Applovin at Goldman Sachs Communicopia: Strategic Growth Insights

On Wednesday, 10 September 2025, AppLovin (NASDAQ:APP) outlined its strategic vision at the Goldman Sachs Communicopia + Technology Conference 2025. The company’s leadership, led by CEO Adam Foroughi and CFO Matt Stumpf, discussed their focus on growth opportunities, capital allocation, and the expansion of their recommendation model. While AppLovin is optimistic about its trajectory, challenges such as increasing data center costs and competitive pressures were also acknowledged.

Key Takeaways

• AppLovin aims for a 20% to 30% long-term growth rate, leveraging its MAX mediation platform and AI advancements.
• The company has allocated $5.5 billion to share buybacks over the past three years, reflecting its commitment to returning value to shareholders.
• Expansion plans include a self-serve ads platform launch and international growth, particularly in e-commerce.
• Generative AI is poised to enhance ad creative, potentially boosting campaign performance.
• AppLovin’s gross ad spend in Q1 exceeded $11 billion, underscoring its substantial market position.

Financial Results

• Gross Ad Spend: Over $11 billion in Q1, positioning AppLovin as a major player in the advertising space.
• Long-Term Growth Target: Aiming for a 20% to 30% growth rate, supported by strategic investments and technological advancements.
• Share Buybacks: Approximately $5.5 billion invested in buybacks over three years, utilizing 100% of free cash flow.
• EBITDA Margins: The company expects to maintain margins between 80% and 85%.
• Data Center Costs: Growing at about 10% of overall revenue growth annually.

Operational Updates

• MAX Mediation Platform: Experiencing double-digit growth in gaming supply, driven by impression growth and revenue per impression.
• E-commerce Expansion: Applying its growth framework to new verticals, including e-commerce.
• Self-Serve Ads Platform: Scheduled for launch on October 1, 2024, with international expansion ahead of schedule.
• Generative AI Integration: Enhancing ad creative, campaign analytics, and customer support.

Future Outlook

• Growth Drivers: Reinforcement learning and model enhancements will fuel growth, alongside new market expansions.
• Target Audience: Focus on companies interested in performance-based advertising, from discovery to transaction.
• Potential Market Expansion: Aiming to significantly increase its customer base from thousands to millions.
• Gaming Market Influence: Potential revenue uplift from collaborations with gaming giants like Activision.

Q&A Highlights

• Competition: Seen as beneficial, with AppLovin’s platform growth catalyzing broader market expansion.
• Capital Allocation: Emphasis on organic growth and slight scaling back of buybacks in the future.
• Efficiency: Commitment to automation and lean operations to maintain high margins.
• Technology: Powerful recommendation engine and AI advancements drive non-linear scaling.

For more detailed insights, readers are encouraged to refer to the full transcript.

Full transcript - Goldman Sachs Communicopia + Technology Conference 2025:

Eric, Interviewer: In the interest of time, we’re going to have to keep the conversation going. It’s my pleasure to have the team from AppLovin here. We’ve got Adam Foroughi, CEO, and Matt Stumpf, CFO. Adam, Matt, thanks so much for being part of the conference. Adam, thank you for coming back. You were here two years ago. We had a great conversation. We’re excited to have you back at the conference.

Adam Foroughi, CEO, AppLovin: Thanks, Eric.

Eric, Interviewer: Okay. Adam, let’s start with you. The business has evolved a ton since you were here two years ago. Talk a little bit about the journey the company’s been on and set the stage a little bit, maybe big picture, on what the priorities are for where you want to take this company in the years ahead.

Adam Foroughi, CEO, AppLovin: Yeah. First of all, thanks everyone for joining. I think the last time I was sitting down with Eric, we had like five investors in the room. This is a much bigger audience. We have had one of the wilder rides in the public markets. If you’ve listened to me talk, and I did podcasts in 2017 all the way through to now, you will have always heard a consistent message. We exist for one reason and one reason alone: help an advertiser find a new customer, engage that customer, run an advertising campaign from top of funnel all the way to bottom of funnel, get a conversion, and price it on revenue. That is a really, really hard thing to do. With the modern era of neural nets, you can do it much better than you could have done it years ago.

