Spotify Technology S.A. (SPOT) September 9, 2026 Conference Transcript

Okay, so if everyone can find their seats, we're going to go on to our next fireside chat. Good afternoon, everyone. It's my pleasure to welcome Gustav Soderstrom, co-CEO of Spotify, to our Communicopia and Technology Conference, his first investor conference in that role. For those of you who haven't met him before, Gustav is an engineer by training and a founder by instinct. He built ConnectWorks, which was acquired by Yahoo in 2006, and 13th Lab, an augmented reality startup acquired by Oculus. He joined Spotify in 2009, spent the next 15 years as chief product and technology officer. In that time, he called the mobile shift early, bet on machine learning more than a decade ago, and helped transform a music app into the place where nearly 800 million people listen and watch podcasts and books as well. He became co-president in 2023 and stepped into the co-CEO role alongside Alex Nordstrom this January. Gustav, welcome to the stage at Communicopia.

Thank you. Thank you. Appreciate it.

I always love welcoming people to their first Communicopean Technology Conference. Thanks so much for taking the time to do it. I really appreciate it.

Thank you for having me. It's exciting.

Let's start with the future and the technology shifts that are underway. You've been a key architect of Spotify's evolution through a number of technology shifts. Drawing on that experience, how are you thinking about Spotify positioning itself for where the industry landscape is going over the medium to long term?

The way I think about technology shifts, often called macrowaves, that's what we talk about internally, is that in times of no change, market shares tend to stay stable. In times of change is where you have opportunity. So Spotify was born out of one of these opportunities. It was the shift to low latency, high bandwidth broadband. This is why Spotify could start in Sweden, because Sweden, fortunately for us, happened to have that before everyone else. It was one of these technology shifts that enabled Spotify. And then Spotify grew. We had our desktop and mobile proposition. And then the smartphone happened. And it was a big change. It was scary. But the outcome was that Spotify figured out how to reposition itself for this age. And once we did, we started growing very quickly. Then the connected home happened. We figured out, we'd connect, and we started riding that macro wave or surfing the wave, riding the tailwind. So this is how I think about AI and where we are right now. The job for me and Alex, the other co-CEO, is to be able to understand technology deeply, try to understand what it means for Spotify for the business model, reposition the company to turn it into a tailwind, and then write it. So, you know, I always tell everyone, I've been at Spotify for almost 18 years now, and I haven't had this much fun since the smartphone, because now everything is changing. So now there's opportunity to do something, to get bigger, to raise your ambition, and to increase market issues.

Well, let's build on that last answer, because you guys as a team hosted an Investor Day just a couple months ago in May. You laid out a lot of key strategic priorities that are anchoring the business going forward.

