Music streamer Deezer says more than 50% of daily uploads are AI-generated

The Great Dilution Has Arrived

I’ve been watching the collision of generative AI and content platforms for a while now, and the latest numbers from Deezer are the strongest signal yet that we’ve reached a breaking point. For those who haven’t been tracking them, Deezer is a major European music streaming service that holds a decent chunk of the global market. They just dropped a bomb: more than 50% of the music being uploaded to their platform every single day is now AI-generated.

We aren't talking about a few experimental tracks here and there. We are talking about 90,000 files a day. This is the industrialization of creativity, and it’s happening faster than the platforms can build the infrastructure to manage it. As a founder, you need to look past the headline and see what this means for the future of digital assets, intellectual property, and what we actually define as a "creator."

Quantity is Not Quality

The problem isn't that AI music is uniquely bad—though a lot of it is. The problem is the sheer volume. In the old world, the bottleneck for music production was human effort. Even the most prolific bedroom producers had to physically sit down, play an instrument, or at least arrange a sequence in a DAW. That took time. That time acted as a natural filter.

Now, that filter is gone. You can script a process to generate thousands of lo-fi study beats or ambient noise tracks in an afternoon. These aren't being made for the love of the art; they are being made to capture a sliver of the royalty pool. If you upload enough tracks, you start to pick up pennies from the algorithmic playlists. Multiply those pennies by 90,000 tracks a day across multiple accounts, and you have a business model that is essentially a DDoS attack on the music industry’s bank account.

The War on Functional Audio

Deezer, along with Spotify and Universal Music Group, has been trying to fight back, but it’s an uphill battle. They’ve started focusing on what they call "functional audio"—white noise, rain sounds, and basic loops—which are the easiest for AI to replicate. The platforms are trying to de-monetize or reduce the payout for this kind of content because it’s drowning out the actual artists that people pay subscriptions to hear.

From a builder perspective, this is a classic game theory problem. If the cost of production drops to near zero, the value of the output trends toward zero unless you have a mechanism for curation or verification. Right now, the platforms are relying on reactive moderation, but that’s not sustainable when half your data stream is synthetic.

The Opportunity for Identity and Provenance

This is where the intersection of AI and crypto becomes interesting. Not the hype-driven version, but the practical application. We are entering an era where we can no longer trust the source of a digital file just because it exists on a reputable platform. If 50% of the music is AI, how do we prove what is human? How do we track the training data used to create those tracks so that original artists get their cut?

For developers, the real opportunity isn't in building another AI music generator. That market is already saturated and commoditized. The opportunity is in provenance. We need better ways to watermark human creativity and much better ways to track the pedigree of AI outputs. If you can build the tool that helps Deezer distinguish between a professional artist using AI as a tool and a bot farm pumping out garbage for pennies, you have a massive business.

The Founder’s Dilemma

If you're building in the creator economy space, you have to ask yourself: are you building a tool that adds noise or one that adds value? The Deezer stats show that we are currently being buried in noise. The platforms are getting desperate to find ways to protect their margins from being drained by synthetic accounts. They aren't just looking for AI detectors anymore; they are looking for identity solutions.

We are likely headed toward a world where "Artist" accounts require much stricter KYC or some form of verifiable credentials. The days of the anonymous upload being a gold mine are probably coming to an end. Platforms will have to prioritize human-verified content simply to survive the tidal wave of AI-generated content.

What This Means for You

• Platform Risk: If your startup relies on scraping or hosting user-generated content, expect your infrastructure costs to skyrocket as you try to filter out the junk.
• Curation is King: Algorithms are struggling with this volume. Human-led curation and niche communities are becoming more valuable because they provide a trust layer that the big streamers are losing.
• New Standards: Keep an eye on metadata. We need a new standard for how AI assistance is disclosed. Those who adopt transparent standards early will be the ones who survive the inevitable regulatory crackdown.

A Skeptical Look at the Future

I’m skeptical that the current streaming model can survive this. The royalty system was designed for a world where people made records, not where servers generated files. When 50% of your uploads are synthetic, your business isn't a music company anymore; it’s a data storage facility.

The takeaway for builders is clear: Don’t build the engine that makes the noise. Build the filter that makes the noise irrelevant. The real money isn't in the generation; it's in the verification and the distribution of value back to the humans who started the process in the first place.

The democratization of creation was supposed to give everyone a voice; instead, it's given the bots a megaphone that's drowning everyone else out.

Read the original at TechCrunch AI →