In the world today, the largest platform that does this is Meta. There is not a second. We’re the second. We’ve been able to scale much bigger than I think a lot of people realized when they looked at our stock over the last few years until we put out some disclosures in Q1 around the fact that in Q1 we were over $11 billion of gross ad spend. People always look at our revenue and don’t realize that the revenue is the net after we paid publishers. In terms of gross, and obviously we’ve grown quite substantially since we put out that disclosure, that’s the equivalent of Pinterest, Snap, and Twitter total revenues on our platform already. What’s also really interesting about it is that we’ve disclosed that we had hundreds of e-commerce advertisers at the time, and our last disclosure on gaming customers was in the hundreds.

You’ve got a platform that might have 1,500 advertisers with spend at that level. Never happened before. I don’t know of another $200 billion almost company that hasn’t actually launched their product yet. What gets us really excited internally is that we have this product that’s exceptionally high value for the customers that we have. We haven’t launched it broadly, so we haven’t given most of the companies in the world the opportunity to tap into its excellence, this recommendation model that we built. We’re on the cusp of doing that. As we go forward, the way we look at the business in the future is we took 13 years perfecting a product in a niche market, the gaming market. Turned out it was economically a pretty big market. Broadly speaking, in economic terms, the gaming dollar spent versus the total TAM of the economy is not that big.

Now we’re about to tap into the whole world. If we can help every type of business, top of funnel to bottom of funnel, and price on revenue, not only are we going to be a really great return for our shareholders, we’re going to help their economies expand. We’re going to create millions of jobs. That gets us really excited internally.

Eric, Interviewer: Okay, there’s a lot to mine in there. We’re going to touch upon all of those themes as we get through this conversation. Matt, let’s bring you into the conversation.

Adam Foroughi, CEO, AppLovin: Sure.

Eric, Interviewer: As a company, you’ve been very transparent in laying out how to expect to grow going forward. Talk a little bit about the building blocks of growth for this company and how should people think about that trajectory in the years ahead.

Matt Stumpf, CFO, AppLovin: Yeah, I think what’s really important to understand is, first of all, before the 20% to 30% kind of goal that we’ve put out into the public markets in terms of long-term growth, is to first understand that we sit on top of a large amount of gaming supply. We’ve talked about the MAX mediation platform that we own and implied historically in earnings calls that that mediation platform, that that supply is growing double digits on an annual basis. That’s coming from a combination of both impression growth as well as dollars per impression. First, that’s the supply that we have access to. You look at the 20% to 30% growth on top of that coming from technology, and that’s comprised of two components.

First, reinforcement learning, which is super important in these AI technologies that the model continues to improve from the recommendations, the predictions that it’s doing about what’s going to happen when it shows an impression, and then also directed model enhancements, so improvements to the model over time. What’s really important to understand about that 20% to 30% goal is that that also applies to e-commerce and these new verticals that we’re going into. As you think about the e-commerce expansion that we’ve talked about in Q4 and then in 2026 of next year, that framework also applies there. More importantly, we’ve created this fantastic recommendation engine that historically we’ve only had a certain amount of demand, right? Only mobile gaming demand to feed into that recommendation. We’ve kind of kneecapped it.

By providing more demand, demand diversity from e-commerce, you can really, really allow the recommendation engine to get better over time. That 20% to 30% should be a baseline of growth going forward. We’re very optimistic and excited about the potential for growth in the future.

Eric, Interviewer: Great. Okay. Adam, you have one of the more unique perspectives. Whenever I get a chance to ask you about competition, it’s interesting. You always reflect back on just the execution of the platform itself. We get it a lot from investors. Who do you compete with? How do you think about competition rather than just executing instantly on the opportunity in front of you?