How would you frame those key priorities and how you're thinking about the execution roadmap ahead for the companies when measured against those priorities? and so Alex and I we obviously you know we've seen this AI wave coming but everyone sees it coming the interesting question is what does it mean and what we saw happening we saw a few things happening a few years ago and we had a few answers to what it would mean one thing that would that it would mean would be that for the first time in history computers would start to speak English so there was about 1 million people globally who could speak to computers, they were called developers on GitHub. It has been evident for some time that anyone will be able to instruct the computer, and today you can do that. So we realized that all 8 billion of us are going to be able to talk to computers. So the question then is, what does that mean for consumer services? So we started rebuilding Spotify to be the first truly intelligent media service. You know, we could see consumers sitting there with chat GPT over here, you know, basically AGI. And then over here with the media service that at least a few years ago was Dumb as a Rock. That's not going to last. Of course you're going to want to be able to talk to your media service the way you could talk to ChatGPT. when a consumer heard a song on Spotify, if you go back a few years, they might say, this is a great song. Now I want to see the music video. I have to go to YouTube to do that. That's not great for us. Today you have all the music videos on Spotify. But you may also say, who is this what is this band? Who is it? When are they touring? All of those are questions where we sent them away. So now we rebuild Spotify. So you ask all of those questions to Spotify right there and it tells you who the band is, who the drummer was, who produced the song, when they're touring, where they're touring. It has both all of our proprietary information that we've licensed or bought even who you sample this song from. But it also has access to all the world information of what people are doing right now. So we create a new vision for the consumer product and we think we were early with that and we think right now we're quite far ahead in the capabilities that Spotify has and what you can ask it to do so that was one change the other thing that happened was that because more and more features would start using AI and inference the internet would start going from what has for the longest time been a fixed cost amortization business into a marginal cost business what I mean with that is since before software since Silicon Silicon Valley has been about fixed-cost upfront investment amortization, and then after Silicon Software had the same dynamics, this is now changing. Now you're starting to have marginal cost per user in terms of inference. Fortunately, so we looked at this and said this could be a challenge and a headwind for a company like Spotify. Let's turn it into a tailwind. Fortunately for Spotify, we're one of the few companies that always had marginal cost. We've always had per stream fees in the free tier. If we had grown too quickly through our journey, we could have gone bankrupt. So we were never fortunate enough to just let growth run wild and then monetize later. We had to grow and monetize. So we developed a lot of discipline in this genre over the years. So Alex and I looked at this, and we said, how can we turn this into an opportunity? But yesterday, in May, we presented this power law. And it's been true forever that usage of services is not a linear curve. It's a power law. We have a long tail of users that use the service a little, and you have people who use it more and more. But then you have a head who do not use it twice as much. They use it literally 10 times as much. Now, if you think about having marginal cost, in a fixed cost environment, those 10x users, they don't matter that much. It was just the upfront software cost. If you have marginal cost, now these 10x users will cost you 10x much. So you have to have a business model that supports this, or you're just going to have to cut them off and ship them off to service. so we had started another business a few years ago called audiobooks and audiobooks has that same dynamic we have some amount of cost per listening hour so what we did was we put a large amount about 15 hours of audiobooks listening into premium an amazing offer the average audiobook is only seven hours so for most consumers this is all they have they can listen to in a month this was very well received that the audio books businesses is going like crazy right now but you have these power users who did not want 15 hours they wanted a hundred hours so we created add-ons for them where you can buy additional hours we were very hopeful that this would work well but it worked much better than we thought this add-on is already on a hundred million error after only about a year so they're encouraged by this we said this is the model for inference-based products as well. We put a bunch of inference into the free tier. We put more inference into the paid tier, but we control how much. We control our margin. The margin is a managed outcome. And then if you want more Spotify, instead of saying, no, go away, go somewhere else, go to a competitor, we say, sure, you can pay for more. You can pay for as much more as you want. So this is how we think about our business now. We have a free tier that maximizes the opportunity, right? I think all subscription businesses eventually are going have a free tier we were just early to that if you want to maximize the opportunity we think the time for music is literally 8 billion people if you want to maximize that you need a free tier then we have one of the world's largest subscription tiers which lets us use bundle economics to enter new areas audiobooks etc but what we didn't have until the spring was the ability to also monetize the rest of this power law the head which we now do with add-ons with audiobooks with inference add-ons and several other add-ons that are coming along the way. So we feel that this is the right business model for a modern company in the age of AI. You need to let users buy as much as they want of your service.

Okay. So against that, you know, probably a question I get a fair bit from investors as well is how do you see your own competitive positioning in the audio landscape? And when you think about some of the growth opportunities you laid out at Investor Day, how do you think you could possibly change the positioning you have competitively and build even larger amounts of scale around the business, mostly tied to engagement is what I thought the messaging was coming out of Investor Day.

