Duolingo Aktie

Good evening, everyone, and welcome to Duolingo's Q3 earnings webcast. Today after market closed, we released this quarter's shareholder letter, a copy of which you can find on our website at investor.duolingo.com. On today's call, we have Luis von Ahn, our Co-Founder and CEO; and Gillian Munson, our CFO. They will begin with prepared remarks before we open the call for questions. Analysts may ask a question by using the raise hand feature. Please note that this call is being recorded [Operator Instructions].

Before we begin, please note that we will make forward-looking statements regarding future events and financial performance. These statements are subject to risks and uncertainties described in our SEC filings and are based on our assumptions we believe to be reasonable as of today. We undertake no obligation to update them. We will also discuss both GAAP and non-GAAP financial measures. Reconciliations between the 2 can be found in our earnings materials, and we encourage you to review them when evaluating our performance.

And now I will hand over to Luis.

Thanks, Debbie, and thank you all for joining. We had a great Q3. Daily active users accelerated for a second quarter in a row to 29% year-over-year growth, up from 23% in Q2. Bookings, revenue and profitability were also ahead of expectations. At the beginning of the year, we laid out a plan to prioritize teaching better and user growth, knowing it would take time to see results. We're now seeing the user acceleration we hope for, and I'm proud of how our team has executed against our strategy. We expect this momentum to continue with year-over-year DAU growth in Q4 similar to Q3 and Q4 bookings slightly ahead of previous guidance.

Every new learner expands the audience we can monetize over time. While teaching better and user growth remain our top priorities for this year, we're gradually leaning more into monetization in ways that preserve the quality of our free experience. Our ambition is to ultimately reach 1 billion learners, and build a much larger business over time. The more people we can teach, the bigger that opportunity becomes.

With that, I'll turn it over to Gillian.

Thank you, Luis. I'm glad you're all here this evening. We started the year with an ambitious plan to reaccelerate our growth, and our Q3 results show we're executing well. The numbers from the quarter are pretty straightforward, so I'll double-click on just 2 items. First, a look at how broad our DAU and bookings growth was. Year-over-year, DAU growth accelerated across 8 of our 10 largest countries. Additionally, year-over-year subscription bookings growth was in the double digits in the majority of the regions we serve.

Secondly, Q3 operating expenses include an $11 million bonus accrual. This bonus rewards our team for reaccelerating user growth. We're on track to hit the targets, so we've begun to record the expense. A second portion of the bonus is included in our Q4 guidance. and we expect to pay the full bonus in early 2027.

Turning to Q4. We expect year-over-year bookings growth of 13%, revenue growth of 14% and an adjusted EBITDA margin of 26.1%. To help you build your models, I want to give you just a couple of items to consider. We expect gross margin to fall slightly from Q3 as we roll out Video Call to all Paid Subscribers. Operating expenses will likely grow faster than revenue because of our continued investments and the second portion of the bonus I just mentioned. As a result, we expect adjusted EBITDA to be roughly flat year-over-year at $84 million.

For the full year, we expect to deliver 12% year-over-year bookings growth and 18% revenue growth paired with a 26.5% adjusted EBITDA margin. We plan to provide our full outlook for 2027 in February. We are encouraged by the momentum we're seeing in our business and look forward to answering your questions.

Back to you, operator.

[Operator Instructions] Your first question comes from Ygal Arounian.

I guess, I want to maybe start with the growth that you guys are seeing and we expect to continue into 4Q. A number of levers you guys are pushing here. And maybe if you could just kind of quantify or qualify what the biggest drivers are between the new partnerships and the games and the Junior's Toy Hunt and those kind of engagement features that you're driving?

And then just a second part of that question, any difference in terms of monetization or retention or other KPIs that you're seeing from these partnerships? And should we expect you to do more of those?

Okay. Yes. Thank you for the question. So first of all, the majority of our growth for this year comes from product improvements. We've, in the past, discussed our notion of The Green Machine, which is basically an optimization system that we have that just makes the app better over time. And basically, every week, the app has hundreds of new A/B/B tests that make it a little better. And that's the vast majority of our improvement in active users comes from improvements in retention because of that. The product is getting better.

Now over the last few months, we've had a few events that are also very important and important to call out. One of them was the Streak Revival last quarter. This quarter, we had the partnerships with Brawl Stars and also Eggy Party. And they did bring users, but I really want to emphasize that we just believe that the majority of it is from improvements to the app, which is why we think that this is going to continue going forward.

In terms of the uses that we have from these partnerships, they retain very well. And part of the reason for that is because we did some product work to make it. So the way these partnerships work is that you're a player of Brawl Stars or whatever the game that is partnering with us is. You come to Duolingo to get rewards for that game, for Brawl Stars for example. But in order to get the rewards, we get you to stay multiple days on Duolingo. And what that does is it builds a habit. And so we see that these players that are coming in have very good retention.

And I think as you think about both DAU and bookings, I think what we've been talking about is how broad-based -- the growth was in the quarter, both from a geographic perspective, in terms of how the product is working. We're feeling really good about the broadness of the growth we're seeing in the business across both DAU and monetization.

Yes. I mean we're basically growing across all regions, and we accelerated across 8 of our top 10 markets in terms of DAU growth. So we're very happy with that.

Your next question comes from the line of Wyatt Swanson with D.A. Davidson & Co.

Could you guys give some color on the Paid Subscriber additions relative to daily active users? Paid subs growth has continued to slow year-over-year despite DAUs picking back up. Is that predominantly a function of extended free trials? Or could you give any additional color on that and whether you expect those trends to change at all?

I'll cover the subs numbers and then maybe you can talk a little bit about the trends in there. Wyatt, the way to think about subs numbers is the largely follow the revenue. So they're kind of an average in there. So the subs numbers tend to follow that revenue growth. So in terms of subs growth, probably that bottoms in a couple of quarters as we work through the bookings from prior to this quarter.

So as we rebuild the bookings growth, that revenue growth will come after and those subs growth numbers will come as well. So it's just -- that's like sort of a math thing mostly. And then Luis, maybe you'll talk a little bit about sort of how bookings are going near term.

Yes. I mean, we started the year with a very clear strategy to prioritize teaching better end user growth. And we've -- and it's very hard to reaccelerate user growth, and we're very proud of the fact that we have done that. But we also, at the beginning of the year, we said, look, our bookings are going to fall within this box of somewhere like 10% to 12% year-over-year growth. We're in the higher end of that. We feel pretty good about that.

But we just have not been concentrating very much on monetization. But what we have done in monetization this year is try to invent new ways of monetizing that are more DAU aligned. And the good news is that we have found a number of them. One of them is longer free trials, like you mentioned. There are others, for example, monetizing the Avatar suits like we talked about, that does really well with IAPs.

And there's just a number of other things that we're going to be doing, and we're slowly starting to monetize more that way. Another thing that we're really excited about is Chess is completely under monetized. We're going to be monetizing it Chess. We're very proud. We just hit 10 million daily active users in Chess. So it's now a relatively significant amount of our users are doing Chess. And we think there's a lot of low-hanging fruit there.

Your next question comes from the line of Ryan MacDonald with Needham & Company.

Luis, just wondering if you could kind of provide a little more color on sort of the early days of the testing of moving Video Call called the Super Duolingo, what you're seeing in terms of user engagement there or even conversion? And then how that sort of informed your view about the decision to sunset Duolingo Max here?

Yes. It's a great question. I mean when we first started Duolingo Max, the higher subscription tier, what we did -- we have to start it because we had some AI features that were just very expensive to serve. I mean the cost of these AI features was high. We did say back then that if we managed to bring the cost down, we may put these features in other tiers. What happened was that the premium feature of Duolingo Max, which was Video Call, when we first launched it, it was costing about $0.30 per call, just generally on AI inference.

We've managed to reduce that cost to well under $0.01 per call, and that's mainly by moving to open source models or open weight models. So we feel really good about that, and that allowed us to put Video Call in Super. And that was actually a really good thing. When we A/B tested that, it's good for revenue, but also it allows us to give Video Call to about 10x as many users because we just have many more Super subscribers than we have Max subscribers. And that's good because that means 10x as many people are learning better, they'll tell their friends, et cetera, et cetera. So we just view this as very positive.

And because of that, and when we tested just -- we actually just tested a sunsetting Max, it was very clear that the right decision was to sunset Max. So we've started doing that. We don't think -- in the short term, we don't think this is going to have a major bookings impact. It's slightly positive. It's not a major bookings impact. But in the long term, we think this is really going to help us because more and more people are going to get Super. I mean, this is one of the main reasons why people are going to get Super. So we feel pretty good about that.

Yes. And I think what's interesting about the process we've taken here is that we've worked down the exposure to Max fairly significantly. So the vast majority of our bookings now is not Max. It's single digits in both in terms of the number of paid subscribers and the bookings there. So it's become a very small thing that we're going to manage, and we've obviously shown this year that we can manage these things quite well.

And Gillian, maybe just assumptions for sort of like churn as you're going through the process? And is this something that you're going to push out sort of all at once to existing Max subscribers? Or are you just going to do it gradually upon renewal?

So we have essentially stopped selling Max to new users. We will be moving Max individual to Super Family at a similar price. So they will be getting the ability to add their friends and family on all the great features that you have in Max moving over to Super. And then that will just leave us with Max Family which we will be ending the year with, which is the smaller portion of bookings and paid users that I mentioned before. So we've really managed down that process.

And again, kind of to what Luis' point was, as we go test this, it ends up being net positive for us. And then on the long term. In the short term, we manage through it, but you've seen we've really managed down the exposure there, and we feel really good about our ability to manage it.

Your next question comes from the line of Shweta Khajuria with Wolfe.

Can I please try 2. One is a follow-up on bookings growth. So how do you think about this lag between DAU growth and bookings growth as we think about bookings growth catching up, how should we be thinking about it as we think of '27?

And then the second one is on gross margins. Is 72% the guidance that you provided, Gillian, a good run rate now that Video Call transition will happen. When we think about costs coming down, but at the same time, usage for Video going up? Is it a good run rate sort of a margin to think of?

Yes. For -- I'll let Gillian answer the gross management question. For bookings growth, our users monetized over time. Generally, when we have a new user, there's some chance that they're going to sign up and pay for the subscription. But many of them, it just takes sometimes years to monetize. So we're just very happy with the fact that we have a much larger user base, now active user base, that will eventually try to monetize.

We're not giving guidance for '27 just yet. But I think the thing we're saying is we feel pretty good about our bookings growth, and we're starting to concentrate more monetization now that we have really managed to reaccelerate DAU growth.

Yes. And in terms of the gross margin, obviously, not guiding to next year, but -- yet, but if you think about this year, gross margin, we thought it would be about 2 points lower. The AI costs have not been as high as we thought, which is great, and we're still putting AI out, AI-based features like Video Call out to all of our paid users by the end of the year. So we feel really good about our ability to manage that.

I would say our appetite to give AI-based features to our customer base is very high. We think there are really creative things we can do with it. So sort of like you saw this year, we'll manage this to an overall plan, but again, not giving next year's guidance. I would just say we do have an appetite to give as much as we can to customers, and you've seen us do that this year.

Your next question comes from the line of Bryan Smilek with JPMorgan.

Just 2 quick ones from my end. Last quarter, we had talked about CURR improvements at all-time highs. And new cohorts are retaining better. Just curious any updates you have there, especially as the velocity of DAU growth starts to accelerate would have the overall curve trends look like on a cohort basis? And then, Gillian, just briefly on paid marketing as well. I believe up nicely year-on-year as you gain greater efficacy. Can you just talk about some of the LTV to CACs that you're seeing and where you're deploying marketing spend by channel?

Yes. I can talk a little about CURR. I mean we mentioned CURR last quarter, it is a metric that we're very proud of. And it is the main metric we optimize to increase daily active users to increase the retention of the app. We don't want to get into the business of guiding CURR and giving updates in CURR every single quarter. So we're not talking about the precise numbers, but it basically keeps increasing. And that is a majority of the DAU growth that you see comes from increases in CURR.

And then on paid marketing, this has been a real success story of our team this year. So just to put it in perspective, paid marketing is about 60% of our noncomp marketing spend, which is a bit over $100 million. We are profitable on that marketing spend, but we continue to believe we have more room to build upon the LTV CACs that we look at for the business.