Adam Foroughi, CEO, AppLovin: There’s a lot of companies that are trying to service this mobile gaming audience because it’s big. It’s a billion users. We as a marketplace allow competition. We want competition to thrive, which is counterintuitive in most marketplace businesses. If you take a typical ride-sharing market, if Lyft gets a ride, Uber’s going to have a ride that’s lost and vice versa. Most investors are trained that most markets end up this zero-sum equation where competition takes from another. In this mobile gaming market, two things have occurred. One is our platform’s growing so fast, but it’s catalyzing the growth of the impressions and dollars per impression in the marketplace. You see other competitors like Unity, Liftoff, Moloco, Mobisa, all these companies are growing 30% plus. At the same time, we’re huge and we’re growing 70% plus. You’ve got a company that’s the whale in the market growing that fast.

If it was truly constructed like these other ecosystems, you would think that the other companies should be suffering, but they’re not. It takes a while for investors to then ask the next question of why. What’s different about this marketplace? What’s different is that everyone’s built a recommendation engine on top of a model. In the recommendation engine space, you’re built on a neural net, so similar architecture to a large language model. When it comes to these businesses, if you get more data into your model, the training data and the reinforcement data is what creates an output that becomes more compelling. In all the large language companies, you hear training data matters, more data, more compute, retrain, better output. Our world, it’s the same thing. No model has all of the training data. There’s no way to take everyone else’s training data.

The model’s only as good as the data it’s trained on. What you end up having happen in a big audience is that every company’s model has a reason to exist. When we lose an impression, someone else may have had a model that makes more sense in their recommendation to the user. What’s important about our economic value proposition is that we’re the market and the lead market maker. This is a concept that most people haven’t understood and matters a lot. When we serve an ad, sometimes the model is really, really predictive on the outcome. We can bid a huge number. At the highest end, let’s say $2,000 CPM, $2 for a single impression. If the model is doing that, no one’s going to beat us in that moment. It is very predictive about the outcome.

Most of the impressions we serve, the model doesn’t know really what to do. It’s overconfident about what to serve. In fact, we’ve said historically we’re 99% wrong. As good a business model as we are, we’re 99% wrong. In most of those impressions, we’re probably 99.9% wrong. What happens when another model gets data where they’re not overly confident they’re actually right? They can buy some impressions away from us. As the house in this market, we now tax that competitor 5%. When you can get 5%, which may not sound like a lot, if you understand typical ad network margins, a substantial amount of the margin of your peers, and you make a 5% with 100% flow-through instead of losing money, that’s a really good economic transaction. We’ve fueled growth in the economy. Every time growth happens around us, we grow with it.

That’s a fantastic equation to sit on top of.

Eric, Interviewer: Okay. One more question on where we are today before we talk about the go-forward. Matt, you guys obviously are a very profitable business. How do you think about the allocation of free cash flow and the priorities for capital as the business stands today?

Matt Stumpf, CFO, AppLovin: Yeah, yeah. First and foremost, we’re focused on ensuring that we’ve got enough capital allocation towards our organic growth initiatives. That takes the form of continuing to support data center growth for capacity, GPUs, et cetera, to continue to support model expansion and scale, and then headcount. Headcount, we grow, I mean, as we’ve talked about quite a bit, we just have a culture where we don’t expand headcount materially. We’re adding headcount, but very minimally. I think about that as more kind of a fixed cost general. We’re left with quite a bit of free cash flow. What are we doing with that? We don’t believe in holding cash for potential hypothetical M&A in the future. We believe that the best investment for our capital is then investing back in our own stock. Over the past three years, we’ve done that.

We’ve invested almost $5.5 billion in total, which represents about 100% of the overall free cash flow over that same period in our own stock through buybacks and withhold to cover, which has been a tremendous return for our shareholders. Going forward, we do plan on scaling that back very slightly, but we still think that with these growth opportunities that we’ve been talking about, the best return for our shareholders is to continue to invest in ourselves and to buy back shares.

Eric, Interviewer: Okay. Let’s build on some of those growth opportunities. Adam, you’ve talked about e-commerce. We’ve even talked about wider web advertising. How are you landing on what the right next market opportunities are for you as a platform? How do you think about aligning your strategic priorities to capitalize on those opportunities?

