The List Is the Last Thing You Own

Two numbers explain where dance music is right now.

The UK lost 141 festivals between the start of 2024 and the middle of 2026. Around 800 late-night venues have closed since 2020, roughly a quarter of the country's nightlife. That's one end of a dance career disappearing.

At the other end, Deezer reported in April 2026 that it was receiving about 75,000 fully AI-generated tracks a day. Close to half of everything uploaded. By June that figure had passed 90,000. The company also said up to 85% of streams on fully AI tracks were fraudulent.

The stage is shrinking and the feed is filling with noise. Both things are happening to the same artist at the same time. Whatever we think about direct-to-fan as an ideology, it has stopped being an ideology. It's now the only part of the business still reliably paying the middle.

The math nobody bothers defending

Spotify pays somewhere between $0.003 and $0.005 per stream. We need roughly three million streams a year to clear a minimum wage salary from it. That number has been quoted so often it's lost its power to shock, so here's the version that should: since 2024, any track that fails to reach 1,000 streams in twelve months earns nothing at all. Not a fraction of a penny. Zero, redistributed upward to the tracks that cleared the bar.

Deezer found that only 2% of uploaders had more than 1,000 monthly unique listeners.

We don't think this was designed as an attack on independent artists. It looks more like housekeeping. But the effect is the same, and it's worth naming plainly: streaming has formally stopped pretending to be a revenue model for most of the people uploading to it. It's a distribution channel and a discovery surface. That's it.

Once we accept that, a lot of confusion clears up. We stop asking how to make streaming work and start asking what streaming is for.

We got here first

Here's the part the wider industry keeps missing. Electronic music has been running on direct income for twenty years. We just didn't have a buzzword for it.

DJ fees. Beatport sales. Vinyl and cassette runs. Sample packs. Splice royalties, which paid producers $15 million in 2020 alone. Production tutorials, presets, sound design courses. Bandcamp as a shopfront, a catalogue and a mailing list rolled into one.

None of that came from a strategy deck. It came from a genre where the recording was always the marketing and the money was always somewhere else. Dance producers have spent two decades building small, weird, high-value businesses around a few thousand people who care a lot.

That's the model everyone else is now scrambling to invent. The IMS Business Report put the global electronic music industry at $15.1 billion in 2025, a record, with merchandise and creator tools among the fastest-growing lines. Not streaming. The stuff that sits directly between an artist and a person who wants to give them money.

What it actually looks like

Fred again.. is the obvious case, and the interesting detail isn't the arena numbers. It's that before playing to 77,500 people at the LA Coliseum he threw a party for 150 fans he'd identified personally. He built the top of that pyramid by working the bottom of it, hosting listening parties for his most engaged Discord members across eighteen cities.

Skrillex, ahead of his 2025 Ultra return, told people to leave him their email addresses. Then he emailed the list a Dropbox link to 34 new tracks the night before they hit streaming. His fans got the music first because they'd given him a way to reach them.

Lane 8 has been taping over phone cameras at This Never Happened shows since 2016. Less reach, more depth. It's worked for a decade.

And Bandcamp, for all its problems, has paid artists $1.64 billion cumulatively. Bandcamp Fridays alone have moved $154 million since 2020, nineteen of it last year.

Different scales, same mechanic. Every one of these artists has a way to reach their audience that doesn't route through an algorithm.

Now the uncomfortable part

Goldman Sachs prices the "superfan" opportunity at $4.5 billion. Luminate says 20% of US listeners are superfans who spend roughly double on physical and 66% more on live.

When numbers like that appear in a bank's report, they're not there to inspire independent artists. Spotify has a premium tier in development, reported at around $6 a month on top of Premium. Universal's Boyd Muir has suggested 20 to 30% of subscribers might eventually take it. The majors are building fan apps. Live Nation is spending on venues.

So the same word now describes two opposite things. To us, "superfan" means the person who buys the vinyl and the ticket and the hoodie. To a platform, it means an upsell.

Mark Mulligan at MIDiA put the risk well: the majors will probably try to harvest too much value too quickly, and you can't harvest fandom if you aren't also nurturing it.

The thing being fought over here isn't the music. It's the contact details. Whoever holds the direct line to the fan captures the spend. That's the entire game, and it's being played right now, largely without us at the table.

Which brings us to the caution that undercuts our own argument. Bandcamp, the flagship of direct-to-fan, was sold to Epic, then sold on to Songtradr, which cut around half the staff within days, then cut more engineers in 2026. Artists have been pulling their catalogues. The lesson isn't that Bandcamp failed. It's that any platform can be bought and hollowed out while we're standing on it. An email list survives an acquisition. A follower count doesn't.

The objections are good ones

Direct-to-fan overwhelmingly rewards artists who already have an audience. Radiohead could give away In Rainbows precisely because it was Radiohead. Patreon converts existing relationships into recurring revenue, it doesn't manufacture relationships from nothing. Most music Patreons earn under $100 a month.

It's also work. Real work, with churn of 5 to 10% a month, which means standing still requires constant recruitment. Recurring revenue is labour with better scheduling. Anyone selling it as passive income is selling something.

And discovery still runs through the platforms. Nobody finds a new producer on a mailing list. Streaming and social remain the top of the funnel, and pretending otherwise is how artists talk themselves into obscurity while feeling principled about it.

So what's the actual claim

Direct-to-fan isn't replacing streaming. It's replacing the label.

That's the more precise version of the argument, and we think it holds. Distribution is solved and free. Discovery lives on platforms nobody owns. What's left of the traditional label function is capital, relationships, and the ability to turn attention into money. Two of those three are now available to any producer with a mailing list and a Shopify account.

The last piece, turning attention into money, is exactly what direct-to-fan does. And it does it without giving away the masters.

Here's where we land. As AI floods every feed with plausible, frictionless, unattributable music, the scarce thing stops being a good track. Good tracks are about to be infinite. The scarce thing becomes a verifiably human scene: people who know whose record they're buying, who turned up, who are in the room.

You can't put that in a playlist. We can only reach it directly.

So you own the list, or someone else owns the relationship.