Spotify Earnings, Individualized Networks, AI and Aggregation

# Spotify Earnings, Individualized Networks, AI and Aggregation

Spotify's nature as a content network means that AI is a sustaining technology, particularly because they have the right business model in place. 

Good morning,

A couple of scheduling notes:

  * First, tomorrow’s interview is actually of me; that means the transcript and podcast will be coming out a couple of hours later as it will be cross-published with the other podcast.
  * Second, Monday is a federal holiday and the stock market will be closed; there will be no Update.
  * Third, just to get ahead of things, I will also be taking a vacation day the following Monday, February 23. Next week is also Chinese New Year, and while that is less pertinent to me now that I am in the U.S., I will have family in town and will be getting some sun.

On to the Update:

### Spotify Earnings

From [Bloomberg](<https://www.bloomberg.com/news/articles/2026-02-10/spotify-sets-user-growth-record-thanks-to-wrapped-promotion>):

> Shares of Spotify Technology SA jumped the most in nearly eight years after the Swedish music streaming giant added a record number of users last quarter, far surpassing analysts’ expectations. The world’s biggest streaming service added 38 million new listeners from October through December to reach 751 million, according to a statement Tuesday. Paid premium subscriptions increased 10% to 290 million. Spotify said it expects to reach 759 million monthly active users in the current quarter. The company credited its annual “Wrapped” campaign, which provides users with a personalized, data-driven summary of their listening habits over the past year in a format designed to be shared and compared across social media. Spotify also cited the launch of an enhanced free tier around the world as contributing to the growth.
> 
> The shares rose as much as 19%, their biggest intraday gain since April 2018. After increasing 29% last year, the stock gave up most of those gains heading into the quarterly report as analysts focused on concerns about the impact of artificial intelligence and Spotify’s ability to continue to raise prices.

[This earnings call](<https://seekingalpha.com/article/4868154-spotify-technology-s-a-spot-q4-2025-earnings-call-transcript>) was the last one for Spotify founder and former CEO Daniel Ek, who officially stepped down on December 31, handing off the reins to Gustav Söderström and Alex Norström as co-CEOs. That actually made for a really compelling earnings call that I thought was worth covering in full.

### Individualized Networks

First up was Ek, who only spoke at the beginning, and while he framed his remarks as looking forward, it was really a recounting of how Spotify has succeeded to date:

> If you want a framework for evaluating Spotify going forward and what to hold us accountable to, I’d point to 3 key things. And then you must also layer on the culture that makes them possible. First, we solve problems at the intersection of consumers and creators. This is where we focus. If something is good for the consumer and also good for the creator, that’s where you’ll find us every time.

There is a broader principle here, which is that value is created at intersections, particularly if that intersection is between two separate networks. For creators and labels, Spotify is one interface through which they can reach hundreds of millions of listeners; for consumers, Spotify is one interface through which they can consume effectively all of the audio content in the world.

> Second, we are first and foremost a technology company. We’ve said for years that we aim to be the R&D arm for the music industry. And if I may say so, nearly 20 years in, I think we’ve earned that. We drove the shift from downloads to streaming and subscription, and we proved the model could work at scale.
> 
> But here’s what excites me the most. Our capabilities now extend far beyond music. Today, what we built is a technology platform for audio and increasingly for all the ways creators connect with audiences. And this identity will matter even more going forward. The next wave of technology shifts, AI, new interfaces, wearables, new ways of interacting with content, these will reshape how people discover and experience audio and media. The hard problems I had in music and podcast and books and video, in live and in things we haven’t even built yet, we’re going to keep building the technology to solve them.

This is the yin to the intersection yang: from the perspective of a creator or a consumer, they are interacting with Spotify the large audio service; it’s a singular entity. The trick for Spotify, however, is that they need to interact with every individual creator and especially every individual consumer as a unique individual: everyone listens to different music, needs different recommendations, etc. That is where the tech comes in: it enables the intersection to appear singular from the outside, even as every single connection from the inside is unique.