In terms of where that money is spent, it's really pretty broad-based, and we move it around fairly actively in terms of what is creating good returns for us. And the team continues to do that really effectively. So I think the net on paid for Duolingo is it's probably never going to be the majority of how we go to market. That's just not how we think about it. But it's a really helpful tool and our team has done a really good job. And we're feeling that we're getting better and better returns on it and getting smarter on it. So it's been additive to us in 2026 and will be probably again in '27.

Yes. I should say for internally, this is one of the biggest wins we've had this year that we became actually intelligent about performance marketing. And I'm very happy about that.

Your next question comes from Mark Mahaney with Evercore.

Let's try 2 questions. I know this Video Call functionality as you're rolling it out across more of the Super Duolingo users subscriber base. What have you seen so far in terms of the boost that's had to engagement and retention or satisfaction? I would assume that it's had a material impact. It's one of the reasons why, as the costs come down, you're spreading it out. But what have you actually seen in the data so far? Is it too early to tell?

And then secondly, I know you talked about Chess earlier, so how about my second favorite topic, which is math. And how many users -- and I didn't see anything in here, but maybe it's in your press release, but the adoption you're seeing there, is there an opportunity there to also do a separate or combined price action?

Yes. Thank you for both questions. So Video Call, I mean, look, we feel really good about video call. It is a feature that actually teaches this conversation, and we see adoption of it continue to grow. The team just keeps making it better and better. And we see -- one of the ways we measure it is just a number of words spoken per active user, not practices or Video Call, but practices are of Duolingo. And that number just keeps going up. It's kind of -- I think it's roughly doubled over the last year or so. So we feel pretty good about that.

And the other thing is we know that when people try video call, during the free trial, for certain types of video calls that we tailored to the user's difficulty, then they are more likely to subscribe to actually pay for the subscription. So we feel really good about it. This is a feature that just keeps getting better.

In terms of math, look, this is also my favorite topic. I love math. I love our math product. It keeps getting better and better. It's also a huge opportunity. There's about 1 billion people in the world learning math. So I really think it's a huge opportunity. It is growing about at the same rate as languages. So I'm happy that it is growing, but I would hope that it can start growing at the rate that Chess is growing, which is growing faster. So we're going to continue making a better teaching tool. One of the things that we have found, by the way, is -- this is not that big of a thing to learn, but people prefer to play the game of Chess than to learn math.

They're both good.

I think they're both good, but hey.

Your next question is going to be audio only from the line of Omar Dessouky with Bank of America.

I wanted to elaborate a little bit on the gaming partnerships. You highlighted Brawl Stars and Eggy Party, which you said, brought millions of new users. So how scalable is this approach? Have you approached a meaningful percentage of like the top 100 mobile game developers out there? And what is the economic incentive for game publisher to collaborate with you, particularly that you're kind of competing for the same consumer attention? And then I have a follow-up along the same lines.

Yes. So we're very happy with these partnerships. I mean generally, for a lot of parts, people come to us as opposed to us having to go to them. I mean, just historically, people really want to partner with us. So we're very happy. We think we're going to continue doing some of these. Now it's not the case that there are hundreds of such partners that we're going to do. So we're seeing it more. And this is probably going to happen a couple of times a year.

And they do bring users, and we're very happy with that. But most of our events, if you look at what the plan for next year, most of our events are first-party events. These are our own things that are -- we're going to do things like Avatar suits and stuff like that. In terms of what is the incentive for the game publishers, there's no money being exchanged in these partnerships and we didn't pay them and they didn't pay us to do that. What we're finding with them with these partnerships is that it's a win-win for both. They get a bunch of users, and we get a bunch of users. So it doesn't seem like we're competing for that.

And generally, the incentive is just cross-promotion and for certain games, certainly the case for Supercell and Brawl Stars, they do a lot of things to keep their game fresh. And one of the things that they did when they did the partnership with us is for about a period of a few weeks, the entire games was just teamed with Duolingo, and the players like this. It was a very positive thing in their community. So that's the incentive for doing that. I mean they gained also a bunch of active users and revenue.

Okay. And my follow-up is, does this somehow dovetail with your performance marketing? Because as we all know, mobile games are a major driver of just the performance marketing within mobile gaming apps in general. So do you view this approach as sort of complementary to that? Are you looking to penetrate more of the gaming kind of audience and bring those over through performance marketing and using this experience to a -- how should we think about that going forward for you? Is it a pretty big audience.

Yes. I do think they're related. Because you're right, a lot of the performance marketing ads are for games. Certainly, the ads as we show in our product as well. So we do think they're related. We did see when these partnerships were going on, we did see our performance marketing actually benefit from the partnership. So we do see that. I don't think we fully understand the relationship with these 2, but I do think they are additive to each other.

Your next question comes from the line of Ron Josey with Citi.

This is [ James Michael ] on for Ron. Given experiments per week grew faster than head count last year, and Green Machine have driven 2 quarters of acceleration. Can you provide more color on the underlying changes in the architecture of the testing process? And how much more throughput there could be an AI to run more of the experiment design and analysis?

Yes, we're very happy about that. If you look right now, our -- the number of experiments that we ran, call it, the last month, is up about 90% year-over-year compared to last year. And our employee number grew by like 10%. So per capita, this is a major improvement. A lot of that -- not entirely, but a lot of that is because of AI. We're just seeing -- it took some time for the promise of AI in coding to actually turn into improvements in speed, but we are now seeing that.

And we're seeing not just significantly higher experiment count, which I believe will continue going up and up and up. but also the latency, the time between an idea to it being in front of the users is much shorter now. And that allows us to iterate a lot. And I think that's extremely compatible with our process of improvement because we just run a ton of A/B/B test.

And what's good about AI is it's particularly good at these smaller changes. Larger changes, a lot of times, it still kind of messes up. But the smaller stuff, it's pretty good at it, and we mainly run on hundreds and hundreds of tiny changes. So we're finding it quite compatible with the way we operate. And this is why we feel pretty good about our ability to continue growing both our DAUs because of growth in retention, but also our monetization because we're going to -- my sense is next to you. I don't know the exact quantity, but we're going to run a hell of a lot more experiments next year than we did this year.

Helpful. A quick follow-up on that. Last quarter, you flagged shorter attention span and tech shorter [indiscernible]. Has anything shifted here? And are you seeing down-man specifically from session frequency as well as retention?

Yes. We're working -- this is something we're really working on. We're not disclosing any of that, but we are actively -- we have staffed teams now to try to increased engagement throughout, and we feel pretty good about that work.

Your next question comes from the line of Justin Patterson with KeyBanc.

Great. Well, first off, congratulations just with the DAU progress this year. You've certainly come a long way within there. I guess as we head out of this year into '27, how do you think about just the right degree of friction and step to put back to the app and steps to monetize, so you don't repeat the lessons of 2025 within there? And I'll have 1 follow-up.

It's a great question. I mean we're -- we have a strict guardrail inside the company to not degrade the free user experience anymore. The good news is that we really -- we took this year to figure out how to monetize without degrading the free user experience. And I think we just found a number of levers that work that we feel pretty good about that. But yes, I mean, this is a lesson learned. We do not want to mess up the free user experience anymore. I don't believe that's going to happen.

Got it. And then thinking about just the decision to deprecate Max in this way, putting Max users into the Super Family plan. Talk about just why that's the right choice versus the individual family plan. Or excuse me, the individual Super plan. And then that creates the potential for more paid sharing, if you will, friends, potentially sharing with one another. So would love to hear about how you're thinking of the guardrails in there. So these are truly net new people versus some canceling and just getting on a friend's plan.

Yes, it's a good question. I mean the first thing to say is that at the moment, even though family plan is a pretty good deal when compared to the individual. Even though that's the case, we like it because it has extremely high retention. So over time, it just -- we like it. The retention is really good. And it makes sense that the retention is really good. If you have people as long as one of them is still using it, they kind of keep paying for the plan. So that's why we like it.

The reason we decided to move Max individual to Super Family was because they're roughly the same price. So that allowed us to not have kind of a onetime decrease in bookings from these people. And for them, we think it's a good deal. I mean, it's basically the identical features just you can invite 5 more people. And that will happen. I'm sure there's some amount of cannibalization, but we're not really seeing that in the numbers. It's not material enough for us to like see anything that is like, oh, this is actually bad for us because we've been A/B testing this for a little while. And we just it looks good.

And I think it speaks, Justin, to our confidence and the size of the market opportunity here. to be able to do that, and we're really not concerned that we're going to impact our market opportunity. We continue to be extremely bullish about what's ahead for us and the billion learners that we think we can touch, it's going to give us a much bigger business over time.

Your next question comes from the line of Alec Brondolo with Wells Fargo.

I think the other change this year to touring beyond the deprecation of Max is the light tier. I've been seeing lights here. I've been seeing light tier be kind of merchandise to me and my testing of the app pretty frequently in the third quarter and the beginning of the fourth quarter. So just any thoughts about your expectations around kind of light penetration next year? And the state of the advertising experience in Duolingo right now would be great.

Okay. So it's light on that. We are testing it. We have not pushed it very far, and we may try to push it more. I mean we're mainly driven by the results of our EBITDA. And I think our EBITDA for light have just not been home runs. I think there's probably some wins there, but they just haven't been home runs. So I don't actually know what will happen with light. It's a much cheaper plan that also has ads. I just don't know what will happen. It may be the case that ends up only being served in certain regions, kind of probably regions with lower GDP or capita. But I don't actually know.

At the moment, this is pretty small, and we just have not seen it be a major home run. What is there, it's still being tested. In terms of ads, we feel very good about ads. In terms of the team, we now have an actual serious team working on our own ads business. So we feel like next year, our ads business is going to do well.

And just to remind you about our ads business, we've had historically not invested much in our own ads. And so what we did is we just used these ad networks with kind of the lowest common denominator of ads. And what we're doing now is we are switching those to much higher quality ads. Now that takes some time to do, in part we're doing it through the direct sales, we now have a direct sales team. It's not a very large direct sales team, but it's a very good direct sales team. But we're doing some of that.

And also, we're just switching the ads to higher-quality ads. So the transition is going to take some time. But once we're done with the transition, you're just going to see ads that are significantly higher quality on Duolingo with kind of more well-known brands and higher quality brands and in addition, that pay us more. I think over the long term, we feel actually pretty excited about our ads business.

The team made some really nice progress during the quarter. So if you look at the full quarter, the ads business was sort of flattish but actually, it exited the quarter growing nicely. So we think the team is making good progress there.

Your next question comes from the line of Andrew Boone.

I wanted to go first in terms of following up on Justin's question. As we think about 2027, you guys clearly have transitioned from monetization now into engagement. As we go back to a monetization format, should we expect a change in terms of how you guys are going after that? Is that different in-app purchases? Are there any things that are now off limits? Like is it a restarting of the engine that was previously up and going? Or are you guys going to approach this differently?

And then just my second question, another follow-up. But as we think about AI unlocking Duolingo broadly, Luis, is the biggest needle mover for AI, is it internal? Is it new products that become available? And what happens as we've seen AI deflation kind of come through the model? Like what does that unlock for you in terms of internal processes that now become available maybe in a year or 2 years as we've seen with video calls?

Great. In terms of monetization, I mean, I don't know exactly how to answer your question in terms of like is it a restart? Is it -- I mean we've been working on monetization this year. It's just we told the monetization team, look, you cannot degrade the free user experience anymore. We have to come up with new ways of monetizing. And they have. And that's working really well.

So I think you're going to see us -- first of all, the way you're going to see us work is similar to how we've been working the whole time, which is we're going to run thousands of A/B/B tests to try to monetize better. The one thing that you're not going to see us do is add friction to the free user experience, but we're going to do all kinds of things. Certainly, there's IAPs in there. Certainly, there's better ads in there.

But there's also other things that get people to subscribe. Video Call is a good thing. I mean, Video Call for the longest time was just available in Max. Now that it's in Super, I think we feel a lot better about merchandising it more often because it's a lot more affordable. I mean the reality is that Max was a pretty expensive plan. So we feel pretty good about that. But generally, the type of work they're going to do is just a lot of A/B/B tests.

