AI Music Flood vs. Artist Survival: The Door Nobody Talks About
Sometimes I sit here and laugh at the headlines. “AI is the future of music.” “AI will help artists.” “AI will democratize everything.”
Really? Help artists? Open doors? I’ve been in this game long enough to know the problem was never about doors being closed. The problem is: the door ain’t getting bigger. And that door is called time.
Flood vs. ceiling
Every day, around 100,000 new tracks hit streaming (btw: industry-reported 2024 figure). That’s already overwhelming. And now AI has entered the room, multiplying that number by the thousands. Anybody with a laptop can crank out ten songs before lunch. Some even drop three albums in a month (e.g., Velvet Sundown in June 2025 - ~500k monthly listeners; multiple tracks flagged 100% AI by detection tools).
So yeah, uploads explode. Feels like growth. But here’s the catch: listeners don’t get extra hours in their day. Nobody woke up with 26 hours instead of 24. Which means… if you’ve got one hour for music, that’s maybe 20 songs. Period. Doesn’t matter if 100k or a million new songs dropped today - you’re still only fitting twenty.
That’s the bottleneck: human time. And once you see it that way, you realize it’s less about how much music exists and more about who actually makes it through the door. And that leads naturally to a bigger picture: if everyone’s stuck outside, how do you even get picked to step in?
Think of it like a club line
Imagine a club in Manhattan with a tiny door. Day one: 100 people waiting. The door lets in 10 at a time. Fine. Day two: 1,000 waiting. Same door. Chaos. Day three: 10,000 waiting. Same door. Nobody moving.
That’s streaming. Door - attention. AI-generated music = a flood of new bodies in line. Uploads grow. Time doesn’t. And here’s the twist: when the line clogs like that, it’s not talent that decides who gets in. It’s the bouncer. Which, in streaming terms, means the algorithm. And this is where AI shows its other face - sometimes useful, sometimes harmful.
The algo loves “fit”
Playlists don’t care who you are. They care if your track fits: tempo, mood, key, vibe. Humans write from feeling. AI writes from the algorithm’s checklist. It can spit out track after track until one lands exactly on target.
That’s why ghost bands are showing up. Velvet Sundown is suspected fully AI – yet racks up ~500k listeners (btw: 10/13 songs 100% AI per detection tests). People don’t notice or care. They’re not searching a name, they’re searching Deep Focus. If the vibe matches, they let it play.
But here’s the nuance: AI itself isn’t evil. The danger is in how it’s used. Used wrong, it floods playlists with soulless filler. Used right, it can help:
Vocal cleanup for singers without expensive studios.
Mastering assistance so indie artists sound professional without huge budgets.
Idea sketching where AI generates rough drafts that humans turn into finished songs.
Accessibility tools for disabled artists who can’t physically play instruments but still want to create.
So it’s not AI vs. artists – it’s junk use vs. smart use. And once you understand that, the next question is: how are people actually listening inside those playlists?
Skips and shorts
Five years ago, people still searched for artists. Now? Playlists dominate. And people skip – constantly. The average Spotify user skips once every ~4 minutes (btw: usage data). Gen Z often bounce before halfway.
So songs shrank. Hits went from ~3:30 average in 2019 to ~3:00 in 2024. At the 2024 Grammys, almost 20% of nominated songs were under 3 minutes. Viral tracks like Lil Yachty’s “Poland”? 83 seconds.
Shorter songs mean more plays per hour and less risk of a skip. And AI thrives here, because it can A/B thousands of hooks overnight and keep only the winners. For artists, AI could also be a tool to sketch ideas quickly or cut production costs, but the danger is when quantity replaces quality.
And if skips are the listening currency, what does that mean for the money? Let’s follow the chain.
How Spotify Really Pays Artists (in simple terms)
Here’s the part most people don’t see: the payout system itself. If you don’t understand how money flows, the whole “slices getting thinner” thing feels abstract. So let’s break it down like I’d explain to a friend.