Note that this isn’t unique to Spotify: this is the exact same trick that every successful network business pulls off. Everyone might be on Facebook or Instagram or TikTok, but the actual reality is that every single person’s experience of those products is completely unique and individualized; if it were the same for everyone there would, in the end, be fewer users because of much lower retention.

> Third, we play the long game. When we went public in 2018, I talked about long-term value creation. While I know many of you focus quarter-to-quarter, that’s not how we grade ourselves, and it’s never have been. We chose growth over profitability for many years. And I know that was painful for some of you, but in order to scale, it was the right thing for consumers and creators and ultimately, for the business we’re running today.
> 
> We acquired Echo Nest back in 2014 when most people didn’t understand why a streaming company needed a machine learning AI company. And that bet gave us personalization, something that’s now core to everything we do. We built our Ubiquity Play that’s called Spotify Connect starting in 2011, right as we launched in the U.S. At the time, every major tech platform was building their own Walled Garden for audio. The conventional wisdom was pick an ecosystem and live inside it. We bet the other way. We decided Spotify should work everywhere in your car, your speaker, your TV, your gaming console, regardless of whose ecosystem you’re in, Apples, Googles, Amazon, Samsung, Sonos, all of them seamlessly. And today, Spotify works across more than 2,000 devices from over 200 brands. And you can start a song on your phone and you can finish it on your TV. That doesn’t happen by accident. It happens because we choose ubiquity over control, openness over lock-in, and we stuck with it for over a decade. These weren’t obvious calls at the time, but they compound and that long-term orientation will continue to guide Spotify.

In fact, these were obvious calls, if you truly internalize my second point above: a network that truly scales must be everywhere and meet people where they are. To be large is to be everything to everyone, all at the same time, and that is only something that is possible because of technology.

### AI and Aggregation

It fell to Söderström to make the case as to why AI was a boon for Spotify, and a lot of the reasons rested on the foundation Ek laid out above. First, however, Söderström made a very astute point that I completely agree with: real disruption is about business models:

> My view is that new technology is seldom disruptive on its own. Significant disruption happens when new technologies enable new asymmetric business models. For example, this is what Spotify did to music downloads. This is what Uber did to taxi service.
> 
> So the question everyone should be asking is, does this evolution create new business models? Or are we mostly just seeing new technologies? For example, in SaaS, there is currently a lot of fear that the perceived business model will be challenged by more outcome-based models, which is reasonable. However, in the consumer space that we are in, we believe the dominant business model will continue to be ads plus subscription, both places where Spotify excels. This puts Spotify in an outstanding position because we already have the right business model. Our job then just becomes leveraging these new technologies to our benefit, which is something that we’ve done consistently for the last 18 years.

I wrote about the [Unified Content Business Model](<https://stratechery.com/2023/the-unified-content-business-model/>) back in 2023, making the case that the fundamental zero marginal cost nature of content, which costs a lot to create but is then infinitely copyable and reusable, leads inexorably to some combination of an advertising and subscription model.

People have made the case that AI is different, given the marginal cost of inference, but I think that is only a temporary state of affairs: for scaled content services the marginal cost of inference will approach zero, whether that be through some offloading to local models or cloud models becoming “good enough” that costs approach that of normal compute. What will remain very expensive is the training of the models themselves, but that’s a fixed cost: the key will be to have a scaled customer base over which you can leverage those costs, which means you end up with the unified content business model.