I don't mean to interrupt, but does that imply that conversion rates then don't improve on a go-forward basis? Because you guys then need to grow with users without additional friction?

I don't know the answer to your question, whether we're going to be able to improve them or not. Obviously, we do want to improve conversion rates. I don't know. I feel like there are some things -- some of the things we've done this year have improved conversion rates, things like longer free trials, improve conversion rates. And by the way, at the moment, we're not offering longer free trials to every user. We've kind of given them to some users.

We're going to expand that further. So I do believe there's some increase in conversion rates. I don't know by how much. But what is true is that what you're not going to see us do is kind of squeezing users. You're not going to see us do that.

Yes. And I think, Andrew, you already saw in the quarter, we reaccelerated bookings growth in the quarter relative to last quarter. And I think you're seeing us be careful about how we go do all that. One of the things we've talked to you all about is that top of the funnel matters a lot. Top of the funnel converts in a multitude of the amount that the core base does. And so that's why you've seen us work on things like paid marketing, some of these outreach things growing in new places and new products. And I think that's a piece of the puzzle to on an aggregate basis.

And then your second part of your question was about AI. We are -- there's 2 places and you mentioned the 2 places that AI really helps. One is just internal productivity. And the other one is AI features in the product. And we're very happy about both. In terms of internal productivity, like I said previously in the call, we're just seeing our team be able to ship A/B/B tests much faster. And we're getting to the point where simpler A/B/B tests are now being shipped by non-engineers.

So a person who's a product manager or a designer is now able to fully from idea to it's in the app. They're able to do it now by themselves without the help of an engineer. Now that doesn't mean we don't need engineers. We need engineers for all kinds of things, including the system. We built an internal system that allows us to do that. We're trying to make the system better and better. So we definitely have a major need for engineers.

But the fact is that significantly more people are able to run their own A/B/B tests. And so I don't know the precise increase in a number of A/B tests that we'll run next year, but it will be multiples of what we ran this year. I don't know if it will be 3x. I don't know if it will be 10x. I don't know because all this stuff is changing, but it will be multiples. And so I'm very happy with that.

And that, I think, is one of the things that we feel bullish about. In terms of the product, you'll just see us put AI in a lot more places. We're working on a number of features, particularly practicing speaking, that are not just Video Call, that you'll see there. And in addition to that, you'll just see improvements in how fast we can ship new teaching modules, let's call them. Like, for example, we're just going to teach a lot more Chess, a lot more content. We'll just be able -- we're able to just ship it way faster now.

Your next question comes from the line of Alexander Sklar with Raymond James.

Two-part question for me on gross margin. First, maybe for you, Gillian, just will the sunsetting of Max, have any impact on gross margin beyond what you've outlined in the fourth quarter? And then second, for you, Luis, the mix of direct payments outside the app store, any change to how you're thinking about that opportunity, either just as a result of the rise in personal agents to drive traffic more direct to you or just more than market testing?

So Alex, our guidance for the fourth quarter, which implies gross margin will come down just about point-ish. It is inclusive of the transition we're expecting on Max. I think as you've seen our team do all year, we're always have an eye to how do we get more efficient with our AI spend. But I think we also want to fit more AI in the product and then balance that to our bottom line.

So you will, we believe, see some decline in the gross margin sequentially, but still a very attractive number and we'll be able to put a lot more AI content in the product and do the Max transition as well.

Yes. And in terms of direct payments, we do -- for the longest time, at least in our apps, I mean, we have an iPhone app and an Android app and our website. For the longest time, you can only do direct payments on our website. We can now do direct payments in certain regions and for certain users for our apps.

We do use it in certain cases, and particularly one of the places where we're using it quite a bit is in performance marketing. When we acquire user performance marketing, we actually acquire them to a website and then get them to download the app. That's pretty standard practice. So you'll see us do some of that.

I don't know -- just judging just by the number of people we have working on this, I don't know if this is going to be a major driver for next year. I mean we do use them in different places, but it is not the case that at the moment, we're like we must switch everything to direct payments. That's not currently in the cards.

Got it. And as it relates to kind of the record Q3 gross margin, you wouldn't attribute any of that to the growing mix of driving traffic directly to yourself?

No. The main delta in the gross margin was better revenue and less AI costs than we expected going through.

Your next question comes from the line of Ross Sandler with Barclays.

Luis, I wanted to go back to the big picture with AI. So we can all see that AI is progressing at a pretty massive clip. The models are getting better. The latency is getting faster. Meta even followed you guys with kind of a sassy avatar launch recently. But the question I have is that, like there's some academic research that says one-on-one tutoring is sort of one of the best ways for anyone to kind of learn a new subject.

So you talked a little bit earlier about how you're going to integrate some of these more native AI interactive elements to the app. Right now, it's been like kind of an add-on with video. But I'm curious how you're thinking about like the long term, could these become much more core versus kind of the tap and learn gamified tap and learn use case that we started with at Duolingo. And how does that change, I guess, the business model for Duolingo as you integrate more of that?

Yes. So it's a great question when we think about a lot. We know today -- a few years ago, maybe 4 or 5 years ago, we did a bunch of research that shows that Duolingo was about as good as a classroom for language learning. We've gotten better to the point where I'm pretty sure we're better than a classroom now. But we are today not as good as a one-on-one human tutor. And nothing has gotten as good as a one -- a good one-on-one human tutor is still better than where we're at now.

We think that over the next few years, we will be able to be as good as a one-on-one human tutor. Because we just see how much we're improving every year. I mean, the thing about one on one human tutors is that they're not improving. They're like it's about as good as them, but we're improving. So we think we're going to get there.

Now there's something very important to mention, which is that, to a large extent, even though one-on-one human tutors are very effective at teaching, to a large extent, most people don't want to do that. I mean, there's a small fraction of people that are willing to do that, but most people just don't want to -- even if it's given up for free. It's just -- it happens that it's not that compelling of an experience on your mobile phone when compared to TikTok or Instagram or whatever. It's just not that compelling.

So what we want to do is we wanted to have something that is as effective as a one-on-one human tutor, and we think we can do it, but it will not look like a one-on-one human tutor. It will, in the end, still feel gamified because ultimately, the thing that is most important, even though one on one human tutors are very efficient and effective, the thing that is most important, the major failure mode is people just dropping off.

Like you can have a very effective one on one human tutor, but people just will be like, yes, I'm going to stop going to that. So what we're working on is it will still feel gamified, but you will see a lot more AI in there for us trying to meet you where you are. I mean, by the way, the reason one on one human tutors are better than a classroom or better than what we do right now is that they are really good at adapting to you. They're just -- they just -- and they do even things like they read your facial expression when they're saying something to you, they can see that you're like, you look confused. So you don't even have to say it and they quickly adapt to you.

That's why they're so good. And this -- I believe the technology is just getting better and better, and we'll be able to do that. It will take some time, but we'll get there, I think.

But then, Ross, the sheer number of users of Duolingo and the database we have puts us in a position to do the best that can be out there in the market. I think that's a really powerful thing that we have. The incredible number of people who use us every single day and working off that data to come up with the exact right way to teach them.

Yes. I mean that's a really good point. The data that we have in terms of people learning, I mean, I don't think anybody -- well, we know, we just -- nobody has the amount of data of people learning that we do.

Your next question comes from the line of Nathan Feather with Morgan Stanley.

Two on my end. First, when you talk about focusing more on monetization next year than you did this year, what does that actually mean internally? Is that more people dedicated to monetization pass? Is that changing kind of the hurdles for what you accept? How should we think about what that actually means?

And then second, just given the real success of the Chess course, how are you thinking about new course expansion, both in terms of the time line for when you might pursue that a little more aggressively? Then also the type of subjects you think would be applicable here may be different than what you had tried previously.

These are both great questions. What does it mean for focus more on monetization? I mean, this year was a unique year in that we told the whole company look -- what really matters is accelerating DAU growth. We have a huge opportunity ahead of us. We have to figure out how to accelerate user growth. And by the way, accelerating user growth is really hard. So hey, let's go there. So we put a bonus for the whole company, and we said, let's go and accelerate user growth.

Next year is a little different. What we're telling everybody is we continue to focus on user growth because it's such a big opportunity ahead of us. But you know what else matters, monetization. So I think you're just going to see more A/B/B tests related to monetization, and that's kind of how we're going to operate next year.

In terms of new courses, we're not announcing new courses at the moment. What I will say is when we added Chess, a AI helped quite a bit. And we were able to add Chess from idea to being in the app in about 9 months. What I will say is that we know that we can do that even faster now.

Your next question will be audio only from the line of Arvind Ramnani with Truist.

But nevertheless, yes, I just want to ask about sort of the inference costs and DAU is, right? So you've had a nice jump in your DAUs, up like 30%. What is that doing to influence cost per DAU, right? Is basically -- are you seeing your cost per year are coming down faster than AI usage for user going up?

Yes, we are seeing that. I mean we want to put a lot -- we have been putting a lot more AI in the product, and we are seeing cost per DAU come down quite significantly. The main reason is for the majority of our users, open weight models are about as good as frontier year models. There are some small cases where a frontier model is still needed. But for the majority of our users, the open weight models are plenty good.

So whenever we're able to switch from using a frontier model to an open weight model, the cost savings are pretty high, and we're able to do that quite a bit. So our team has done a really great job at being responsible about that.

Yes. No, of course. And I think you're now kind of at the forefront of it, it seems like that's kind of what the more experience or the more service uses of AI are doing as they start in the frontier model and they eventually migrate. Particularly if you kind of know if you have a level of sophistication. And I guess kind of the other thing is you want to put more AI in the free experience to drive DAU growth.

And I think you've cited at one of the big reasons you kind of the cost of video dropped like 95% or whatever. So the question I want to ask is, where is inference costs were free users stopped making sense, right? Like inference is you can deliver that very inexpensively. But it's a breakpoint where inference cost per free us really doesn't make any sense. And you say, like, hey, we're going to start giving this to the free user.

Yes. I guess the way I'll answer your question is that we are going to spend some on inference cost in free users next year. It's not a ton, but we are going to spend some on that because we do want to give some of these AI features to free users. But you call it single-digit millions that we're going to spend next year on inference costs for free users on kind of live inference costs.

Yes. And sorry, last question on this is. Are you still like -- and I'm not looking for the next quarter, but are you just a few years away where your business becomes large enough where you basically are going to own some of the compute? Or will you -- how are you accessing this, right? Are you using the AI cloud? Are you using the inference providers? Are you doing this internally? Like how are you accessing the inference or the compute to do this?

Yes. I mean it's all of the above. I think the majority right now is inference providers. But we're trying some internally. So it's kind of all of the above, but we'll -- we'll do what makes sense for the business. Yes.

Yes, classic A/B/B testing, which is always been very smart on that.

Yes. The team is really creative on this and has done a really excellent job at managing it.

Your last question comes from the line of Ralph Schackart with William Blair.

I jumped on light, so I apologize if you covered this. But you mentioned on the call that you're going to be running multiple more tests in 2027, they referenced like 90% year-over-year testing year-to-date. I'm guessing that's made on the product side, but can you sort of clarify all of that's on the monetization side. But maybe the more important question is, how are you able to run these tests next year at this scale?

And then also focus on the monetization. Do those experiments need to drop by, I don't know, a rate of 50% to 45% year-over-year growth. But kind of walk us through the path that you can continue to run these experiments and drive monetization next year?

Yes. So the way we improve the product is we work on 3 main things. We do improvements to we call it user growth. So that usually is improvements in retention of the product that increases daily active users. We do improvements in monetization and then we do improvements in teaching better. In all -- in these 3 buckets, the way we work is we basically run a ton of A/B tests in each 1 of them with the aim at improving one of these things.

So this particular A/B test is aimed at improving monetization. This particular A/B test is aimed at teaching better, et cetera. And we just run a lot of these. At the moment, like we mentioned, we're about 90% higher than, for example, the last month, we ran 90% more A/B tests than we did 1 year ago. We believe that, that will continue increasing.