Adam Foroughi, CEO, AppLovin: I mean, look, we have a huge opportunity in front of us just opening up the platform to web-based advertisers. E-commerce being the biggest TAM and the most fragmented industry and the one that we’ve really played around with inside our pilot. We’ve been in this pilot state for a year. We went through and we recruited hundreds of advertisers that we disclosed in Q1. I said a year ago, my scoreboard was, can we get as much dollars from these customers as they would be willing to spend on Meta? Now you have to understand, we’re not asking anyone to take dollars from someone else. We are trying to show them that we are absolutely incremental to the dollars they spend elsewhere. This is a really important concept to understand.

If these companies are spending, let’s say you bring in a company A and they spend $1,000 a day on Meta and they have $2,000 a day of revenue from that $1,000 spend. When we launch our product, we want to measure it by being able to give them $1,000 a day and $2,000 a day of incremental revenue. If we could do that, their business goes $2,000 to $4,000 and it doubles. In an arbitrage-based world where you give them top of funnel to bottom funnel price on revenue, you don’t expect the customer to take from others to add. What you expect is that they’re going to have unlimited budget. In an arbitrage, you want to go as hard as you can. We, with the way our technology and product is set up, allow them to become arbitrage marketers. They get a product sold.

They get a credit card paying them. The money shows up in their account. They see the media costs that they don’t pay down their credit card on until 60 days later. They’re immediately profitable on the media costs, and then they get the new customer. You provide that value, which is exceptionally hard to provide. Again, only Meta has unlocked this value in the past, and now we’re that number two. You provide that value, you have a very scalable business. There are a couple of reports that were put out there by Northbeam and Triple Whale that will back up my point that I believe we’ve reached the level where I’m confident in how we measure up in the marketplace. The Triple Whale and Northbeam reports, and these are two analytics providers in e-commerce, they process billions of dollars of transactional value.

They know where the media spend is. They put out two tables. One showed 70% of the market spend went to Meta, 20% went to Google, and we were around 4%. We were bigger than TikTok and everyone else. We were third biggest. The other one was like TikTok a little bit ahead of us, and we were like 3.7 or something. If you look at that at face value, you’d go, Adam doesn’t know how to score his own product. They’re not even close to Meta. What is he talking about? You have to remember that these companies, we don’t work with most of the companies that they represent. If you understand that we probably work with 5%, maybe at most 10%, you’d take that 4 and divide by 0.1.

If you do that and you reduce everyone else by the expansion of us in the budget share of the like-for-like customers, you’d all of a sudden see we’re pretty high up on the list. Also remember that we are just in the midst of opening up international. Our business is half U.S., half international. We can agree Facebook and Google and everyone else on that list is global and works with probably everyone at these companies. We work with a small fraction and we’ve limited our traffic, yet we’re indexing that high, which tells you we have perfected a product to a cohort. As we open up the platform, we should just get better.

Eric, Interviewer: Okay, let’s stick with that theme for a minute. I know you’re laser-focused on the e-commerce opportunity and building for some of the go-to-market strategies around it. We’ll talk about it. Take a step back for a second and talk about how wide the aperture of what this could be applied to over time. We get asked about connected TV. We get asked about a wider array of the open web. What will you be looking for from the e-commerce initiatives that could possibly widen it out beyond that over time?

Adam Foroughi, CEO, AppLovin: We want to work with companies that want to serve advertising on a performance basis defined as discover a new customer and be able to price all the way to the point of transaction. That doesn’t mean that it’s always revenue. Some companies in lead gen might collect an email address. Financial services might have a different goal. The reality is the company has to have a desire to do performance the way we define it. We’ll never go into the expansion area of brand advertising. When we get asked that, it’s a simple no. That’s just not our business model. We’re focused on building economies for our partners. On the supply side, what’s interesting is today, obviously, we don’t have a demand limitation. We only have hundreds of customers. We’re extracting this much spend from them.

They’re not going to go spend five times what they’re spending on Meta on us. We don’t have unlimited budgets. Let’s say we open up our platform, fast forward a bit of time, and we go from hundreds to hundreds of thousands of customers. If we get to that point, you can imagine at some point, we will want to go out and place the ads for these customers anywhere we can find future customers for them. In theory, at that point, we’ll have excess demand. Where would we go first? It’s not going to be connected TV or even where people would think. It’s actually going to be other game apps. A lot of the gaming apps in the market monetize with in-app purchasing. These are the biggest games. The $100 billion TAM monetizes with transactions to Google and Apple. These companies traditionally don’t run ads. Why?

