The RIAA's year-end report, out in March, put US vinyl revenue at $1.04 billion for 2025, up 9.3 percent. It's the first time vinyl has passed a billion dollars since 2000, and the nineteenth consecutive year of growth. Vinyl shipped 48.5 million units against 29.5 million CDs. Luminate, which counts point of sale rather than shipments, has vinyl at 47.9 million units, exactly half of all US album sales.
The number I find most interesting is the nineteen. This isn't a spike or a trend piece. It's been growing for nearly two decades, through the entire period in which recorded music became functionally free and infinitely portable. The convenient version won completely, and the least convenient version grew every single year alongside it.
Canada is smaller and the data is less granular. Luminate has Canadian physical albums at 3.8 million units for 2025, up three percent, while total album sales fell two percent and on-demand streams grew to 150.8 billion. Vinyl isn't broken out separately, so I can't tell you the Canadian vinyl number and neither can anyone else.
The usual explanation is nostalgia, and I think that's mostly wrong. A meaningful share of vinyl buyers are too young to have owned a record player the first time, and a lot of what sells is new. Nostalgia doesn't explain a nineteen-year curve either. It explains a fad.
The explanation I find more convincing is that the format's inconveniences are the product. A record costs about thirty dollars, holds around forty minutes, has to be turned over halfway through, and can't follow you into the car. Every one of those is a defect if you're optimizing for access. If you're trying to actually listen to something, they're the whole design. You paid, so you're invested. It ends, so you have to decide to continue. You can't skip easily, so you hear the track you would have skipped, which is often the one that turns out to matter.
Streaming removed all of that friction and, along with it, the sequence. Albums were built as arguments with a shape, and the shuffle dissolved the shape into a supply of songs. I'd guess most people under thirty have heard far more music than I had at their age and fewer complete records.
There's a marketing observation in here that I keep circling. We spent twenty years treating friction as the enemy, and the whole discipline of conversion optimization is built on removing steps. That's correct for a transaction. It's not obviously correct for an experience, and we've applied it to both. The vinyl number is a billion dollars a year of people paying a premium for more steps.
The condition is that the friction has to be doing something. Nobody misses waiting on hold. What people are buying with a record is not effort in general but attention made physical, an object that occupies a shelf and requires a decision. The cost is the mechanism.
I don't own many records and I'm not about to argue that the format is superior, which is a conversation I'd rather avoid. What I notice is that the last few times I listened to an entire album, start to finish, without doing anything else, it was because the format made stopping the harder option. That's not a fact about music. It's a fact about attention, and I've stopped being surprised by how much of mine gets decided by whichever choice was slightly easier.
Sources
- 2025 Year-End Music Industry Revenue Report (RIAA, March 2026): riaa.com
- Luminate 2025 Year-End Music Industry Report: luminatedata.com
- Luminate year-end report 2025, Canada (Billboard Canada): ca.billboard.com