So what sort of AI might Spotify want to invest in, then? Well, that depends on what side of the market you are considering. Start with the consumer side: I just made the case above that the reality of modern scaled networks is individualized experiences, and that’s exactly what AI can provide through two vectors. The first is giving individuals more control over what they hear, which Spotify is investing in. Söderström said:

> Back in 2021, we saw the potential of AI that would be able to think and speak at the level of human. So we acquired AI voice platform, Symantec in 2022. And this put us on an early path to introduce agentic experiences to Spotify users. One example of this is the wildly popular Interactive DJ, which we introduced in 2023 and have continued to enhance since then. About 90 million subscribers have used IDJ so far, driving over 4 billion hours of time spent on Spotify, and this keeps growing.
> 
> More recently, we also launched Prompted Playlist, a new tool that has instantly taken off with power users. So if Interactive DJ is the chat interface to Spotify, where you can talk casually, Prompted Playlist is the deep research mode of Spotify. It lets you describe and set rules for your own personalized playlists, literally writing your own algorithm. It taps into your entire Spotify listening history, reflecting not just current obsessions, but the full arc of your music taste and integrates up to the minute culture pulled from the Internet.
> 
> There is nothing else like it. So all of this teases the next evolution of Spotify, delivering the world’s most intelligent agentic media platform, one that you can literally talk to that fully understands each individual listener and puts them in the driver’s seat. It’s about moving from a passive experience to an interactive one. This is a stark contrast to most media services today.

Still, only some users are going to do this work; what is more compelling is the second impact, which is using AI to reason across the vast amount of data that Spotify can capture from all of those individualized experiences to further hone the individualized experience for everyone, even if they never actively use Spotify’s AI tools.

> I’m excited about us being the first truly intelligent agentic media service that you can literally talk to. And this is not just a pipe dream. You can already talk to Spotify through the AI DJ casually, but also through prompt the playlist in sort of a deep research way. We’re going to keep investing in that. What that means structurally for Spotify is that we are building a dataset that never existed, which is the data set of language to music, language to podcast and language to books. We’ve had the song-to-song dataset, but no one had the language to song dataset.
> 
> I want to drive on a point here, which is this is a very specific dataset. You may think it is a canonical dataset, meaning there is a factual answer to, for example, what is workout music. There is no factual answer to what is workout music…You can’t just have an LLM commoditize it as a fact, the way you can commoditize Wikipedia. You actually need to have many, many hundreds of millions of listeners across the world’s market constantly telling you what it means for that specific person. This is a dataset that we are building right now that no one else is really building. It doesn’t exist at this scale. And we see it improving every time we retrain our models.

On the creator side, Spotify doesn’t necessarily need to invest in AI to make music; it’s not like Spotify buys the guitars or recording equipment for musicians either. However, what Spotify does need to be ready for is a flood of AI-generated music. This, however, is exactly where those individualization algorithms pay off: the entire point of Spotify is to help individual users manage abundance; they already have access to nearly all music ever made, so adding a bunch more slop doesn’t meaningfully change the problem space.

The people who are harmed by the explosion in supply are the other creators, who now face more competition. This, by extension, actually increases Spotify’s power over supply, because they get to decide what is a hit. Söderström again:

> In terms of net new music, there are tons of companies that allow you to create music using AI. But that’s not where the music breaks. That music, if it breaks on Spotify. That’s where charts, that’s where the culture moment is. So we feel very comfortable about that position.

In short, Spotify believes that AI is a sustaining technology for their business, and will make their position as an Aggregator even stronger.

> There is this fear that software companies are not going to exist anymore, everyone rolls their own products. I certainly don’t think that’s going to be true for consumer products. I think what will happen is something more like what happened with the Internet. When the Internet came along, everyone thought that we would all have our own web pages. What actually happened was there ended up being very few web pages. In times of lower friction, things actually tend to aggregate, not disaggregate. That’s the opportunity we see in front of us. I think companies such as us are simply going to produce massively more software. Up until our limiting factor is actually the amount of change that consumers are comfortable with.

Spotify is not without its challenges, specifically YouTube; I buy the argument, however, that AI is not one of them — indeed, it is in fact an opportunity.

* * *

This Update will be available as a podcast later today. To receive it in your podcast player, visit Stratechery.

The Stratechery Update is intended for a single recipient, but occasional forwarding is totally fine! If you would like to order multiple subscriptions for your team with a group discount (minimum 5), please contact me directly.

Thanks for being a subscriber, and have a great day!