So just overall, we're going to be able -- for monetization, but also for teaching better, but also for daily active user users. We're going to be able to run a lot more experiments next year. I don't know how many more because it depends on a lot of things, how our processes improve, et cetera. But it really is getting to the point where the bottleneck is you're starting to get cookie bottlenecks that we're not the thing that we used to normally think about bottlenecks. For example, a bottleneck now is ideas. I mean, like even coming up with ideas of what to test, we just -- we're -- that has become a bottleneck.

Another bottleneck is when for example, one of the things that we're able to do now is, like we mentioned in a previous answer, we're able to get non-engineers to actually launch an A/B/B test. So like a product manager or a designer is actually able. We built an internal piece of software that allows a product manager to just, without the help of an engineer, do an entire change of them. They're not large changes. They're small stuff. But we're able to do that.

But another bottleneck comes what if that person is a product manager, they introduce a bug to the app, who fixes it? Because I don't think they'll know how to fix this. So it's -- so this is the -- our processes are going to have to change. And we're in the middle of that, like a couple of the things, I don't even know the answer for what is the case, what we're going to do. But I do know that we're going to be able to run some integer multiple of the number of A/B tests that we ran this year.

I'm showing no further questions. This concludes the Q&A section of the call. I would now like to turn the call back to the host for the closing remarks.

Just thank you all for being here, and thank you for the great questions. And especially thank you for -- thank you to the team for the really great work so far this year. We're very proud of our results.

All right. So welcome, everyone. Let's get started. I'm Ron Josey. I lead Internet coverage here at Citi. And I am thrilled, psyched, excited to have with us today, Luis von Ahn, the Co-Founder and CEO of Duolingo. So Luis, thanks for being here.

Thank you for having me.

Duolingo is one of the, in my mind, most recognized consumer learning fun platforms out there. There's a lot going on. And so maybe we'll get started. And first of all, thanks for coming. So excited to see you. Thanks for being here.

I think most people know Duolingo, so we can sort of just jump into it. Maybe I'll start with some icebreakers. Just to tighten the mood, if you will. And the first one to lighten the mood is when I give you this one number or word like what's the first thing that comes to mind, 100 million.

That's the number of DAUs we want to get to in 2028.

Okay. That's the first thing that comes to mind.

That's it. That's our goal.

Chess.

Love it.

Okay. AI is like?

Exciting for Duolingo, scary for me personally because I don't know what's going to happen.

Okay. Green Machine, the name came from how -- where did the name come from? Green Machine?

Green Machine is our process of optimization at Duolingo, and it came from my Chief of Staff, Molly Brean. She invented it.

Yes. There was a -- I think there was a big wheel back in the day called The Green Machine as well. Okay. And maybe the last one, more fun, but Duo is so lovable, but at the same time, it's so demanding. So when you think about the character of Duo, what's your first -- when I say Duo, what do you think?

Well, Duo has a mind of his own, and he just wants you to do your lesson. That's all he wants in life. And so when you don't do your lesson, he will freak up. So that's it.

Yes. Well, he freaks out on me too often.

You got to do your lesson.

I need to do my lessons is sort of how it goes. I was debating whether to do this all in Spanish, but we'll stick to English for now. So we'll go through some questions. And of course, we'll open it up for -- from the audience for questions as well. So let's just kick off here specifically. And so it's been 8 to 9 months in sort of a strategy reset, if you will, around prioritizing engagement and user growth in '26 before we get back to sort of the monetization. Just take us back to late last year. I know we've talked to this a lot already, but take us back to late last year when you all were going through this idea of let's drive engagement. And walk us through sort of why do we do it?

And then more specifically, we're now 8 to 9 months in. And so what are you seeing in the business that has made this sort of like the right approach thus far?

Yes. So yes, about a year ago, we had, I don't know how many quarters, something like 16 quarters or something of incredible revenue growth. Just every quarter was just we were growing quite a bit for a long time. And it was doing well. But one of the things that happened is we are an incredible optimization company. We optimize for things. And it was -- it had become clear to me that we were spending most of our time optimizing for revenue, which I thought was shortsighted.

And what I thought the real thing we should be optimizing for is audience size, especially because we are at a unique time where because of AI, we really are going to be able to teach much better over the next few years. And I think there's a real opportunity of capturing an even larger audience than we originally thought. So I made the decision, which was not an easy decision to mainly concentrate on growing our user base, our active user base.

And so that's what we've been doing for the last few months. It's going really well where our year-over-year DAU growth rate is very good. We're very happy with it. And so we've been able to reaccelerate our growth rate, and we feel quite good about our ability to get there. I mean, right now, whatever the latest -- I don't remember the exact number that we said for the latest DAU numbers. I know the DAU numbers from yesterday, but I don't think I can tell you that.

I think we know them at some point in August.

Yes. We know them at some point in August. But in any case, we feel pretty good about getting to 100 million in 2028. It's looking really good. And the really good thing about our DAU growth rate is that it's really all regions are growing very healthily. And so -- and ultimately, the larger the audience that we have, the more people are learning on Duolingo, the more they'll tell their friends, so we keep growing. But also the larger the audience we have, the more revenue we'll ultimately be able to make. So I think this is just going to be a much larger business than we originally thought.

Luis, that was super helpful and insightful. And in that, I think you said we want to teach better in the next few years or something along those lines. Talk to me about dive deeper into that teach better over the next few years because the product works great. Duo brings me in, maybe not every day because he gets mad at me. But just talk to me what do you mean by teach better or give us any examples to help us understand what that means?

Yes. So a couple of things about teaching. First of all, one of the things that we learned a long time ago was the hardest thing about learning something by yourself is staying motivated. That if you're going to learn something by yourself, it's different than learning something in school. Learning something in school, you kind of have to go to school because your parents are forcing you. It's just -- you kind of have to go. If you're going to learn something by yourself, especially on a phone, it's pretty hard to stay motivated because you are one tap away from going to TikTok or whatever it is.

So the main thing that we have concentrated on for many, many years is getting people to stick around and stay motivated. And we have done a really good job. If you, for example, look at our DAU of our MAU ratio, it's incredibly high. It's better than most games actually. So we've done a really good job of getting people motivated. Now the other thing, of course, we have to do is make sure that we are actively teaching you as well as we can. We did a study a few years ago, and Duolingo was about as good -- I felt a few years ago, Duolingo was about as good as a classroom. And we felt really good that Duolingo was about as good as a classroom.

Now it just turns out that the reason that was the case is because classrooms, particularly for languages, are just not particularly that good. So we were about as good as a classroom a few years ago, and we're like we feel great, but it's like it turns out classrooms are just not that great. What's really good about learning at actually teaching a language is a one-on-one human tutor. By the way, this is true for most subjects. If you want to learn something really well, getting a one-on-one human tutor is usually the most efficient best way to...

Converse yourself.

And with the tutor, like it actually works really well. We are not yet as good as a one-on-one human tutor.

But every year, we get better at teaching. If you just -- we do measurements about how well we teach. And my sense is that in the next -- I don't know the exact time line, but in the next, call it, 2 to 3 years, and again, I could be off on this, but in the next, call it, 2 to 3 years, I think we will be as good as a one-on-one tutor in terms of learning outcomes. And I think that's incredible for a couple of reasons. First, we're going to be way cheaper than one-on-one human tutor. I mean one of the problems with one-on-one human tutors is they're quite expensive. If you want a good one. This -- we're talking $100 an hour easily if you want a good one. So we're way cheaper than that.

But secondly, most importantly, the problem that one-on-one human tutors have is that for most people, they're very boring. So they just don't do it, whereas we are going to be as effective as a one-on-one tutor, but it will feel as fun as a game. So I think that's a real game changer in terms of -- it really will be kind of a no-brainer to use Duolingo to learn certainly, languages, which is where we're most advanced. But I think we're going to get there as well with chess and with math in terms of teaching as well as one-on-one human tutor in the next few years. And I think that will be a real game changer.

So within Green Machine, and I'd love to have any insights on just the data on the path to that 2 to 3 years we get to that one-on-one tutor. Any insights or maybe aha moments on the product that you've seen that says, ah, this is something that will bring people coming back and get those streaks going. Anything in there like...

For coming back, yes. I mean it is an incredible thing when you look at the behavioral tricks, I'll call them, that get people to come back because we have data from so many users, I mean, really literally millions of users every hour using Duolingo. We have data from that, and we use that to get people to come back. And there's just some pretty amazing things. So for example, it really matters what you say in the notification to get people to come back. This is why the owl, for example, has a personality. It turns out that the awl being disappointed in you really works well. I'll be motivated, yes.

Well, it really works well. And by the way, how well that works depends on the culture. In certain countries, disappointment works a lot better than in other countries. So this is all the stuff that our AI finds out or either AI or our team just we find out that in certain geographies, certain types of motivations are better than others. Another really good example is the -- if you start a lesson on Duolingo, the progress bar is actually not linear. It's in fact, what happens is the first few exercises you do, it grows a lot.

So very quickly, you get to halfway. And the thing about getting to halfway is that humans just have an instinctive need to fill a half filled progress bar fully. So we just get you very quickly halfway and then you're like, well, may as well finish the lesson because I'm halfway. Now you're actually not halfway. We lie to you. You're actually just, I don't know, 20% of the way, but you feel like you're halfway so you may as well finish. So there's just -- there are thousands of these little tricks to try to get people to come back. Because ultimately, what we're doing is we're doing something that you say you want to do, but that ultimately, you don't really have the self-discipline to do.

So we're trying to just give you that self-discipline. I mean it's kind of similar to we're just helping you to go to the gym or we're just helping you to do the thing. So we're helping you to stay to be good for yourself long term.

Yes, I would agree with that. And you're building a lifelong skill upon doing it. So I am certainly guilty of the halfway, let's finish it up. So that's -- I try to go faster on the Spanish ones because that's easy for me. But anyway, let's maybe talk a little bit more just on that 100 million path and bring it back into 2Q.

Specifically, I think you said we talked about DAU growth is accelerating. I would love to talk a little bit more about what are the drivers of that accelerating growth. And the question we get is the drivers and the durability. How do you continue and I'm not asking about it I would love to -- but how do you just -- what were the product changes that sort of got to this level of accelerating growth that you laid out last year when you said we need to get to this metric?

Yes. I mean -- so when people ask us, is this durable? I mean the answer is an easy yes because these are not one-off things or anything. These are just simply the product is stickier. We have done a number of changes. By the way, in terms of the acceleration, the majority of the reason why we've accelerated and why we're beating our own internal forecast in terms of DAU growth is because the product is stickier. So we've just done an incredible job in terms of running tests that make the product stickier. We -- every single week, we run hundreds of A/B tests to try to make the product stickier.

Now some of them don't work, some of them work, about half of them work. It's hard to attribute to a single thing that why we're stickier, but it's just little things about like we tune the time in which we're going to send a notification. We change the animation of when you get a streak, how explosive it looks. So it's all this stuff that we do, and it just becomes stickier and stickier. At this point, we had it in our last shareholder letter. This metric that we look at internally is called current user retention rate, which is basically if you are here today and you are here another day in the last 7 days, what is the probability you'll come back tomorrow.

That number for us now is 84%, which is extremely high. And it's at an all-time high. It keeps growing. It grew by about a percentage point in the last year. And it's because the product is just stickier. So that's probably the biggest reason. There's another reason that is less impactful, but another important reason is our marketing has done an incredible pivot over the last year.

I think, hopefully, a lot of people are familiar with our marketing. It's pretty unhinged if you go to our TikTok or Instagram, like our social media has been pretty unhinged. And it has really worked very well. But over the last year, we've done a pretty incredible pivot where it remains unhinged, but it's a lot more brand aligned to kind of efficacy and teaching well, et cetera. So we're just seeing that the marketing is doing better than it was about a year ago. And in addition to that, we've diversified our marketing. It used to be the case that the main thing that we only looked at was we just had a really popular YouTube, TikTok and Instagram channels.

We now have diversified. It's not just our own channels. We have thousands of user-generated content that is helping like other people talking about us, that's helped quite a bit. And in addition to that, we've started doing some performance marketing. It's not a huge amount, but we've just gotten much better at performance marketing. And I think the combination of this just makes our marketing more effective. And so that's why we're accelerating.