Yeah, so there are a few different dimensions that we can compete on. As I said, the business model itself is one competitive advantage, the freemium business model that we've now added add-ons to. So we can both play with growth, we can play with price rates of the premium tier, but we can also play with finding niche audiences that have much higher ARPU, like audiobook listeners, heavy AI users, other things that are coming up. So this is one dimension. The other dimension that we've been playing with is to add more verticals. We went into podcasts because we saw that there was a lot of engagement there, but no one was really innovating on product. We went in there, and we became market leaders. Then we saw audiobooks, which we also thought was a great fit for Spotify. And if you look at the Scandinavian countries where I come from, audiobooks is not a niche feature. It's a math feature, which is why we put it into premium. So we are going to go into more and more verticals. And one of the things that we think is unique about Spotify versus many other consumer apps is that most consumer apps, you take a video app, that's a use case. It's a situation. Spotify is not a use case. Spotify is your studying, your running, your dinner with friends, your car commute. We're everywhere. So we have so many use cases to go into. One that I'm very passionate about, for example, is fitness. Almost everyone uses Spotify when they're running and when they're in the gym. So we think we have the brand promise to go into many more areas than other brands. So many people look at Spotify and say, like, okay, maybe they have three verticals, music, podcasts, and audiobooks, but how many more can there be? We think there are many. We think our brand is strong enough and has enough elasticity for a lot of areas, and we think we're already in those use cases. The last thing I would say is, you know, Alex and I, we look at others, we try to learn from the best, but you would like to have your own strategy, right? You would like to be differentiated. So one of the bets that we made that we talked a lot about at Investor Day was that it's clear that short-form videos is eating enormous amounts of engagement. It's very effective. It monetizes advertising extremely well. You have an ad break like every 30 seconds. It's also very addictive. And it's also true that most of the consumers regret most of the time they spend on these services. This is not something I made up. What we did was we asked someone to survey our user base, 50% our user base, 50% non-Spotify user base, They asked across all the big surveys, all of them, Spotify, Apple, YouTube, TikTok, all of them, two questions. How much of the time you spend do you regret, and how much of the time you spend do you value? Which is the same question as two ways, just to control for the answer. And what turned out in this survey, which we talked about, was that Spotify was the lowest regret content on the Internet, at least in this group. So people regret less than 10% of the time they spend on Spotify. And conversely, they value more than 90%. This is not true for most of the other services. Some of them people regret up to 70% of the time they spend. This was an insight for us. I previously thought that, you know, I have kids on my own. When they sit there and scroll and scroll and scroll, my view was that they're probably trapped and they love it. Turns out, no, they're trapped and they know it and they don't like it. They hate it. So we decided to turn this into a strategy. because we were the lowest regret we said we are going to be the lowest regret or the most the most valuable time spent on the internet why is that a good strategy for spotify it is because spotify still sort of 90 subscription revenue if you are if we were 90 advertising revenue going after max engagement would probably have been the best strategy and it probably is for some of those companies but if you're not subscription you should maximize for subscription retention so the question is at the end of the month when a user decides that they want to pay again when they vote with a wallet what do they ask themselves do they ask themselves like where was my max engagement where did I scroll the longest no they don't they ask themselves do I regret the time I spent or do I feel good about it so that's what we're trying to maximize and I think that's differentiated both subjectively, because I haven't seen anyone else say it, but also structurally, because if most of your monetization is from ads, it's going to be very expensive to follow up. Does that make sense?

It does indeed, yeah. Really interesting. Let's come back to the analyst day for a second, because you see the investor day. You threw out some targets around where you want to go over the next couple of years out to 2030, thinking about mid-teams growth, gross margin ranges, and 20% or better operating When you guys think about all that you want to accomplish over the next four or five years, how do you think about striking the right balance between investing in growth, making sure you capture the growth opportunity that sits in front of you, but also delivering on that margin trajectory?