Because they don’t want to show their competitors in a game that monetizes exceptionally well. The TAM is $100 billion. Now, Bobby Kotick, fantastic guy over at Activision when he was the CEO, launched ads on the King franchise, Candy Crush. The ads became 15% of the revenue of King, roughly. If you could take a $100 billion TAM and lower it to $50 billion, let’s say we can convert half of it and say 15% revenue uplift to them for not running competitors. I’m not a great seller, but I can sell that. On the other side of that, that’s $7.5 billion extra of publisher scale all to our non-gaming product. These are already our customers that literally plug and play with the MAX platform into our demand. That’s the first place you go. That’s a massive uplift in our business.

If you gross up the pub spend to gross, you can look at that as basically a double to what we do today. You go further than that. As we keep expanding from there, you’re starting to look at, can we expand into CTV? Can we expand into serving ads in other social networks, into utilities? We’ll want to go everywhere over time, but we’re not in a rush. Our first priority is get the customer on. We have tons of eyeballs, tons of supply, expand inside gaming, and then broaden out.

Eric, Interviewer: Okay. Understood. Matt, bringing you back into the conversation. Adam laid out where you want to go as a platform for the longer term. How as a team do you think about balancing growth investments against incremental margins against some of the capital allocation we talked about earlier on the go-forward?

Matt Stumpf, CFO, AppLovin: Yeah, I mean, we have fantastic EBITDA margins today. We’ve laid out a framework for investors in the past that with the existing business, the basic kind of concept is data center costs have grown at about 10% of the overall revenue growth. That’s what we’ve seen on an annual basis. We continue to get GPUs and support the scale of the platform. It’s about 10% of that directly from revenue growth. We’ve been adding headcount, but very minimally. We almost think of that as a fixed cost, to be honest. Going forward, we will continue to grow both the headcount for our engineering team and business development to support those new initiatives, but very minimally.

As we push into things like e-commerce, where we are going to see new costs that investors should be aware of, those are for things like API calls, as we use LLMs for agentic customer support, campaign analytics, and management. We’ll also be building in tools for advertisers to use generative AI for their creatives. Those things will also cost incremental amounts for the company. As well as on our last earnings call, we also mentioned that we’ll be doing performance marketing to also improve the visibility of our brand and expand into new markets with advertisers who are not aware of the AppLovin brand and now Axon. Those costs will increase, but given the fact that these new costs will be revenue-driving costs and will be performance-based, we don’t expect that the margin profile of the overall business will change materially from where we’re at today.

For investors in terms of expectations for the remainder of the year and the next year and in the future, we do still expect that we’ll have between kind of 80 and 85% EBITDA margins which are very healthy.

Adam Foroughi, CEO, AppLovin: There are two things Matt said that I want to cover from a business perspective. One thing that I don’t think is priced particularly well in advertising companies today, and this isn’t limited to us, is this notion that generative AI is going to allow an explosion of the ad creative. I’ve seen hundreds of thousands of ad creatives put live on our platform, and a good one versus a bad one can be over 2x the performance for the same customer. Literally, a customer that spends all this time building a product can double their performance, which doesn’t translate to double the scale. It might be 10x the scale from investing $5,000 into an ad creative that’s better than the one they currently run. That’s a huge lever. In fact, it’s the biggest lever that advertisers will have going forward.

If you live in a world of static, manual, human-built ads, and our world is a video advertisement plus some follow-on to the video advertisement where the video ad on average engages the user for 33 to 35 seconds, you have to be creative. You have to build a lot of content in. The human being is just not as creative as a machine. If the generative AI tools get good enough and VO3 is on the cusp, we’ve seen this from Google. They just rolled out portrait mode and we’re vertical video for the most part. If they can create a great ad out of the box, we will pay for that.