Yes. Super helpful. One of the things that I wanted to ask a little bit more is I think the comment was you run hundreds of A/B tests, and we'll see what works, maybe it's a 50-50. Talk to us a little bit more just the internal process to run those tests. So specifically, I was a product manager a long, long time ago at an old company called AOL. And we would.

I was the user, what's the year?

Most people were [ back and ] they're dead. And now everyone knows how old I am. But we used to very much manage every single release. And then, of course, you just measure everything. And so talk to us more about the hundreds of A/B tests. I'd love to hear about how that has maybe ramped up over time. And how do you manage that process? AI, I'm assuming, has really helped to understand what's coming through.

Yes. I mean, so we have -- there's a new version of Duolingo every single week. So our iPhone app and our Android app. Every single week, there's a new version of Duolingo. That version of Duolingo, every time it comes out, comes out with some large number of new A/B tests, hundreds that we're testing. The way this works is we have -- there's 3 big things we work on Duolingo. One is user growth, another one is revenue and another one is teaching better, whatever the subject that we're teaching.

So all A/B tests are based on one of these. So they're trying to do one of these things, either it's meant to teach you better or it's meant to keep you more engaged with the product or it's meant to make more money. There's large -- each of the teams that own one of these is about 100 people. So it's about 100 people working on Duolingo that are either engineers or designers or product managers that are working on teaching better and about 100 that are working on monetization, et cetera. And the way it works is the combination of these 100 people, one of them comes up with an idea for an A/B test. Of late, it's either the 100 people or AI also comes up with ideas for A/B tests.

Then we sketch it. It gets reviewed by a group of people in the company that reviews all changes to see if this is a good change that we're going to do or not. Then it gets built. These days, it's mostly built by AI, and then it goes into the product. From idea to product, it's about a month that it takes from idea to product. And then we run it and literally about 50% of them are successful in whatever the metric they were trying to improve. So either it improves revenue or improves how we teach or it improves daily active users. That's kind of what happens. But the velocity that we're getting A/B tests out is very high. By the way, the number of A/B tests that we run per capita, so per employee of Duolingo that is working on the product has significantly improved over the last few months. And that's a lot thanks to AI, not entirely, but a lot thanks to AI. It's not quite doubled in the last few months, but it has -- it's close to doubled in the last few months.

And I think we saw that in 2Q with The Green Machine testing optimization came through. And also in 2Q, you benefited from maybe not test, but events, streak revival being one of them, I think, video calling now available in Super Duolingo. Was that part of the testing as well? Or that was the idea and it just sort of stuck? And I had one more question on The Green Machine, then we'll move on.

The streak revival was just an idea that one of our product managers had. Oh, maybe we can -- there's a lot of people who have lost their streak. Maybe we can have a onetime event where we let them revive their street. It was very popular. We had about 15 million people revive their street worldwide. That helped for sure. But that was a minority. It sounds like a really big event. But in terms of when we do the actual calculations of how much this helps with DAUs, it helped, but it was a minority. The majority was just long-lasting changes.

Got it. When we think -- now we'll take a step back, I guess, and talk bigger picture. And I know this year is not the year we're maximizing monetization, but it's helpful to hear those 3 sort of buckets, if you will, between user growth, revenue and teaching better. Talk to us about how you think you balance those 3? Or maybe I get the question all the time, how do we get back to revenue growth, right, if we're optimizing user growth and -- I guess, user and teaching growth here in the short term?

I guess that question only from Wall Street. So let's see. This year, our goal really has been -- from the beginning of the year, we decided we were going to set some financial kind of box or financial boundaries that we're going to say, look, this year, we're going to grow whatever we said at the beginning of the year, bookings by something like 10% or something that we said something. And we're going to operate like this with this year, we're going to have adjusted EBITDA in this range, et cetera. And subject to that, we're going to try everything we can to try to increase our audience size.

So that's how we're operating this year, and that's what we've been doing. I mean, in fact, every quarter, you've seen we're within the box and trying to accelerate our user growth. And I think it's been very successful in accelerating user growth. In terms of kind of monetization, we're not giving any actual guidance for next year or anything. But what I will say is that we'll probably have not just be concentrating on user growth next year. I think that's safe to say that we understand -- obviously, this is a business, and we understand that the ultimate goal is really revenue, but we really thought that this year, we should spend it on user growth.

As we think about different forms of monetization, the question that I'm sure you get all the time in terms of us and we do as well is just the tiers of subscription. So we've made changes to Max, of course. We've made changes to Super. We've included Video now in different tiers. Talk to me about how you think the subscription evolution, how we evolve subscriptions going forward?

Yes. So historically, we've had one subscription tier, which was Super Duolingo, which is basically just a premium version of Duolingo. A few years ago, we added a new tier, which was Duolingo Max. It was a higher tier. It was about twice the price of Super Duolingo. One of the reasons we added a higher tier was -- there are 2 reasons. One is we saw there actually was a desire to pay more or there were people who are like okay with paying a lot more.

So we're like, okay, well, let's have a more expensive tier. The other reason we added it was a couple of years ago, we introduced this new feature called Video Call, which is basically being able to practice conversation with an AI. At that time, each time somebody did a practice of the conversation with an AI, the call cost us $0.30. So that's what it costs. So we pretty quickly did a calculation. We're like we can't really put this in Super Duolingo because it's not expensive enough and then with a calculation.

And basically, we need to charge about twice as much for that to make sense in terms of our margins. So we put -- we added Max and this tier, and then we put Video Call in it. And we're like, if you buy Max, you can have access to Video Call. We said back then -- if we manage to decrease the price of these, we may start putting that feature in lower tiers or maybe even give it for free because ultimately, what we want to do is teach as many people as we want. I think that's really what's best in the long term for the business. What happened is that over the next few years, our team did an incredible job mainly by switching mostly to open source AI models to decrease the cost of the Video Call.

At this point, the cost is well below $0.01 each time you do it. So it had -- it went from $0.30 to well below $0.01. And so we now are able to put video call in Super Duolingo, which is the kind of the cheaper tier that we had, and we've started doing that. And today, not quite all of our Super Duolingo subscribers have access to Video Call, but at least all new super subscribers have access to Video Call and some fraction of kind of older super subscribers we've started giving a Video Call. That calls into question, what the hell is going to be Max or are you going to keep Max? And the answer really is we don't know. We're not entirely sure what we're going to do with Max. It's still there. People are still paying for it, but we're not entirely sure there are some options.

One option is that Max actually does go away. Another option is that we start -- maybe we give unlimited access to Video Call and Max and limited access in Super Duolingo, et cetera. So there's some options. And we just don't know. And the reason we don't know is not because we're sitting there kind of thinking about it, it's because we're actually running tests to try to figure out what is actually revenue maximizing for us. And we're running a lot of tests with different cohorts of users to try to figure out what it is -- what we do. And in addition to that, if we just turn off Max today, that would involve a hit in revenue, so we can't quite do that. So we're trying to be very thoughtful about this. Ultimately, we're going to -- what I'm pretty sure of is that we're going to be able to -- whatever it is we do, whether it be shut off Max or whether it is that we end up doing some changes to how Max works, whatever it is we do, I'm pretty confident we're going to be able to do this in a way where we don't lose revenue.

So I think I'm pretty happy with that. And if we're able to do that, that's really great because that suddenly, we are not losing revenue. I think we're making the same amount of money as we were before. That's awesome. And we suddenly gave access to Video Call to 10x as many people. And that's good for us because if 10x as many people have access to Video Call, that means 10x as many people are learning better, which means that there will be more word of mouth for Duolingo because they'll tell their friends , look, how much Spanish I learn on Duolingo, et cetera. So it will just be good for us if we're able to do that.

Talk to us about the adoption of Video Call. How important is it to the subscribers that are using it? When I say important, how often are people coming back to it as is that the killer app or a killer feature that is really driving usage?

Yes. So Video Call is a very popular feature. When you have access to it, you continue using it. What is clear is that learning a language requires learning SWRL, speaking, writing, reading and listening. Without video call, you pretty much -- it's very hard to learn speaking. You can learn some pronunciation. It's very hard to learn actual speaking. So anybody who actually wants to learn a language really, really likes the feature because it is the difference between not really learning how to speak and learning how to speak and people are able to do this really well with Video Call. So the feature is very successful itself.

I should have asked in the icebreakers, open source, closed source and just leave it there.

Open source all the way.

Open source all the way.

And by the way, that does not mean that today, 100% of our AI usage is open source stuff because there are cases where the frontier models are better than open source. But what we're finding is that for most of our use cases, we can substitute for the open source, and we are very happy with that.

And do you see a similar $0.30 to $0.01 sort of like result here?

It depends on what the usage is. It may not be -- I mean that's pretty striking. I don't know if that's like a 50x or something. It's pretty striking. I don't think we're seeing a 50x, but we're seeing pretty good improvement.

Yes. Yes. I mean, stark, it's unbelievable, right, from $0.30 to $0.01.

It is, it is. Two things happened there. Open source is one. The other thing is that just everything got cheaper. Everything got more efficient in terms of the AI spend. So I think those are the 2 things that happened.

It's funny in this world of AI spend, we hear about efficiency and then we see sort of the opposite at this point. But yes, that is the goal.

We are getting significantly more efficient.

Yes. So I'm going to come back to that in a second. I did want to ask one more on the subscription side, the free trial. So I think there are some -- you're always running and optimizing what works, what doesn't work. 7 days, back-to-school, 30 days. Talk to us just about the -- what are you seeing -- what you saw from 30 days? And how did that or does that -- is that a reason why the CURR is doing so well?

Yes. So we historically have had a free trial for Duolingo. Historically has been a 7-day free trial. This year, we are switching most of our free trials to be 30-day free trials. What we're finding is that, that is actually good for revenue as in more people end up subscribing. And it's also good for DAUs because you put some of these people in the 30-day free trial, they become more hooked to Duolingo because they can use it more, no ads, et cetera. So it's actually -- this is an incredible change that makes it so that it's good for DAUs and good for revenue.

I think that has helped a little bit on the CURR, but I don't think this is the biggest reason why our DAUs are growing. It's probably a little bit. It's not the biggest reason. But what we really like about it is that it's just better for revenue. And one of the things that is important to understand about this when we switch from giving 7-day trials to 30-day trials, there is an amount of bookings that gets delayed a little bit because when you -- if you have a 7-day free trial, you started today, you pay us in 7 days, whereas with 30 days, you pay us in 30 days.

In that time, that bookings -- those bookings get delayed for us. So we've been careful about not just doing that all at once because that probably create a little bit of a bookings hole. So that's what we're doing. But ultimately, we'll make it so that most of our users are going to be getting 30-day trials. We may start trying longer than 30 days. So it may be 2 months, it may be 3 months. For example, Spotify gives a 3-month offer. I don't know what will end up winning, but these are all tests that we're going to be running.

When you did the 7 to 30, were there tests along the way? Why was 30 the right number? Was it 21?

Yes. We have definitely tested a lot of different numbers. 30 is a nice number that gets people like, oh, 30 days. And for some reason in people's mind, that it's a free month.

And 3 months is...

3 months sounds nice.

A quarter.

Yes, yes, yes.

Let's see. We have about 5 minutes left. I've got a ton more questions, but if there's any in the audience, if that's okay, I would love to open it up. Is there anything in the audience? Any questions? If not, we'll keep going here, at least tee up your question.

I used to be a professor. You got to wait about 60 seconds for people to feel bad for you...

That's clearly not enough.

People eventually feel bad for you after about 60 seconds.

We're just on the clock here. We got a packed schedule, this is a treat to have you.

So let's talk a little bit more just about beyond languages, if you will. Chess, math, music, all coming on here. I think Chess added some additional capabilities more recently on seeing what maybe you should have done in game matches or whatever. Just talk to us about the road map and how the -- what Chess got right and how it informs newer things coming down the path.

Yes. Chess is -- I mean, we have other subjects. We have math, music and chess. These are the 3 other subjects that we do on top of languages. Chess definitely got something more right than the other ones. Chess is the fastest-growing subject. It is really -- I mean, it is growing -- it continues growing pretty fast. It is still growing faster than including languages right now.