Yeah, so I think what is nice about Spotify, I'm sure you've heard a little in the effect, what has happened before is very likely to happen more. So Spotify has almost 20 years of the Lindy effect. And what I like about this is sometimes when I talk to investors, I say the following. Actually, I'm sort of paraphrasing Alex Nordstrom here. We have something like between 3% and 4% of the world's population paying us every month today for subscriptions. If you look at a market like Sweden, 50% of the population are paying for Spotify. And that includes babies and the elderly. You could imagine that the conversion percentage is very high, and it doesn't stop. It just keeps going up. We have never seen it stop. Conversion has never stalled out in any market, right? So then you say to yourself, maybe Spotify won't convert 70%, 80%, 90% of the world to premium, but is it unreasonable that we would convert 15% of the world? Then you have a 5x story already. So then the question is, what do you have to believe to believe in a 5x story the truth is just more of the same so Spotify started in Sweden there were zero paid users it's gone from 0% to very very high percentages but this is true for every market conversion is a functional time the longer you stay in the market the longer conversion goes up so Spotify has something like 40% conversion on average but that average doesn't exist it's a blend of super high conversion will be been for a long time lower conversion were nascent and so for the longest time people said okay I buy that Sweden converts well but Swedes are different and then it worked in Scandinavia and they said well I buy the Scandinavians are different but it will never work in Europe and then it worked in Europe and people said but never in the UK right it's a different market guess what worked in the UK and then we came to America and people said well not America surely it's the world's biggest music market same story and then they said okay but never ever in Latin America right cannot be guess what exact same story we're converting like crazy in Latin America and now they're saying like but what about India can't be true right of course it's going to be true and it's going to be true for Africa as well so really what you have to believe is just more of the same which I think is a very good base now Alex and I are not content with saying like we're just going to do the same thing for another 20 years. But it is important to know that you don't have to believe more than just the same for a few more years for Spotify to get, and the music industry, to get many X bigger. Now, on top of that, we are going to do a lot more. As I said, we're going into more and more verticals. We are trying to become something very different for consumers, a truly intelligent media service. We are going from, what we said at Investor is we're going from passive single-player services interactive multiplayer services what do we mean with that well multiplayer means we have a feature called jam there's over 50 million users where you play music together to you know in the same room but also remotely this means that if you're in college it it matters which music service you have you have to be on Spotify or you can't be at the party so we think multiplayer network effects are very important especially at our scale but we also think interactivity is very important so Spotify started as a pretty passive consumption service the thing you did though was you curated playlists that was very valuable for Spotify users helped tell us like these songs go well together in the playlist that was the basis for machine learning what they're doing now since a year back since we had a mixing is they're not saying just which tracks go together in a playlist they're also saying how they transition well exactly in which millisecond so if you listen to our automatic mixes a year ago they were okay but not great if you listen to them now they're really good why is that it's because millions of people have told us exactly how these songs go together so they're curating the catalog for us now why are they doing that not because they're working for us it's because they love to interact with the music so this is why we're so confident about the thing I'm sure going to ask about remixes and covers, using AI to be able to play much more with the music, to take your favorite song and transform it, not just beat, stretch it, and do a nice DJ transition, but actually change the whole song, do a remix. So we're betting a lot on interactivity. These are some of the questions of how we're planning to grow. I think the baseline is incredibly strong, but the ambition is much higher than baseline.

Okay. That's super interesting. And I do want to come back and go a little bit deeper in some of those topics, but let's end this part of the conversation by you talked about the way the consumer thinks about your product. You talked about layering in more engagement, more utility, and obviously the conversion theme that we talked a lot about at Investor Day. How do you think about the pricing dynamic around the product longer term? If you're delivering more utility to the consumer, how do you think about pricing elasticity?

Yeah, so for the longest time, Spotify held the price at sort of $9.99. and actually if you then come for inflation we actually lower the price and with family plans effective price actually went down over time at the same time we just poured value into there more machine learning the catalog went from I think 2 million tracks to 200 million tracks like more and more value kept the price the same so the price the value gap increased and then a few years ago we started taking back some of that price capturing some of that value but the rule that Alex and I have is the value always has to be vastly bigger than the price you're paying. This should be a no-brainer. This should be the best proposition on the internet. But we've not proven several times that we have more value. This is the question, like how much prices do you have? The answer is what happens when you price it? Do you get churn or not? We do not get churn. We've been very comfortable in these price races. The second test is do you dare to price lead? What happens if you're more expensive than your competitors? We're more expensive than our competitors in all markets, and we still don't see the churn. We feel very good about the value we have there, but we're not complacent about it. So we have literally a strategy called value in premium, where we're trying to make sure that we always put enormous amounts of value in premiums. All these mixing features I told you about, free in premium. You don't have to pay for those, right? So there are a few levers. One is pricing of the premium tier, where we think we have more room to grow already, but we also keep putting in more value, so that gives us more room to grow. And then, of course, we have the add-ons, where we think we could have much higher, and we do have much higher ARPU for some users. So we're going to create more of these add-ons in music, in podcasts, in audiobooks, and in other verticals.