The expansion of our business going from a fixed set of ads to this human being being able to tell the machine, "Build me 20 variations of this best ad I’ve got," and the machine doing a very good job of that is not going to be a small amount of uplift in terms of consumer response on our platform. It’s going to be a very large amount of uplift. We will better use the impressions to better personalize on the other side. That’s something that we can’t size. People ask, "How big could that be?" No idea. From my experience, it’s going to be really big. The second piece that he said is we might performance market. I mentioned this on the last earnings call and people were so confused. What is he talking about? We’re a B2B company and we can agree that we have fantastic economics.

We have a great business model. The only risk we have in the future is that there’s not a domain in front of our ads product. The ads product is fantastic, but we’ve got to go recruit customers. There’s a massive world of 10 million plus customers. We get a lot of shots on goal to do this. If we go from the very low thousands to the hundreds of thousands to the millions eventually, we’re going to be a much, much bigger business than today. How can you do that in the absence of a domain in front of your property? You can market.

I fundamentally believe we can run a Super Bowl ad in a couple of years if we’re doing a really good job as a product and serve it to the audience that watches the Super Bowl, which is going to be inclusive of a whole bunch of small business owners and employees of other businesses who can benefit from a platform like us and may just not be getting that benefit because they haven’t heard of it yet. If we can do that, if you start seeing Axon ads promoting the Axon platform everywhere out there, that’s going to catalyze even more growth because we’re really good at performance marketing. We have fantastic economics. Our LTV, the cost of user acquisition function, should be as good as anything the world’s seen. These are things that we’re really excited about. We don’t run at 81% EBITDA margins because we’re supremely cheap.

We are pretty cheap and we like to be very optimized, but we do want to reinvest into growth opportunities. If we can pull up revenue growth, which already we’re starting at a really phenomenal baseline, if we can do it in automated ways, we are going to invest behind that.

Eric, Interviewer: Okay, I want to build on that and I want to come back to the underlying tech maybe before we run out of time. The shorter-term narrative that’s getting a lot of investor attention is you’re in the process of launching a self-serve ads platform on a smaller scale to begin, and then you’ve articulated in the public domain that it’ll be on a much wider scale as you get into the front part of next year. What was the genesis of moving in that direction towards launching self-serve, and how should investors think about that opening up the array of advertisers that will likely be partners with you as that scales in the years ahead?

Adam Foroughi, CEO, AppLovin: Look, our business aspirations have grown because the technology is obscenely powerful, and we’re disservicing what our engineers have built on the business side. Let me unpack that a little bit. We’ve got this recommendation engine, and a year ago, the only thing we could recommend to the consumer were game ads. It’s as if TikTok was just golf instructional videos. Nobody but a very narrow slice of the world would use TikTok because it’d be boring, and they wouldn’t realize the true potential of their tech. If the user is telling us 99% of the time, "We like the game we’re playing," like, "Stop showing me game, stop showing me game," we’ve done a disservice to our technology on the business side. Where did we go? We said, "Let’s prove that we can make it work across other industries, transactional industries, where we’re not showing the user another game.

We’re recommending them products that they can go transact on and then come right back to the game that they were playing." We went out and we launched this pilot. Obviously, it worked really well. There’s a lot of excitement around it. We proved to ourselves that the opportunity is much bigger. I got a funny email the other day. There was a consumer that actually found my email and emailed me. She was complaining, and I chuckled a little bit because the complaint was, "You’ve served me 250 ads in a row for an e-bike. Stop serving me an e-bike. I don’t want an e-bike.

Like, what are you doing?" I felt like I should just buy her the e-bike because when I looked at the model and I was like, "What is it doing?" It thinks that she needs something that’s a mode of transportation or something in athletics, but we don’t have the density and diversity of the customer yet. The model goes, "I don’t have anything else to show. Here’s the e-bike 250 times in a row." Think about that. That’s an embarrassment, but that is a disservice to our technology.