The Pool. Spotify collects money each month – mostly subscriptions and ads. Roughly 70% of that revenue goes into a royalty pool for music rights holders. The other ~30% is Spotify’s cut to run the business.
The Split. That big pool doesn’t get handed out evenly. It’s divided pro-rata – meaning it’s based on total plays. If you got 1% of all streams worldwide, you get 1% of the pool. If you got 0.001%, you get 0.001% of the pool.
Labels vs. Indies. Here’s where it gets messy. Major labels (Universal, Sony, Warner) cut special licensing deals with Spotify back in the early days. That means:
They often negotiate higher effective rates per stream than indie distributors.
They sometimes get minimum guarantees or advance payments before streams even happen.
They also took equity in Spotify early on, so they benefited when Spotify went public.
Independent artists usually come through a distributor (DistroKid, TuneCore, CD Baby, etc.), which passes on whatever Spotify pays out minus its fee. No special perks.
The Content Types. Not all streams are equal:
Premium subscriber streams pay the most.
Ad-supported free streams pay much less.
“Functional” content (sleep sounds, white noise, etc.) still takes a cut of the pool unless the platform blocks it – which is why majors lobbied Deezer to demonetize it.
The Average. Because of all those factors, the average Spotify payout per stream to rights-holders hovers around $0.003 (a third of a cent). But:
A major label artist might effectively see closer to $0.004-$0.005 once you factor their deals.
An indie might see $0.002-$0.003 depending on distributor cuts.
Out of that, remember songwriters and publishers get their piece too – so what lands in the artist’s pocket is usually less.
Subscribers. By early 2025, Spotify reported ~248 million Premium subscribers, up around 14% year-over-year. That growth helps expand the pool, but not nearly as fast as the explosion of new tracks being uploaded. So yes, the pie gets a little bigger – but the number of people fighting for a slice grows much faster.
Why slices get thinner
Let’s connect the dots. The total pool doesn’t change much month to month – it only grows when Spotify gets new subscribers or more ad revenue. But the number of streams keeps rising as uploads explode and listening habits shift to shorter songs.
So if the pie is $100 and there are 100 plays, each play is worth $1. If the pie is still $100 but now there are 1,000 plays, each play is worth $0.10. That’s dilution. And that’s why more AI tracks in the system = thinner slices for everyone else.
This is why the majors are nervous, why Spotify keeps deleting spam, and why independent artists feel the squeeze the hardest. Because the math doesn’t care who made the song - it just slices by stream count.
Labels vs. sludge
“Majors will fix it.” Half true.
Universal struck a deal with Deezer (2023) to demote “functional noise” (rain, static, sleep loops) from royalties and reward engaged listening. Early results = ~+10% payouts to real artists (btw: Deezer CEO comments).
But majors also lean into AI for themselves. AI wellness/sleep playlists with their signed artists are already live. Why? Because functional music = ~15B streams/month across platforms. So yes, they’ll fight junk – but they’ll also build a cleaner version they can control. That shows AI isn’t going away; the question is who uses it, and for what purpose. And if that balance tips too far, we hit the breaking point.
The collapse point
If uploads keep exploding and listening time stays flat, the math only bends one way. More tracks + same time = thinner slices.
At some point, slices get so thin it’s not worth it. Nobody gets fed. Not the indie kid. Not the superstar. Not the AI bot.
And the cracks are already here: shorter songs, diluted payouts, platforms purging millions of junk tracks. Which leaves the question – where do we put our energy as artists?
Where I put my chips
For me, the math just proves what I’ve felt for years:
Streams are discovery, not foundation.
Playlists don’t build careers. People do.
The only thing that survives is what AI can’t fake: your story, your shows, your truth.
AI can still play a good role if we use it smart – as a tool, not a replacement.
Because when the door’s too small and everyone’s pushing, the only thing that matters is the handful of people who stop and say: nah, I’m here for you.
That’s time. That’s loyalty. That’s the currency AI can’t mint.
