In the first quarter of 2026, a data point emerged from Deezer that should unsettle anyone who believes in the scarcity of human expression: over 90,000 AI-generated tracks now flood its library every single day. This is not a prediction or a hypothetical—it is a measured reality. As a macro watcher who has spent years tracking the intersection of technology, finance, and cultural infrastructure, I see this number as a seismograph for a deeper tectonic shift. The sound of 90,000 algorithms humming in parallel is the sound of a trust mechanism breaking. But trust, like liquidity, does not survive a vacuum for long.
The hollow resonance of digital ownership in art is no longer a philosophical debate—it is a structural problem that demands a new layer of verification. To understand its magnitude, we must first examine the context. Deezer, a streaming platform that holds a fraction of Spotify’s global share, is not an outlier; it is a bellwether. The AI music models powering this deluge—Meta’s AudioCraft, Google’s MusicLM, along with platforms like Suno and Udio—have matured from research projects into production-grade tools. Their outputs are now cheap, fast, and indistinguishable from human composition to the untrained ear. But the key insight lies not in the technology of generation, but in the absence of an equivalent technology of verification.
Based on my experience auditing the fragility of decentralized systems during the 2020 DeFi Summer, I recognize a familiar pattern: a promising abstraction (decentralized finance, then; AI-generated content, now) that scales rapidly while ignoring the foundational trust infrastructure required to sustain it. The 90,000 daily tracks are a liquidity event—a flood of tokens with no provenance. And just as liquidity mining APY in DeFi masked the hollow core of subsidized TVL, these AI tracks mask the hollow resonance of creative abundance. The question is not how they are made, but how we will know what they are.
The core of the crisis is not technological but epistemological.
When I led a six-month audit of SWIFT versus Ethereum-based settlement layers in 2017, I saw how hidden fees eroded trust in cross-border transfers. Today, Deezer faces an analogous erosion: every AI track that passes as human makes the next human track suspect. The platform cannot afford to simply ban AI content—that would cut off a massive source of user engagement and future revenue. Nor can it embrace indiscriminately without alienating the creators who form its cultural bedrock. This is a prisoner’s dilemma at scale, and the numbers are about to get much worse. If 90,000 tracks per day is Deezer’s baseline, consider the global volume across Spotify, Apple Music, and Tencent Music. We are likely looking at millions of daily AI uploads by the end of 2026.
The macroeconomic implications mirror the stablecoin debates I have followed in Geneva’s regulatory corridors. Just as PayPal launched PYUSD not out of technological altruism but to hedge regulatory risk, music platforms will soon be forced to align with either a regulatory framework or a technical solution for content provenance. The EU AI Act already mandates transparency requirements for AI-generated content—but those requirements are based on self-reporting. Without a neutral, tamper-proof layer of attestation, the law will be as effective as a paper wall in a hurricane.
This is where a contrarian angle emerges: the decoupling thesis. Mainstream commentary assumes that AI will destroy human art, but the opposite may be true. The very abundance of synthetic content will create a premium for authenticated human origin. In finance, after the 2022 liquidity freeze, survival metrics (audited reserves, real-time solvency) became more valuable than growth metrics. The same logic applies to music: platforms that can verify the human source of a track will command a trust premium that translates into higher subscriber willingness-to-pay and lower friction with label partners. The metabolic cost of synthetic creativity—in terms of energy, legal risk, and brand dilution—will become a liability that only rigorous verification can offset.
During the 2021 NFT mania, I tracked the energy consumption of Ethereum’s Proof-of-Work and calculated that ten thousand high-profile art pieces consumed more carbon than 100,000 households in Geneva. I saw the hollow resonance of digital ownership then—people paying for a token that proved nothing about the art’s origin or the artist’s intent. Today, the 90,000 daily AI tracks are a similar token of emptiness, but with a far higher volume. The difference is that today, we have a viable counter-technology: blockchain-based provenance, combined with zero-knowledge proofs, can timestamp and verify the creative authorship of a track without revealing the underlying training data or infringing on privacy.
In 2026, living in Geneva’s regulatory hub, I facilitated a roundtable between EU regulators and AI developers. The consensus was that 70% of AI training data lacks provenance—a gap that blockchain can fill via transparent, immutable logging. Deezer’s data point is the canary in the coalmine. Every platform will soon need a provenance layer, not as a luxury but as a survival metric. The architecture of attention in a post-authenticity era demands that we know not just what we are listening to, but who—or what—created it. The absence of that knowledge is already creating a silent crisis: listeners are beginning to distrust the entire platform.
Yet, a deeper risk remains unaddressed: the legal status of DAO-like collectives that manage AI music rights. Most DAOs have the legal status of “no legal status,” and when things go wrong, members face unlimited personal liability. If a Deezer user uploads an AI track that mimics Drake’s voice and generates royalties, who is liable? The user, the AI platform, the model trainer, or Deezer itself? The law has no answer, and the 90,000 tracks per day are stress-testing that silence.
The resilience of the music industry will depend on its ability to shift from a scarcity-based model to an authenticity-based one.
The takeaway is forward-looking and, I admit, uncomfortable. We are witnessing the liquidity of creativity evaporate as trust fractures. But just as the cross-border payment industry learned to embed compliance into the transaction layer, the music industry must embed verification into the production layer. Regulators lag, capital moves—but technology can bridge the gap. The question is not whether AI will replace artists, but whether we can build a verification layer that restores the value of authenticity. The hollow resonance of digital ownership in art finds its echo in the silence of a platform drowning in its own data. The future belongs not to the largest catalog, but to the most trusted one.
The architecture of attention in a post-authenticity era will be built not on algorithms that generate content, but on algorithms that verify it. And the first platform to implement that with rigor will own the next decade of music. The 90,000 daily tracks are a warning, but they are also an invitation: to build the trust infrastructure that the industry has always needed but never had. The metabolic cost of doing nothing is already calculable. The question is whether anyone will listen.
