

CISAC (the International Confederation of Societies of Authors and Composers) partnered with PMP Strategy to model what happens next. Their finding: nearly 25% of creators’ revenues are at risk by 2028, potentially €4 billion.

The mechanism is brutal. Streaming platforms split royalty pools among all rightsholders based on stream share. If AI-generated tracks flood the pool — even if they’re mostly not being listened to — they still dilute the payout per stream. Real musicians get less for the same amount of actual listening.

Deezer has taken the most aggressive stance of any major platform. Since January 2025, it’s been running its own AI-detection tool, tagging AI-generated music, and removing it from algorithmic recommendations and editorial playlists. It’s even stopped storing hi-res versions of AI tracks. In June 2025, Deezer became the first (and still only) streaming platform to explicitly tag AI-generated music.

Deezer is now licensing its detection tech to the rest of the industry. Whether anyone takes them up on it is another question.

The Suno/Udio Problem

Deezer’s detection tool can identify music from the most prolific generative models — Suno and Udio being the big ones — and the company claims it can add detection for practically any new tool as long as it gets relevant data samples. It’s also made progress on “generalizability” — detecting AI content without needing a specific training dataset for each new generator.

But this is an arms race. Each new generation of AI music tools produces more convincing output, harder to distinguish from human creation. A Deezer-commissioned Ipsos study found that 97% of people couldn’t hear the difference between AI and human-made music. That’s the core problem: the technology has already crossed the perceptual uncanny valley. If you can’t hear the difference, the badge becomes the only signal.

What This Means for NZ

New Zealand has a disproportionately strong music export scene relative to our size — Lorde, Fat Freddy’s Drop, Six60, Benee, and a thriving independent scene. NZ On Air has funded local music for decades, operating on the assumption that human-created music has cultural value worth preserving.

But here’s the existential question NZ On Air and Creative NZ are facing: should public funding support human musicians creating AI-generated music? If a musician uses Suno to generate backing tracks and adds a vocal, is that “New Zealand music”? What about fully AI-generated tracks with Kiwi accents trained on local music?

The answer isn’t obvious. NZ On Air’s funding model was built in a world where “making music” meant human skill, time, and creativity. In a world where anyone can generate a passable reggae track in 30 seconds, what exactly are we funding?

Meanwhile, Kiwi musicians trying to break internationally face an even more crowded discovery landscape. Streaming platforms were already a lottery. Now there are 75,000 AI tracks per day competing for the same algorithmic attention.

🔍 THE BOTTOM LINE

44% uploads but 1-3% listens — AI music is a supply flood, not a demand revolution. The platforms are drowning in synthetic noise while real musicians see their royalty pools diluted. The badge is a bandaid, not a solution.

❓ Frequently Asked Questions

Q: Does AI music actually sound real? Yes. Deezer’s own research found 97% of people couldn’t hear the difference between AI and human-made music. The perceptual uncanny valley has been crossed. Detection now requires technical analysis, not human ears.

Q: How does AI music affect real musicians’ pay? Streaming platforms use pro-rata royalty pools — all subscription revenue goes into one pot, divided by stream share. Every AI track, even if not listened to, dilutes the pool. More content chasing the same revenue = less per stream.

Q: Can this be stopped? Detection technology can identify AI-generated music and exclude it from recommendations. Deezer has shown it’s possible to reduce AI fraud to “a minimum.” But enforcement requires industry-wide adoption, which hasn’t happened yet.

Q: Is anyone actually listening to AI music? Hardly. AI tracks account for 44% of uploads but only 1-3% of streams. And 85% of even those streams are fraudulent — bots generating royalty payments. Real human listening to AI music is a rounding error.

📰 SOURCES

• Deezer Newsroom — “AI-generated tracks now represent 44% of all new uploaded music” (April 20, 2026)
• MusicTech — “44% of music uploaded to Deezer every day is AI-generated”
• NME — coverage of Deezer’s AI music statistics
• Billboard — reporting on AI music’s impact on streaming
• CISAC/PMP Strategy — AI study projecting €4 billion creator revenue at risk by 2028