Probably the biggest thing that Chess got right is that it happens that it's a game and people like playing games. Now we're not -- people sometimes ask us like, well, are you going to now start teaching poker? Or are you going to start teaching like Mahjong or something? We're not going to do that. We're not a gaming app. We are an education app ultimately. The reason we added Chess is because it really actually is good for your brain. There's just -- there's a lot of studies that show that if you are a student and you do chess after school, that's like the best thing you can -- other than doing your homework, the best thing you can do after school is learn Chess to improve your grades, et cetera. So this is why we have Chess. It's -- we're all positively surprised about how well it's going.

I'm very excited about math. The math has a huge opportunity. We're not -- I mean, we do have millions of active users, but it's still a small number compared -- in math compared to the opportunity. There's about 1 billion people in the world learning math. And we're pretty excited because literally, every month, it's just much better at teaching math. If you haven't tried our math course lately, it's quite good. If you were not very good at math in the tenth grade or something, you should try it because it's quite good. So I'm pretty excited about that. And I think it will -- this just opens up a lot more audience than just languages. Although my sense is that for the foreseeable future, languages are going to be the biggest thing. It's just -- there's a lot of people wanting to learn languages and a lot of people wanting to learn languages outside of school.

Yes. Well, we communicate with one another every day. So that's helpful.

Yes, English, in particular, if you want to learn.

So maybe we'll keep going here on just teaching better. And so teaching better, teaching quality, the Video Calls we talked about, all that is coming in here. What I want to try to understand is the link between teaching better into just greater usage and the ability to pay more. And so I know we've gone through a lot in this 35 minutes or so, but would love to understand what -- if there's a secret -- not a secret, but some killer app that has helped the learning better drive those -- that DAU growth.

So it's funny about teaching better. I wholeheartedly believe that we need to teach better. And if we teach better, our word of mouth increases because you see more and more people that are like, "Oh, look, I now am able to watch Netflix in French entirely because of Duolingo." I mean if you're not teaching well, people just can't say that. So -- and the more you say that -- the more people say that, the more you grow.

So I wholeheartedly believe that teaching better will increase our audience. That said, our evidence that teaching better increases our audience is not as tight as you would like it to be as in like we run many A/B tests that, for sure, teach better, and we don't see increases in DAUs. I think part of the reason for that is that it's very hard for people to know -- people don't realize when they're learning better, at least in the short term. They may realize that after a year passes, like, "Oh, wow, I can't believe I can now speak French."

But on a week-by-week basis, if we are a little better at teaching like people, a lot of times just kind of don't realize it. So I do believe the link is there. It's just much harder to prove certainly on the kind of A/B test that you're only going to run for 2 weeks. But I think that this is really a major unlock for us, and this is why we're continuing to work on this.

We have about a minute left. We haven't talked about AI as a risk -- and I'm not going to -- I don't want to end on it, but I do want to hear, we're seeing DAUs reaccelerate. There's been this concern for a few years around AI risk, but frankly, we're not seeing it in the numbers. Talk to us about how you view the environment here.

Yes. I mean I've never really thought that AI is a risk for us. For a while, the market kept on asking us like basically investors. I don't know, if you look 9 months ago or so, they were like the questions we would get from investors like is AI going to kill you.

They've stopped. I'm glad because I don't think -- I mean, it just -- there's just no real evidence that that's going to happen. Our user growth keeps accelerating. And I think the reason for that is, usually, when people say is AI going to kill you, there's kind of 2 buckets that people used to say that for. One is, are people going to need to learn languages because language translation is really good. I think people are going to continue learning languages. If you look at our users, half of our users are hobbyists. I don't think they care whether language translation is really good or not. And the other half are learning English, and they actually want to learn English. I don't think they'll care that you can translate.

By the way, language translation has been really good for like 10 years. And in that time, we have, I don't know, 100x our users. So I don't think that that's the risk. Then there's the other one, it's like, well, can somebody just like tap one button and create all of Duolingo comes out of the box. And I don't know, that may be the case that it will happen, but it's -- we have unique data that allows us to teach better and keep people more engaged, et cetera. So I just -- that's something else that we're just not particularly worried about. So I'm glad that people have stopped asking about it because I'm like, yes.

Yes. Well, Luis, this is wonderful. Thank you very much for the time.

Thank you very much for all the questions.

Congrats on all the success. Thank you.

Good evening, everyone, and welcome to Duolingo Second Quarter Earnings Webcast. Today after market closed, we released this quarter's shareholder letter, a copy of which you can find on our IR website at investors.duolingo.com.

On today's call, we have Luis von Ahn, our Co-Founder and CEO; and Gillian Munson, our CFO. They'll begin with prepared remarks before we open the call for questions. [Operator Instructions] And please note, this call is being recorded. [Operator Instructions]

Before we begin, please note that we'll make forward-looking statements regarding future events and financial performance. These statements are subject to risks and uncertainties described in our SEC filings and are based on assumptions we believe to be reasonable as of today, and we undertake no obligation to update them. We'll also discuss both GAAP and non-GAAP financial measures Reconciliations can be -- of the two can be found in our earnings materials, and we encourage you to review them when evaluating our performance.

And now I will turn it over to Luis.

Thanks, Debbie, and thank you all for joining. Q2 was a strong quarter. DAUs grew 23% year-over-year, accelerating from Q1 and coming in slightly ahead of our expectations, and we're encouraged by what we're seeing so far in Q3. The vast majority of that growth came from the work we do every day through what we call the green machine. We test hundreds of product changes, measure their impact and double down on what works. Most changes are small, but they compound over time. As I discussed in our shareholder letter, that's also what's driving CUR, a measure of user retention to an all-time high.

Another highlight of the quarter was Streak Revival. A onetime campaign we ran in June. The idea was simple: give learners who lost their longer streak. A chance to earn it back by completing three lessons. More than 15 million learners survive their streaks. And what's particularly encouraging is that these users are also showing better retention than a typical reengaged cohort. We brought learners back to a product that keeps getting better at teaching languages, chess, math and music [ understanding ]. We're still early in executing our strategy of prioritizing user growth and teaching better, but Q2 gave us more confidence that we're on the right track.

With that, I'll turn it over to Gillian.

Thanks, Luis. Welcome, everyone. As Luis said, Q2 was a strong quarter. In addition to the DAU acceleration Luis just mentioned, [ timing ] results were in line with our expectations and profitability was slightly ahead of our plan. As we look at the remainder of 2026, I want to reiterate how we're managing the business. We are investing deliberately in the opportunities that we believe can make Duolingo a significantly larger business over the long term.

At the same time, our team continues to operate with discipline. We continue to execute to our full year bookings and revenue target ranges of 10% to 12% bookings growth and 15% to 18% revenue growth. We have increased our target adjusted EBITDA outlook to 26.5% from the 25% we outlined at the start of the year.

As for point estimates to help you build your models, please keep in mind [ that ] the following: for the full year, we expect bookings growth of approximately 11% and revenue growth of roughly 16%. At constant foreign exchange rates, from our last call, the bookings growth rate would be about 0.5 point higher. For gross margin, we now expect to end the year closer to 70% as compared to the 69% we initially expected as we drive more AI content into our products offset by AI cost savings.

We expect adjusted EBITDA to be approximately $320 million at the margin of roughly 25.5% I just mentioned. And we expect to generate over $375 million of free cash flow this year. While it is not included in our 2026 guidance, we do want you to know that the company has a bonus plan that will trigger if Q4 DAU growth is 25% or higher and would be paid out during Q1.

Since it's currently uncertain whether that threshold will be met, we have not included it in our 2026 guidance. If it were achieved, we would expect the payout to be roughly $10 million in cash, potentially higher if DAU growth is higher.

For Q3 itself, we expect bookings of approximately $307 million or growth of 9%. Revenue of $302 million, representing growth of 11%. We expect gross margin to be 71% and adjusted EBITDA of roughly $76 million, representing a margin of 25.2%.

Our balance sheet and cash flow potential remains strong. We ended the quarter with $1.3 billion in cash and investments and generated $79 million in free cash flow. We repurchased about $44 million of stock during the quarter, bringing cumulative repurchases under our authorization to $72 million or approximately 700,000 shares.

Putting it all together, our user momentum is strong, our business model continues to generate significant cash flow, and our team is executing well in an important investment year. We remain focused on reaching 100 million DAUs in 2028, and we believe the path there can create a significantly more valuable business for our shareholders.

And now I'll turn it back to the operator, and we're happy to take your questions.

[Operator Instructions] Your first question comes from Wyatt Swanson with D.A. Davidson.

Given DAUs are now expected to be above that 20% year-over-year growth in the second half of the year, could you maybe give some color as to why the full year bookings guide wasn't raised more? Like I realize you're in experimentation mode, but [ sort ] a larger amount of users on the platform technically translate to increased bookings even if you're not pushing for monetization.

Yes, I think that that's a great question. I mean the first thing to note is that our users don't monetize immediately. I mean some of them monetized -- it takes a while for them to monetize because we have this premium model. We do expect that higher DAUs will imply higher revenue. But it's going to take some time.

And the second reason is exactly what you said. We said at the beginning of this year that we were going to operate in this box in terms of revenue, and we are going to continue with this box. Basically, roughly 11% year-over-year bookings growth and then the rest of the efforts are in increasing DAUs and in teaching better.

Got it. Okay. That's helpful. And then you noted with the extension of free trials that helps improve both engagement and monetization. Could you just talk to some of the underlying mechanics as to how exactly that works going from 1 month to 2 months?

Yes. So the -- we're -- just to put it into context, our monetization team, their goal this year is to find things that monetize that are not at odds with user growth. I mean, historically, some of the ways that we have monetized have been by adding friction to the [ freeze ] of product and that is at odds with user growth.

So one of the things that has worked the best is longer free trials. And in particular, the main one that we're trying is historically Duolingo -- free [ trial ] in Duolingo and we say, "Hey, you can try for free. It's been 7 days. So we give you 7 days for free.

That's what historically has happened. We are now shifting most of our free trials, not quite all of them yet to a 1-month free trial.

And what -- the way that works is basically significantly more people decide to take that trial because it's just a better deal. And because of that, we got a larger number of people actually converting to payers. And the other nice thing is that as soon as they agree to go on the free trial, the experience just gets better because we turn off energy, we turn off the ad. So this actually increases daily active users as well. So we like it very much, and it's having a lot of good traction.

Your next question comes from the line of Andrew Boone with Citizens.

I wanted to ask about just your role in terms of influencers, how do we think about international marketing and kind of the changes that you guys highlighted in the letter? And then secondly, as we think about the U.S., Luis, understood the strength in the quarter and kind of resurrected users. Can you talk about top of funnel though? How do you feel about trends there and kind of the broader opportunity of attracting new users that may be new to Duolingo?

Yes. So I should say our marketing team, just to put it in historicals, most of our growth has been organic through word of mouth. In fact, for the first several years, it was 100% organic [ word ] of mouth. Then we added mainly one form of marketing, which was social, our own social media accounts. And that's still going really well.

We are -- I mean we're getting more than 1 billion impressions per quarter on our own social media accounts, which is incredible. But our marketing team is really starting to expand to having other important tools. One of them is creators like groups of creators. And that's working quite well, actually, specifically in certain countries, countries China, Indonesia and India.

Something like 2/3 of our total social media impressions come from influencers. And what we're finding is that we are able to not spend a ton of money on this and what's nice is that these people have different audiences than us. And so this really just brings in a lot of new users.

And this is something that we're very happy with. We use influences differently in different countries, some countries more than others. But that's working really well. I should also mention that our marketing team has just become significantly more sophisticated on performance marketing. So while still the majority of our growth is organic, we're getting good traction on both usage of influencers and performance marketing. And we are seeing that in top of funnel. So it is increasing. So we actually feel pretty good about our top of [ funnel ]. It is increasing.

And you asked about the U.S., in particular, the U.S. growth has increased quite a bit in the last quarter. and that's something that we're happy with. By the way, if you look at our growth, our DAU growth is really broad-based. Basically, all regions are growing and their all [ regions ] are growing faster than they were before. Asia is still the fastest growing what U.S. has increased.

One thing I might add to that, Andrew, is the top of the funnel is part of the story of the quarter. It got better. As you know, that's been a focus area of ours. We've talked about how we want to get better there. And in almost every region, the rate of growth there improved in top of the funnel in the quarter, which is a great accomplishment for the team.

