three years ago, a chicago producer named marcus told me he couldn't make a music video because he had no budget. he'd finished a track that deserved one... something visual, something that moved. but videos cost five to fifteen thousand dollars if you want them done right. and marcus had rent.
i told him what i always tell musicians in that position: save it for after you blow up. make the song so good the label will fund the video. that was the path. that was always the path.
marcus never blew up. the song is still sitting in a folder somewhere.
last month, i watched him generate a music video using runway in about four hours. it wasn't perfect. but it moved. it had rhythm. it looked like something someone made on purpose, not something pulled from stock footage. he spent maybe sixty dollars on credits. the rest was his time and taste.
this is the moment everyone's talking about... and everyone's getting wrong.
The Bottleneck That Wasn't Actually the Bottleneck
the narrative right now is "AI just democratized music video production." and technically, that's true. five years ago, you needed either money or connections to make a video that didn't look like you shot it on an iphone in a parking lot. now you don't.
but here's what nobody says out loud: music videos were never the problem.
music videos were the excuse.
"i'd release this but i need a video." "i'd submit to playlists but my visual identity isn't dialed in." "i can't pitch to sync because i don't have professional video content." these are all real sentences i've heard from real musicians. and they're all partially true. but they're also the comfortable version of a harder truth.
the truth is... getting onto a playlist is harder than making a video. getting sync licensing is harder than making a video. building actual discovery is harder than making a video. and those problems don't have a sixty-dollar solution.
when you remove the video excuse, you're left with what you've always been left with: a song that needs to reach people who don't know it exists.
What AI Actually Solved (And What It Didn't)
let me be specific about what changed, because the specificity matters.
before 2023, the production pipeline for an independent artist looked like this:
write track ... record track ... mix track ... master track ... need video for release ... don't have money ... release without video ... song gets no playlist adds ... repeat.
now it looks like this:
write track ... record track ... mix track ... master track ... generate video in four hours ... release with video ... song still gets no playlist adds.
the second pipeline is objectively better. you have an asset you didn't have before. spotify's algorithm does see music videos. they do matter for social sharing. a finished-looking music video is better than no music video, period.
but here's what the ai tools didn't touch: they didn't make your song better. they didn't make it more discoverable. they didn't change the fundamental economics of music distribution or the way streaming royalties work or the gatekeeping that happens at playlist editorial.
spotify's algorithm is opaque on purpose. editorial playlists are still curated by humans with taste hierarchies that favor certain label partnerships and certain regions and certain genres. and if you're an indie artist in chicago making experimental hip-hop, those systems are not designed for you.
the ai video tool did not change that. it only gave you one less excuse for not dealing with it.
The Real Bottleneck: Discovery Without a Deal
here's the number that matters: spotify released 70,000 songs per day in 2023. by 2024, it was probably higher. maybe 80,000. maybe 100,000 soon.
your song is one of them.
making it beautiful doesn't change that math. making it look expensive doesn't change that math. the discovery problem is structural. and it exists because the incentive system in streaming is broken.
spotify pays an average of $0.003 to $0.005 per stream. a song needs about 250,000 streams to generate $1,000 in revenue. that means if you want to make minimum wage from music in 2024, your song needs to reach approximately 7.5 million people who actually listen to it all the way through.
that is not a production problem. that is not a video problem. that is a discovery problem. and discovery without a label behind you is functionally impossible at scale.
this is what nobody tells you: the label budget excuse existed for a reason. labels have discovery infrastructure. they have relationships with playlist curators. they have marketing teams. they have money to run ads that actually move the needle. a music video was one component of a much larger ecosystem designed to get songs in front of people.
when you're indie, you don't have that ecosystem. you have a song and a thirty-dollar ai video tool.
that's a gap that doesn't close.
Sync Licensing: The Problem AI Can't Touch
there's one revenue stream that actually scales for independent musicians, and it's not spotify. it's sync licensing.
sync licenses are when your song gets placed in a commercial, a film, a podcast, a video game, a youtube creator's video. one placement in the right place can generate more money than a million spotify streams. one indie musician i know got their track into a net flix series and made $12,000. they had made maybe $3,000 from streaming in their entire life.
but here's the thing: sync licensing is relationships. it's having an agent who knows music supervisors. it's being part of networks that music supervisors actually check. it's having material that's professional enough to be taken seriously.
and nobody talks about this, but the "professional enough" bar got higher, not lower, when the barrier to entry went down.
five years ago, a music supervisor would maybe hear 500 submissions a month. now they hear 5,000. the ai video tool means your song has visual content... which means more people are submitting through the same channels... which means music supervisors are even more overwhelmed.
you didn't fix the bottleneck. you just added more people to the queue.
the actual sync licensing pipeline still requires either an agent (who takes 15-20% of revenue), or direct relationships you built over years, or luck. none of those things changed. the ai video tool helped your song look finished. but it didn't help your song get discovered by the people whose job it is to place music.
What Indie Artists Actually Need (And What AI Can't Give)
i've been thinking about this wrong for a while, and i want to correct it in real time.
we've spent the last three years talking about what ai can automate. content creation, video generation, artwork, metadata. and those tools are real and useful. but we haven't spent enough time asking: what does an independent musician actually need that money or algorithms or tools don't provide?
the answer is infrastructure.
not the kind you buy. the kind you build or join or inherit.
a label is infrastructure. they're a distribution center connected to playlists, connected to sync agents, connected to music supervisors, connected to radio stations, connected to press outlets. they're not magical. but they're connected.
an independent musician, right now, has to build that infrastructure themself. or find a community that has it already.
this is why chicago's scene still matters. why the chicago electronic music conference matters. why knowing the right promoters at green mill or metro or empty bottle matters. why getting booked at certain venues instead of others matters. these are all infrastructure. these are all the actual things that move music.
an ai video tool doesn't connect you to those things. it only makes your song look more finished when you try.
The New Game: What Changes When Excuses Disappear
so if the video excuse is dead... what actually matters now?
first: the song has to be good enough that people listen past thirty seconds. spotify's algorithm weighs that heavily. a beautiful video doesn't save a mediocre track.
second: you need strategy around release timing, playlist pitching, and community building. this is the boring work. nobody wants to do it. it's also the only thing that actually moves streams.
third: you need to understand who your actual audience is. not "everyone." not "people who like experimental hip-hop." but like... the specific three thousand people on earth who would listen to your music on purpose.
fourth: you need to think about revenue differently. streaming is not your path forward unless you're already large. sync is. direct fan support is. live music is. merchandise is. teaching is. building in public is. multiple revenue streams connected to one audience.
none of these things changed because of ai. they were always true. the ai video tool just removed the excuse to delay dealing with them.
marcus is going to release his song with a generated video. it's going to look good. it's going to be finished. and then... he's going to hit the actual wall. the discovery wall. the visibility wall. the "how do i reach people" wall.
that wall doesn't care about his video.
Where This Actually Leads
here's what i think happens in the next two years.
more musicians release finished-looking music with AI-generated videos. the visual bar rises. playlists get more crowded. sync agents get more submissions. the signal-to-noise ratio gets worse. and the actual differentiator becomes: is this person connected? is this person building something? is this person part of a community that matters?
in other words... we're going to trend back toward what actually worked before streaming tried to convince us it didn't matter.
relationships. reputation. consistency. community. the things that can't be generated or automated or solved with sixty dollars.
the ai tool was useful. but it was useful in the way a nice camera is useful. it's necessary but not sufficient. the song was always the thing. discovery was always the thing. connection was always the thing.
we just had more excuses to avoid it when the production was expensive.
now the excuses are gone.
the work begins.