If in a year from now there are 250 other things in the category that the model predicted for, and instead of one over 250, it’s 250 over 250, not only is she going to be less annoyed, hopefully I’m not getting emails anymore from her, she might even buy something, but she’s certainly going to have a different response rate to the 250 things. Remember, we’re a large model where everything we do is a data point that goes right back into that model. If 250 times our data point became one, it’s devalued. If 250 times it became diverse, we got the demand, we got the revenue, we got the increased conversion. More importantly, to me, as someone who is proud of the technology our team has built, we brought more data back into the model to better understand what this woman wants.

If we do that on the next 250 impressions, we’re going to be much stronger in what we show her. That is what gets us really excited about the business opportunity. Getting to a point of wanting to open up the platform is simply that our engineers have built phenomenal technology, and now we want to take it to the world and not continue to do them a disservice on the business side.

Eric, Interviewer: I know, Matt, you sort of answered this earlier, but I just want to put a finer point on it so investors can understand it because you talked about the efficiencies as well as a team that you like to drive. At the beginning part of this year, you laid out sort of this is where we think the platform can go. Here are some of the initiatives. Some of them are going to happen this year. Some of them are going to happen on a longer term. Even ad creative was one we talked about a couple of earnings calls ago. How much efficiency can there be in the business to strike that balance on margins and making sure you don’t miss out on the growth opportunities? Please talk a little bit about continuing to find ways to strike that balance.

AI at the center of the company also seems to be a driver of some of those operating efficiencies. It’s not like you’re reducing heads for the sake of just reducing heads. You’re getting better operating efficiency from what we could tell.

Matt Stumpf, CFO, AppLovin: Yeah, I mean, it’s partially the culture of the company, I would say, in the first place that Adam has built. The way that the company runs and the way that our employees and management team thinks is that before we do anything, we think about what’s the best way to do that. Can we do that in an automated way? When we were talking before about adding costs for LLMs to do agentic customer support, that’s one example where rather than hiring on hundreds or thousands of people in the business development team to support growth in customer count, the team wants to get ahead of that and build a way that we can do that in an efficient manner. I would say that that applies to the company across the entire company.

I think that there’s a lot of room for us to continue to drive these really high margins in that fashion as long as we can continue to retain that culture of the company of doing things in an automated way, reducing process, and that applies directly to cost. I don’t know if you want to add anything on.

Adam Foroughi, CEO, AppLovin: Yeah, I mean, it’s totally aligned. We never sacrifice future opportunity for lack of investment. We sacrifice future opportunity because our culture demands that we be lean and only have eight players on the team. Something that I get constantly asked is, you’ve got this massive revenue opportunity in front of you. It doesn’t take an idiot to see that. Why aren’t you investing in the salespeople to go bring some of that revenue up? The challenge with that is as you start scaling out your sales force, you end up with people that are selling the product the engineers built, but they’re probably not as high on the IQ side as your core engineering team. All of a sudden, your culture starts expanding, headcount starts expanding, processes build, and it’s really hard to keep that team eight players for what has built you to the point you’ve built.

We think about the next 10-year timeline. We may not get all the customers that we would get as fast into the present, but I’m certain if the product is good enough, we’ll get them over time. If we can get them over time and if we can automate the recruitment of those customers through investments around tooling and advertising where we don’t have to invest in people, then the margin structure and the net dollars that we will make the next 10 years is going to be much greater because, one, you don’t have to let go of a bunch of people that you hired in your ramp-up phase, which is really, really hard to do, frankly.

Two, you retained your eight players because those eight players solely want to work with other eight players and you didn’t signal to them that we can’t continue to do what we did the whole time we built this company to this point.

Eric, Interviewer: I do want to end on one bigger theme, which is a topic we get asked a lot about by investors. Since there’s a wider audience here, I want to give you an opportunity just to make sure people understand it. I think people generally ask a lot of questions about the tech you’ve built and how the tech you’ve built scales. How linear versus non-linear are the outputs of the tech you’ve built? I think there’s been linear performance and then there’s been big upswings in performance as the company has scaled over the last couple of years. How do you think about what you’ve built and how investors should think about what they’re underwriting in terms of the yield or the output from that platform in the years ahead?