Very clear. Okay. When you think about the opportunity set you're laying out for AI to transform the industry, How do you think about what you probably want to do, which is being innovative and moving the platform forward, but also recognizing that there's an industry structure, there's relationships in this industry, and bringing the industry along and making sure that the relationships with the industry are fostered and grow alongside while you're trying to build an innovation curve?

Yeah, so the way I think about AI is in three different buckets, right? if you think about Spotify and AI there is productivity but we could either be the most productive or we could fall behind and be non-productive and non competitive there's productivity then there's a product proposition where I just said I think consumer products will change drastically in the next few years and we're held them from being the ones that change it and then there is content generative content and music so if we actually start with generative media and and and content which I think is you as investors probably think a lot about because if there was one thing that that could be dangerous for companies like Spotify should be if the core economics of of content somehow shift and and there are you know hundreds of millions and billions of tracks being created on these generative services so what's happening today is all music generation is only about net new music you're only making new music on these on these services out there not on Spotify we don't do that but on these other services most of these services are sued by the label you're making new music there and that is because it's not because they can't make a remix or a cover it's because the remix and the covers are copyright protected so they can't legally right so you're making a new song and and you know Spotify has always had the view that for a creator we do not want to decide we should not decide what tools they're allowed to use I think if you look historically there was a big upheaval when the electric guitar came along, right? It was cheating or different, and then we got rock music and then in the 80s the synthesizer came along and clearly people were cheating now but then we had synth music then, you know, close to my heart in the 90s Avicii came along and because he couldn't play any instrument or sing establishment said that's not a real artist, right? Turns out one of the artists we're the most proud of now. So I think of generative music models as tools and it is a fact that most creators are using these tools to some extent at least in ideation today so then the question is what does that mean for Spotify well to the extent that people are using these tools and creating songs the way for them to get distribution monetization is to upload them to Spotify so so we are we are participating in if there's a hit made it's very likely to be on Spotify so in that sense for an aggregator the fact that the catalog is bigger is a positive but what we think is is wrong with this scenario is that AI is not only a story for for new creators and for the existing artists it's only a store of replacement from here on so what we decided to do is we're saying that no one is solving this the other problem of making remixes and covers for existing IP because it would be too hard for them they would have to get every artist on earth signed up and start paying out to them guess who has every artists on there signed up and it's paying out to them already Spotify so we are in the perfect position to solve this problem how all the existing artists participate in AI instead of getting replaced by AI so we have taken the laborious path of training our own generative music model legally asking for permission slowly and boringly instead of just running ahead and doing them illegally it's slower but we think it's the right thing to do it's just that's what we did against piracy you know we didn't cheat with piracy we actually went and licensed all the music many labels were hesitant artists were hesitant but it was the right thing to do so we're taking the long path training our own models which i feel very good about because we've been working on this for some time but the interesting thing is and then we go and get opt-in from the artists who want to participate we are not forcing artists to participate instead what we're saying is you probably want to participate because if if one of your fans creates a remix of your song that actually increases the share of the streaming pool that you get so if you're an artist today you know from the 80s or 90s the way that you would make more money today is to hope that some other artists would create a remix or cover of your song because then you get more royalties now you can hope that all your fans actually create remixes and covers of your songs so the economic incentives are aligned with the artists and we're not taking anything of that all those royalties are going to the artist so even more than if someone made a cover because then the artist would have had to share the revenue with a cover maker you get all of it the the the fan doesn't get anything we don't take you just get all of it so this the incentives are aligned for artists but we're not forcing them we're asking them to opt in so we think this is the way it's just like with piracy the way forward is to also make a

