The charts are full of old songs, and new artists should be worried

Is the music industry's biggest competitor a rival label, a TikTok trend or a generative AI model? Increasingly, it looks like something much more mundane: Fleetwood Mac's back catalog, and the habits of people who just want to hear what they already love.

The numbers are blunt. According to the BBC, citing Luminate data, streams of current music (released less than 18 months ago) were significantly dwarfed by older music in both the UK and the US in the first half of 2026, with about 60% of the music played released five or more years ago. Eighteen of the year's top 20 bestselling songs came out 12 months earlier or before. Zara Larsson's "Lush Life" (2015) is the year's eighth biggest hit, Michael Jackson's "Billie Jean" sits in 22nd place and Fleetwood Mac's "Dreams" in 30th. In R&B and hip-hop, the two most-played genres, about 60% of streams are catalog hits.

This isn't nostalgia, it's design

The easy reading is that listeners are sentimental. Luminate itself notes that "2026 is the new 2016" has become a dominant social media rallying cry, so nostalgia is real. But nostalgia alone doesn't explain why the pattern is so consistent across genres. We think the better explanation is structural: a service where practically every recorded song is one tap away rewards the safe choice, and the safe choice is the song you already know.

Then come the recommendation systems. As DJ and curator Tim Garcia puts it to the BBC, they're based on everyone's listening behavior and data, so you'll probably get a recommendation very close to what you already play. Qobuz's Dan Mackta says the technology "is never going to want to throw a curve ball at you." Whatever you think of the source (Qobuz sells a human-curated alternative), the logic holds: an algorithm optimized for engagement is a familiarity machine. It doesn't hate new music. It just has no reason to take a risk on it.

The winners keep winning, even the new ones

Even current hits are part of the same dynamic. The BBC points out that Sam Fender and Olivia Dean's "Rein Me In" spent a record 23 non-consecutive weeks at number one in the UK, and that Ella Langley's "Choosin' Texas" became the longest-running US number one in chart history. Those are celebrated as triumphs. We read them as symptoms: when a handful of songs soak up attention for months, the space left for everyone else shrinks. The UK's Competition and Markets Authority, again per the BBC, says nearly two thirds of streams are of music recorded by only the top 0.4% of artists.

For a newcomer, that's a brutal funnel. And going viral doesn't fix it. Some musicians say a hit is hard to follow up, and TikTok is built to create viral moments around a single song, not careers. J Erving, founder of Raye's distribution company Human Re Sources, put the problem well in a Billboard interview quoted by the BBC: people need to care about you beyond the song. A streaming economy that rewards passive listening is hostile to exactly that kind of connection.

Discovery can't be left to the market

Here's where we take a side: platforms and labels that treat discovery as a nice extra are eating their seed corn. Apple Music's Oliver Schusser made the point himself, noting that the performers of recent Super Bowl halftime shows all started somewhere before the biggest stage. If the pipeline of "somewheres" dries up, we suspect there won't be a new catalog worth monetizing decades from now.

There are encouraging experiments, though we wouldn't oversell them. Sofar Sounds is pushing a "Human Discovery Movement" with voice-note recommendations from tastemakers, and an East London rapper, J Casa, argues that live performance builds a lasting connection with a new artist. These are small, human-scale answers to a problem of enormous scale. The harder part is AI-driven noise: Spotify says it removed 75 million "spammy tracks," and independent artist Chiminyo told the BBC he suspects one popular relaxing jazz piano playlist is full of unverified, near-identical acts. Deezer, to be fair, says AI accounts for just 0.5% of its streams, so the real threat for now looks like crowding, not replacement.

Our verdict

Don't read the charts as proof that people stopped caring about new music. Read them as proof that the default settings of streaming favor the familiar. Platforms should build discovery that deliberately risks a curveball, labels should invest in artists beyond the single viral hit, and listeners, yes, you, can do a little work: follow a human curator, go to a small show, press play on an artist you've never heard of. Catalog will always be there. The next catalog won't build itself.