Your next question comes from the line of Nathan Feather.

My end, [ both ] on the monetization angle. I guess, Interested to hear with the learnings you've seen so far from giving video call to new super subscribers? And how is that influencing your pens for video call and MAX generally as we go from here?

So video call is an interesting thing. We're -- I mean, we love this feature. It's an excellent feature for learning conversations, practicing conversation. It really works. We have, in fact, research that shows that if you use video call, you get better conversation. So it really works, it's really good.

When we first started adding video call to the platform, the first time that we put a video called in the platform, I remember the team that was working on it told me, okay, we can give this to users, but it's going to cost like $0.30 per call. to give to users. And that was expensive. And this is why we decided to put it behind our most expensive plan, which is Max but we said back then is if we can decrease the cost of this, we're going to try it in different places because it is our aim to give video call to as many people as possible because it really helps in learning a language.

The good news is that through a lot of really hard work, we've been able to bring down the cost of video call. It is now under $0.01 per video call. And the reason for that is mainly a move towards open source models. It's just a lot cheaper to do that, and we don't see a loss in quality in there. So we're very happy with that. And because of that, we're able to now give video call to super subscribers. The states that we're in right now is most new Super subscribers get video calls.

So if you buy Super right now, we will get video call as well. We expect that over the next few months, we're going to give video to existing user subscribers. We haven't quite done that to all of them. We expect that we're going to do that as well. And that's really good because more people are going to be able to practice conversation.

Now that calls into question what what are we doing with Max? And my answer is I don't know yet. We're -- there's a few possibilities. And one possibility could be that will super subscribers get a limited version of video call, like limited number of video calls and Max subscribers get unlimited. That is a possibility. Another possibility truthfully is that we may actually sat Max.

But -- what I will tell you is that we're going to have an answer to this in the next couple of quarters. And in addition to that, we're going to try to do this without a loss of revenue. And that's partly why we're not going super fast here because we're trying to figure out how to do this without losing much revenue. But again, our intent is that we give it to as many users as possible.

Okay. Great. That's really helpful. And then one thing we've seen reports that you're testing in ad [ supported ] there. So I guess help us think through, do you see any opportunity for maybe a lower price here, something in between super and the free model and from a user segmentation perspective, what are you really going after there?

Yes. We are testing that. It's called Super Lite It is being tested. And now I really want to emphasize the word tested. We are -- I don't know what's going to happen with Super Lite in the end. It is cheaper than [ Super ] about half the price depending on the geography. It is ad supported. So you get ads. And in addition to that, basically, you don't get unlimited energy, you get twice as much energy.

So that's the idea. At the moment, it's still a small fraction of our subscribers are in super light. But part of the reason is that we're early in the testing. We're just not advertising it very well. And of course, the goal with a plan like this would be, if we're pretty sure that you're not going to buy super then we should try to sell you super light. That's kind of the idea. And you're going to see us experiment with that over the next few months. I don't know what will end up happening, but it's something we're trying.

Your next question comes from the line of Bryan Smilek with JPMorgan.

Good to see CURR at an all-time high. Just curious, could you share more color on just overall retention by cohort. Are these new free trial users that are engaging on a daily basis, exhibiting higher engagement trends? Just curious anything you can add there from a retention perspective going forward.

Yes. So like we mentioned in the letter, our retention, our user retention, pretty much all metrics of user retention are an all-time high. The one we look at the most -- this one called CURR, which is current user retention rate. We're very happy that it is an all-time high and that it has increased by about a percentage point in the last year because [ time ] changes to CURR end up implying pretty large changes in daily active users over time because it really compounds. This is pretty broad-based.

Basically, you see it in every region, you see across all types of users. The reason for this is just that the product is stickier. [ I ] mean basically, we have added a number of things, and it's hard to pinpoint to a single one because again, every single -- we have a new version of the app every single week and every version of the app has approximately 350 changes.

So it's hard to point to something that actually did it. But it's just generally our product is just stickier and we're very happy with that because that's kind of the best type of growth you can expect that is just pure, it's a better product.

Great. That's super helpful. And then also kind of building on Nathan's question around Max overall. More from the speaking angle as well, can you just share more color in terms of overall engagement with more intermediate and advanced learners, which would, in my view, likely start to take on more speaking practice within the app.

Yes, we're getting -- I mean, -- so there's two things to say about that. The first one is that if you have access to video call, the engagement is very good. And in fact, that feature has gotten significantly better. One of the main metrics that we have for this feature is a number of words spoken by DAU who has access to that.

That graph is a beautiful graph because it is entirely open to the right over the last couple of years. It's just every month, it is a little better than the previous month in terms of us getting you to speak more. And ultimately, this will just translate in people learning better and being engaged with it. So there's a video call.

We're also in the free tier, making people speak more. Just there's more speaking exercises and there's also more ways to answer, [ exercise ] that before you had to tap on the phone. Now you can answer it with your voice. So we're pretty happy with that. And it's exactly what you said.

This is much more important to advanced and intermediate users, and we see that as something that will over time really help with certainly monetization, but also word of mouth because if people are learning better, they're going to tell their friends, et cetera. So we're very proud of that.

Your next question comes from the line of Ryan MacDonald with Needham.

Maybe just to ask on the bonus incentive comment sort of at the end of the prepared remarks. Obviously, the team internally is operating towards trying to get to that 25% rate. How should we think about that translating to, let's call it, the pace of new experimentation between now and the end of the year and if that picks up, are there any sort of features or initiatives that you'd call out that you're most excited about? Or is this really about just letting the changes in the experience you've already made continue to sort of mature within the market and sort of let them sort of produce results?

It's a bit of both. So certainly, our rate of experimentation will continue. I mean it's pretty high this year. So if you look at actually a number of experiments that we're putting up per week, that keeps growing over the last year that has [ thus ] significantly increased. In fact, it is -- has increased faster than our head count. So that means that per person, we're actually putting out more experiments every week.

So we're -- you will see that throughout the rest of the year.

There's a lot of experimentation that's going to happen. And generally, that's across all the things that we experiment, teaching better and the daily active user growth are where we're spending our focus, but you're also going to see experimentation and monetization, et cetera.

And it's related to the bonus in some way, but mainly it's just this year, our goal as a company was -- we decided we are going to focus most of our efforts of our incredible experimentation machine to just make it so that we can continue growing daily active users because we think that if -- as we said many times, if we can get to a company that is 100 million daily active users, that's just -- that's just a much bigger business, and that's where we want to get to. But you will see experimentation. You'll see a lot as you usually see from us.

Excellent. And then maybe on Max. You obviously are sort of working on a lot of things with Max right now and then one being sort of super or [ voices ] being rolled out more to super. Is there any sort of, I would call it, new feature product development that's sort of maybe geared or aimed towards the Max here at this point that could potentially sort of continue to extend the life of that tier? Or is really everything more focused sort of at the lower tier experimentation right now?

So I won't rule that out. It may happen that we develop something that we end up putting behind Max. We're developing a lot of features, and there may be some that because of cost of them or something we end up putting them behind Max. But that is not the goal. The goal really is to try to give speaking -- particularly speaking, features which are the most expensive ones to provide.

The goal is to provide it to as many users as possible because, again, we just believe that the more people have access to this, the more word of mouth, the more users we have and the larger this business becomes. So that's the goal. But -- so I guess, what I'll say is that's not -- it's not what we're trying to do. But maybe that in 2 months, I come back and say, "Hey, Max actually got a new feature." And it's not because I'm trying to be secretive to you. I just don't know what will end up happening in terms of cost of certain features that we're developing.

Your next question comes from the line of Shweta Khajuria with Wolfe Research.

First one is on attention span for users as attention span, I guess, decreases, how is the engagement and session time across the app trending? And I guess what -- how are you addressing that? And then the second is on math and music. Could you please talk about just the product road map as you see math and music developed through the year and how that could drive contribution to DAU growth.

Yes. Thank you. I think part of your question is something that we've said in the past, which is just the world, in general, attention spend is going down. I mean if you look at how people use things like social media apps, they use them in like 10-second bursts. Whereas for Duolingo, our lesson on Duolingo was maybe 2 minutes. That is something that we need to address. And we are working on that, trying to make the minimum session length longer.

We are -- we have not yet -- we're testing stuff there, but we have not yet released anything. Part of the issue there is that if you make the minimum session length shorter, that's good, but you've got to make sure that people actually also come back many other times in the day so that it compounds. And that's kind of what we're trying -- what we're experimenting with -- but it is something we're actively experimenting with.

In terms of math and music, so we're pretty excited about both of those. I should say there's still -- certainly when compared to chess. They still -- even though they do have single-digit millions of DAUs, they're much smaller than chess. So we do expect growth from both of them, but the contribution to our overall -- I mean we have 60-some million daily active users, even a 50% growth here doesn't contribute all that much -- but we are -- we do expect that they're growing -- we expect them to grow.

Probably the easiest one to talk about in terms of strategy is math because we are we understand that pretty well or at least better than music, I think. With math, originally, when we launched math, I thought, and I was wrong, I thought that we could get the average person on the street to get addicted to learning math. I was wrong. I don't believe we can do that.

But -- so now our strategy is actually to mainly teach math to the people who actually need to learn math, which are basically K-12 students.

Now we may not sell to schools. That's not our goal to sell to schools, but it's at least that the user -- our user base for math is mainly under 18. And as soon as we started working on that, that clarified a lot of things, and I think we started making a lot more progress rather than trying to get your average 35-year-old person who hates math to suddenly love it. I would love to do that, but I just -- I've given up on that one. And in music, it's a bit early. I mean, we are working a lot on music. I'll have -- we'll have more to say about music in maybe a quarter or 2.

Your next question comes from the line of Mark Mahaney with Evercore.

Your comment about video call functionality -- I'm sorry, Luis, were you saying that you could make it available to all Duolingo users or all Super subscribers?

I would love to make it available to all Duolingo users. I don't think we can do that right now. My goal in the moment is to make it available to all Super subscribers.

Okay. And the gating factor to -- for it not to be available to all users to price, right, or cost, I mean, and you bring it out by another 90% next year, and you could do it?

Yes. But there's one other thing that we do have to take into account. This is one of the main things that gets people to buy. So if suddenly all users [ there's ] less incentive to buy and then we would have to find other reasons to get people to buy. So there's a little bit of a trade-off here. So it's both cost and this is one of the main things I get people to buy.

Okay. And then I want to switch gears and ask you about advertising. And I think you've -- you've maybe changed a little bit over the last couple of years in terms of your view on both advertising revenue and on advertising spend. Where are you on that journey? Is there something that you've seen that's made you maybe a little bit more constructive on marketing spend and the efficiency of it and the desirability perhaps of advertising revenue?

You are right that over the last couple of years, I have come -- I've come around to add both in terms of ad revenue and ad spend. A couple of things to say. For the -- in terms of ad revenue, I mean, for the foreseeable future, we will remain a subscription business. I mean our subscription business is much larger than ads, and it will continue being like that for a while.

However, we see a pretty large opportunity with ads. I mean we are -- we have a lot of users. We have a lot of active users and pretty much every app our size or larger than us, makes a ton more money from ads than we do. So we see a pretty big opportunity there, which is why we have become more sophisticated on that over the last -- there was a time at Duolingo a couple of years ago, where our entire investment in ad revenue was at half a person, there's one person whose halftime job was to deal with ads.

So this is how much we were investing in them.

We now have a team it's significantly much more professional. They know what they're doing. I would expect that over the next few quarters, that they're -- that you're going to see some improvements in our ads business. This takes some time get the hang of it. But what I'm seeing is we're just going to be able to give significantly higher quality adds in our product that also make us more money. So we're working on that.

This active work stream that we're doing so there is that.

In terms of ad spend, like for marketing, for marketing for us, we're also significantly more sophisticated. I mean it was a while where I was massively allergic to performance marketing. The way I see it now is I don't think that we want to get addicted to performance marketing because that's never good.

However, it is a good tool to have to complement your other marketing strategies. So our marketing team, which I really think is probably the best marketing team in the world has now started adding as another strategy -- performance marketing and we're very happy with that. It's -- we're seeing the results in our top of funnel for that.