great product not try to stop it but to make a better product that is legal so this is how we're thinking about it we're very excited about it we think all existing artists deserve to participate in AI if they want to but we're not going to force them just just following up on that you know what signs do you think investors should continue to watch for to see you've had some good industry adoption and agreements struck already more of the industry coming along for this as a broader theme and then broadly how it also fits into commercialization longer term as well yeah so we've we've had UMG cobalt modeling sign up we've and we're making really good progress with them the thing

that has been the most important to me is actually that the tools are really good so I've been working very hard on the models themselves I feel very good about where they are right now the next step for us will be a research preview Why would we do a research preview? Well, it is because the way these models work is you can train them in private, but then the way they get better is by lots and lots of examples, right? So what we want to do is we want to ask our users, you know, for a prompt, for a song, then they create a remix, and then we show them two versions, and they say this one was better than that one, and then this is called preference data, and then you can do what is called reinforcement learning on this preference data. this is how you win this is why I'm tropic ran away because they had more uses given in preference data this is why some of these well you have one player that is the best of new music because they got the most users given the most preference data the benefits of Spotify is we have the distribution we have you know almost 800 million users who could give us preference data so I think structurally we're both the ones that are best positioned to do this on the payout side but we're also one of the few companies who would get the preference data with this model to become the best in the world very quickly. So that's the next step. It's already very good, but getting preference data from lots of uses is what is going to improve it quickly.

Yeah, understood. Scale beginning or scale. Certainly understood. You talked earlier about some of the format innovation that's happened on the platform over the last couple of years. Two-parter for me, really.

What have you learned the most as you've introduced a wider range of formats on the platform and specifically what have you learned as more video has become prevalent on the platform and how that changes user behavior yeah if we start with video so there were two things that happened for us one was you know we went into podcast when they were audio but fortunately for us we did if you remember the deal with Joe Rogan and he said I need to have video so we built video because we needed to but that gave us a window in the consumer behavior and we saw what was happening that you know consumers are actually watching this video and dipping in and out of it even though most of the consumption is when the phone is in your pocket and then the rest of the industry of the podcast industry just started becoming video very quickly but then we because we were early to that and we saw it we pivoted and we built out podcast video very quickly and now we have a very good podcast experience which is why it's really only us and YouTube that are in this game because the others missed video if we're going to be honest about it so if you look at video you see two things you see a very clear split between audiences where older people still listen to all the only podcast and younger people if there's no video they contact customer support and say the podcast is broken I don't understand why there isn't any images there it's just so they don't have the concept of all the only anymore yeah right So that's on the podcast side. And we were fortunate enough to act on it technically. But then we actually had one more stumbling or one more barrier, which was the monetization. So the monetization was much better if you were audio only than if you were video. Without going into detail, DAI, dynamic ad insertion, was its whole own ecosystem. It doesn't work for video. So we had this problem where technically we could have video, but the creators were not incentivized to. So back to incentives, we launched something called SPP, Spotify Partner Program, in January, about a year and a half ago, where in premium, we actually removed the Spotify ads and we started paying out from the premium pool instead, which meant that the podcast product got better for consumer, fewer ad breaks, which increased retention and listening for the creator, but the creator still got paid magically. So the product got better because it had fewer ads, and the creator got paid. So this meant that all of a sudden all creators started uploading video. And since then we've been growing. I think video has grown 50% year over year on Spotify. We have over 500,000 shows now. So that's on the podcast side. On the music side, I just talked about us sending our consumers to YouTube a few years ago to see what the artist actually looked like. That never made any sense, but it took some time for us to build out the video stack. So now we have all of the music videos, all the official music videos, But we're also recently getting artists to upload their unofficial music videos that they record themselves. So we have this catalog now, and it's very interesting because in the data, we see exactly what we thought. Like music video is incredibly important the first time you hear a song, and then it's not that important anymore because it's the same video. When you listen in your pocket, you know, the 100th stream of the same song, the video isn't important. But for discovery, it's incredibly important. So actually, I think a song is like 24% more likely to be streamed or saved if it has a video versus not. So it's incredibly impactful for discovery. And Spotify is obviously all about discovery. And we know that discovery directly correlates to retention. So this is why we're betting so hard on video. It's been a huge lift for us over many years, both building and licensing. But now it's there. Unfortunately, we're seeing exactly what we hoped for.