Adam Foroughi, CEO, AppLovin: I mean, look, it’s very hard to quantify it that way because it’s not like it hasn’t. It was a step function in efficiency, and it’s continued to get better. Nothing is necessarily linear or measurable in terms of the rate of advancement. These models are really powerful. Obviously, the math, the technology, we all know how powerful the large language model space is. We can all see Waymos on the road. The model inside a recommendation engine is, again, constructed similarly and very powerful because it’s tied to an economic business model that can become really relevant as it improves that model. The technology itself didn’t exist years ago. If we wanted to build what we built over a couple of years ago when we launched Axon 2.0 and on the continuous iteration of the product, we couldn’t have done it seven years ago. It just didn’t exist.

A lot of it was predicated on research that was more modern. You get to a place where you can drive the value to the customer and predict not only is the user going to engage with what ad, because you’re retrieving that amongst millions of ads, but what they’re going to do in a product all the way to the point of transaction, or possibly for some of our advertisers, to the point of transactions out to 30 days. This is super complex, but these technologies can do super complex and magical things. We are at a place today in the world where the research, we’re sort of on the cutting edge of research when it comes to recommendation engines. There’s no Axon 3.0. Now we have to do research to try to find the next level up around Axon 3.0, and maybe we’ll have a breakthrough there.

Maybe the market will have a breakthrough there. If you fundamentally believe these technologies are going to get smarter over time and the compute’s going to get more powerful and you’re going to have an advancement in these fields, the nice thing is we’re really good at these technologies. We will benefit from that, whether it happens external to us or internal to us. Those continued improvements will make the math more predictive. We’re really strong at driving result to advertiser. Every time we have an advancement in tech, we can take the population of the billion plus daily active users and serve them something more relevant. Even though the advertiser still sees the same return on ad spend because the product works, the scale goes up, which is the catalyst around our growth.

I did want to touch on one thing too that I wanted to state on this before we wrap, because I don’t want to create an IR problem for us as we go into a closed window. We had said we’re launching our self-serve ads platform, 10/1, and we’re opening up international to our web customers, 10/1. We’re pacing well against the 10/1 goal, but we did just open up international to our customers a couple of days ago. If you start looking on Twitter or talk to advertisers, you may see that they’re starting to see good performance outside of just the States. We’ve been testing this product broadly outside of the States with some pilot customers inside the pilot internationally. The tests are great. Results have been phenomenal. We’re ahead of schedule on that. That’s already in market.

I just want to make sure investors realize that that’s three weeks ahead of schedule.

Eric, Interviewer: Okay, that’s a good point. I do want to ask one more. I always like to end with people, if we get a chance to have this conversation a year or two from now, what are the things you’re the most focused on executing on over the next one to two years?

Adam Foroughi, CEO, AppLovin: Look, the thing I’m most proud of with what this technology has done is we’re going to show the world that the game customer is valuable. It’s funny because like 15 years ago, people would have told Facebook, presumably, that social networking users are trash. They’re not monetizable because no one had shown you can monetize them. It turns out everyone uses social networking, and human beings are very monetizable with the right technology. While we’ve built this business, I’ve talked to many brands, like this is a year ago in the past, that have always said game inventory is trash. I went, "Okay, what are you talking about? It’s a billion plus daily active users. These are adults. These are casual gamers. They skew female. The heads of households are here. These people are playing more game time than they’re in social. You’re probably not accessing them.

Certainly, if you’re not working with us, you’re really not accessing them because we’re the whole market here. Why do you think they’re trash?" The answer was simply, "Well, no one can target them for us. We think they must not transact." What was missing was a technology in the middle that could help unlock the value of the customer to the advertiser. If we do our job right in a year or two, people are going to start walking around saying the game user is really valuable. These are human beings, billion plus human beings. They’re all adults. Of course, they’re valuable. I would love to be able to do that for the industry that gave us everything. That’ll create more value for gaming companies. It will obviously create more value for the customers.

If we do our job right, we’ll expand the economy the way we think we can.

Eric, Interviewer: Guys, thanks so much for the opportunity to have the conversation. Please join me in thanking AppLovin for being part of the call.

This article was generated with the support of AI and reviewed by an editor. For more information see our T&C.