I'm going to try to squeeze two in before I lose you. At Investor Day, you talked about getting to 1 billion MAUs on the platform by 2030. Talk to me about the role emerging markets will play in getting to that type of target and how you think about the monetization curve of those types of users relative to what you've seen in developed economies.

Yeah, so we said two things about long-term goals. We said 2030, we hope for 1 billion MAU. And then we said long-term, we hope for 1 billion subscribers, just to make sure that the ambition is very high we want to be the biggest subscription business on this planet and that's that's where we're going we feel very good about the 1 billion MEU like you know Alex and I take nothing for granted every day is a fight but the trajectory is very much on track for that so we feel good about that in terms of monetization you're right that a lot of the growth is happening in emerging markets and as I said we're in this fortunate position where the mature markets just keep converting so we can you know Alex jokes and says you know we're using the west to fund the east right because a lot of growth is in the east but it's also true this is a benefit of us we can keep that average conversion we can keep the average monetization good but we can actually fund our own growth over here because we know this lindy effect we trust so deeply in it so a lot of the growth is coming from there but you also see us actually being very disciplined about that growth so you saw in the most recent quarter what we said that in India we've actually played a little bit with the ad load we've done individualized ad load we played a little bit with the conversion gates to make sure that the conversion stays on the track that we want to have it to feel comfortable about the plan but by and large because we trust so much in this model and it has so much precedent the main opportunity, the main focus for us is to just keep it growing. What we see is that it's so much cheaper to eat market share early in a market than when it's mature. I mean, this is obvious, but it's still true. So when you are growing quickly, you should try to capture that market share because we have so much data saying that we're going to be able to monetize it. It's much more expensive to try to buy market share when a market is already mature.

Last question for you. If I have the opportunity to have this conversation with you 12 months from now, 24 months from now, what are you most excited about building and scaling in this business that you'd want to come back and keep talking to investors about over the next one to two years?

I mean, there are a couple of things. I'm an engineer, so I'm a tech nerd, so I'm going to talk about some features I'm very interested in. But by and large, from a systems point of view, the thing that Alex and I are later focused on now is to prove out this power law or rather the power law is true and we have proof points like audiobooks plus the add-on we want to deliver more proof points so what I'm hoping we can say in 12 months is we don't have one add-on we have several and we can show that it's possible to monetize all of these niche audiences that we have in the youth space so that's one thing I would like to deliver personally one of the things I'm the most excited about is the fact that because of large language models and the fact that these computers not can speak English, Spotify is starting to become something different where you can literally today, the taste profile is rolling out in the U.S. as we speak. You can say to Spotify, you know, Spotify shows you who it thinks you are. And to me, it's just like, I think you're a big, you know, EDM user. It's David Guetta, all of this. And I can say like, well, I don't feel like that anymore. I want to be different. I can instruct the system for who I want to be. I can speak to it in English, both through text and increasingly through voice. So Spotify is becoming this interactive conversation. And so right now, one of the things I'm the most excited about this summer has been something called running mode, where we use our prompted playlist. And you can go in and you can say, I want a running playlist for running a five-minute I want it to be 170 steps per minute. because that's the cadence I want to run in if you don't want to hurt your knees. And Spotify goes in, it looks through all of your listening history since 2008, it looks through your genres, it takes the music that is roughly in that BPM, we have perfect BPM data, which no LLM has, takes it and then it actually beat stretches the music to exactly 170 BPM. So every step lands on the beat. But then on top of that, it actually audio coaches you. So you can say, do you want a pyramid run, do you want an interval run or a straight run? So Spotify is starting to become something very different. It's starting to become an intelligence that can soundtrack your life with high intelligence. So I'm hoping that 12 months from now, this is a highly interactive experience. And you just think of it much more as a friend than a service, if that makes sense.

Well, why don't we leave it there? Thanks so much for the opportunity. Please join me in thanking Spotify for being part of the government. Thank you, everyone.