Your next question comes from the line of Ygal Arounian with Wedbush.

I guess, as I listened to the call and what we're talking about here, feels a little bit like we're starting to maybe shift or start to focus a little bit more on monetization talking about ads and new tiers. Is that fair not to say that you're moving away from the optimization around DAUs, but it does feel like you are maybe shifting a little bit the monetization side. How do you think about that transition? When is the right time? What's involved? And then I have a follow-up.

Yes. I mean it's perceptive of you. It's a good question. I mean the reality is this year and for a while, we really -- one of -- the main thing that we're concentrating on is expanding our active user base because we just think that more users is just better for everything. It's significantly larger business, et cetera. And we have this aggressive goal to try to get to 100 million daily active users. We feel good about that goal. We think we can make it.

And if we do get there, kind of all things get better, just more users, you can make more money, all things get better.

It is true that our experimentation -- when we started the year, we had a little bit of a pause and monetization in the sense that we're like [ a ], any monetization that we're doing that is counter to DAU growth, don't do it. And the reality is we were doing a number of things that were countered to the growth. So we kind of had a little bit of a reset. At this point, enough TAM has passed that we understand levers significantly better.

So we are a little more back into like, okay, this type of monetization is good.

This type of monetization is good. So for example, we mentioned longer free trials. We understand that, and we are kind of going hard on those. Ads is another business that we're perfectly happy if we have significantly better quality of ads and make more money from them, we're very happy with that because that's also DAU aligned.

So we are funding that. So it is true that we're probably taking it more seriously, I don't know if that's the right time. We're focusing a little more on it than when compared to 6 months ago. So that is true. I don't think there's a time when we're like -- a lot of people ask us, "Oh, when are you going to stop worrying about DAUs and when are you going to start making money?

A lot of people ask us that. I don't think that we're going to have a shift like that. we're going to continue really trying to get to 100 million DAUs in 2028. But you'll see probably increased monetization focus kind of gradually over time as we find ways to monetize that are not at odds with DAU growth.

Okay. That's really helpful. And then I want to come back to the comments you made about the open source models and how that's driven down AI compute cost. So a -- how should we think about that opportunity in terms of the margin structure over time? It feels like there's multiple levers you could pull on that. And I mean, it's an interesting theme because you've heard this over and over again, this earnings cycle from a lot of companies in this like lean into open source models. And then how does that change the product [ pilots ] I know AI has been a big component in driving your product velocity. So does that become a bigger opportunity with the cost coming down?

Yes. I mean, so at the highest level, it just turns out that this when compared to a year ago. There's just significantly better open source options now. And while for some stuff, they are behind the Frontier labs, they're not that behind for most things, maybe 3 to 6 months. I mean just think out that for a lot of applications. You don't need the absolute smartest model, like the reality is the quality is indistinguishable for many applications that we use.

For example, serving a call like a conversation practice to somebody who can only speak 100 words of a language. You're going to need to serve -- your model doesn't need to be as good as a philosopher. You're probably going to talk about I don't know, frying eggs, like you're not going to talk about very sophisticated stuff.

So we find that in a lot of cases, we can switch to the open source model.

Internally, of course, we're still using many models from, say, OpenAI and Anthropic, and we'll continue doing that. And it will just always be a decision -- a trade-off between can we do it at roughly the same quality but with the open source? And if that's the case, we'll use the open source. There are still cases where we're not able to and in those cases, we'll continue paying.

But I do expect that what will happen is that the cost per usage of AI like whatever it is, the cost per token, if you want to call it, will come down for us because we will continue moving more and more to open source.

And I also, just like you said, that will probably mean that we're able to offer more and more AI features to more users. And so I'm pretty excited about that in general.

Yes. And the way to think about it is we set out the year to be very patient with ourselves about what the business model is going to be. So we really have been trying to stick to this idea that we're going to grow bookings 10% to 12% revenue 15% to 18%. And when we started the year at 25% adjusted EBITDA margin felt like the right level of investment relative to growth. What we're basically saying is by going up about 1.5 points on that adjusted EBITDA margin, we are seeing AI cost savings that will give us a bit structurally a better margin even with our goal to put voice call out to all super users over the course of the year.

Your next question comes from the line of Justin Patterson with KeyBanc.

First, could you talk through just what's driving the confidence level behind that acceleration in Q4? Is that simply easy comps? Or are you just seeing something with the product side finally starting to click there. And then Luis, I just wanted to go back to making the mission affordable or making education affordable for the masses.

Obviously, as you increase advertising in there, you do start to have attention pension [ point ] in there of people get frustrated with the user experience in it.

Does that start to cause some bad behaviors. So I'd love to hear more how you're thinking through the guardrails of still monetizing across the business while still having education be widely accessible within here.

[ Some ] covered the bookings guidance. So on the bookings guidance, the reality is how the year is playing out hasn't really changed too much. Of course, we've held the guides. Q2 was a little better than we expected. So you're basically seeing us adjust for that in the guidance. I think it's important to remember that our bookings guidance also includes about 0.5 point of FX headwind.

So net-net, you're probably at a little bit of a better place.

But I think the year is playing out the way we thought. We've always thought that acceleration is going to come over time. This is not a slow-end study. That's not the way this place is. But we know that it's going to take a little time for the work we're doing to play through, and that's what you see in terms of us holding which the guidance, which implies Q4 gets better from a year-over-year rate of growth perspective.

Yes. And your question about making education affordable. I mean this is a very mission-driven company. Hopefully, we have shown that we will continue being a very mission-driven company. Our mission is to develop the best education in the world and make it universally available, and that is our goal.

The way we're going to be operating is similar to what we're operating right now to how we're operating right now, which is we are, first and foremost, care about our reach. And that every extra active users that we reach means there's one other person that we're teaching, means there's one other person that instead of scrolling -- doom scrolling on social media is actually learning something that is valuable to them.

So we're -- that's what we're emphasizing. But we also believe that turns out that our mission -- I really do believe that even though we're a very mission-driven company, I think that if we succeed in our mission, that's also aligned with just making this a very large business. The more users we have in one way or another, the more we'll be able to [ amortize ] them.

And your last question comes from the line of Arvind Ramnani with Truist.

Just a couple of questions. Just with China just continuing to become a kind of important [ geo ] for you and you all are running OpenAI, Anthropic to run your models or I think mostly opening eye -- does it create some sort of data residency risk, regulatory risk, there's just -- there's just a lot going on in terms of regulation sort of protection measures and what you just kind of move to a local model in that market? Just how are you thinking about it? Because Luis, I know you're fairly forward thinking about AI and not as active. So we just love to see -- not the to make any changes, but how are you thinking about it?

Yes. So you're right. Thank you for asking about China. I mean, China is a very exciting market for us. It keeps growing. It is our second largest market in terms of daily active users, and we expect it to become our largest market in terms of daily active users in a year or 2. Also China monetizes pretty well for us, monetizes about as well as France. So we're -- it's an exciting good market for us.

When it comes to AI, there's a lot there. The reality is that in China, we simply cannot use the AI models, the kind of U.S. AI models. We have to use local models. That's by law. So we do that. In China, all of our usage of AI uses Chinese models, that is just that you have to do that. And in terms of regulatory risk, it's important to mention that, of course, we [ understand ] that China poses some regulatory risk. That's kind of outside of our control, how they operate the country. But we feel pretty good about what we're doing there in terms of our government relations. So I feel pretty good, but I don't know what the government will decide at any point.

Yes. And just on this topic, right? Can you just give us just a rough estimate, right? Like I'm not looking for anything precise whether -- but just a rough estimate of what your AI costs are? And is it like largely on Anthropic OpenAI? Or are they kind of open wait models? And just in terms of like your cost of revenues at like dose cost you like -- is it 10%? Just directionally, how significant are your AI expenses?

Yes. Our expenses on AI and cost of goods sold are tens of millions of dollars. So they're significant to the cost of goods sold. Hosting is another big cost for us as well. But those are the two biggies going through the cost of goods sold. -- inside the business, we also use AI, and that is more closer to the $10 million range internally.

Yes. Perfect. That sounds great. Yes, I think those are the main questions. I mean unless Luis, if you have any like broader thoughts on just -- there's a lot of debate right now in like [indiscernible] models or close to models. You're able to share anything I love to get your high-level view on it.

I mean at the highest level, our view inside the company is that it is in our best interest as a company to use open weight models as much as possible. So if I had a magic wand, I would try to move everything to [ an ] open weight model. It's not always possible because sometimes the frontier models are more band. But from a company standpoint, it is just significantly better because it's way cheaper to use open weight miles.

Perfect. Perfect. And so would you expect like this have we had a hybrid cloud, where do you think that's where we end up? Like you have some sort of orchestration where you put some traffic towards like the highest-quality models versus open weight. Is that how you expect?

Probably. My guess is that most companies that use AI heavily have a portfolio of models that they use where for some users, you kind of have to pick this one model that happens to be a proprietary model. But for other things, you try to use the open weight models. So my sense is most companies will have a portfolio model like that. That's certainly what we have. What I would say is that our expectation is that over the next some time, that portfolio will be weighted a little more towards open weight models than it is today.

I'm showing no further questions. This concludes the Q&A section of the call. I would now like to turn the call back to the host for closing remarks.

Thank you, operator. I'd just like to thank everyone for joining us, and we look forward to seeing you on the next call.

Good evening, everyone, and welcome to Duolingo's First Quarter 2026 Earnings Webcast. Today after market closed, we released this quarter's shareholder letter, a copy of which you can find on our IR website at investor.duolingo.com.

On today's call, we have Luis von Ahn, our Co-Founder and CEO; and Gillian Munson, our CFO. They'll begin with prepared remarks before we open the call for questions. [Operator Instructions]. Please note this call is being recorded and all participants are currently in listen-only mode.

Before we begin, please note we'll make some forward-looking statements regarding future events and financial performance. These statements are subject to risks and uncertainties described in our SEC filings and are based on our assumptions we believe to be reasonable as of today. We undertake no obligation to update them. We'll also discuss both GAAP and non-GAAP financial measures. Reconciliations between the 2 can be found in our earnings materials, and we encourage you to review them when evaluating our performance.

And now I will turn it over to Luis.

Thanks, Debbie, and thank you all for joining. Q1 was about execution. We said we were going to prioritize teaching better and changes in the growing users, and that's exactly what we did. DAUs grew 21% year-over-year, right in line with what we expected as we make this strategic shift. I want to spend a few minutes on what we shipped this quarter related to language learning because teaching better is the foundation of everything we're building toward.

Speaking practice has historically been the hardest thing to do well on a mobile app. This quarter, we made it a bigger part of the experience for free users and paid subscribers. We introduced spoken tokens, which let learners speak their answers to almost any exercise. We started rolling out speaking adventures, which put learners in real-world conversational scenarios and we launched Flash Cards, which help learners build fast recall by saying words aloud.

And for our paid subscribers, video call keeps getting better. Over the past year, we've more than doubled the average number of words spoken per user in that feature. We also reached a major milestone on content. We now offer courses up to professional proficiency, which is B2 on the C of our scale across all our 9 most learned languages and we got their fast.

In Q1 alone, we published 20,500 course units. To put that in context, that's more than 10x what we were shipping per quarter just 2 years ago. AI has fundamentally changed what's possible for us, and I believe we're just scratching the surface. The product is better than it has ever been, and I couldn't be more excited about what's ahead.

And with that, I'll turn it over to Gillian.

Thank you, Luis. Q1 was a solid quarter. We achieved double-digit growth in both bookings and revenue expanded gross margin and delivered adjusted EBITDA of $83 million, which is about 29% of our revenue. As you consider 2026, it's worth reiterating how we are thinking about the year. We are investing deliberately to set us up to be a larger, more durable, long-term business. This means that for this year, we are managing the business towards the targets that we shared on the fourth quarter call.

Specifically, 10% to 12% bookings growth, 15% to 18% revenue growth and an adjusted EBITDA margin of about 25%. To help with your modeling, we have provided point estimates for the full year 2026 and based on what we can see today, consistent with those ranges. These are bookings growth of roughly 10.5% and revenue growth of roughly 16.1% and an adjusted EBITDA margin of 25.7%. A few things we want to make sure are on your radar as you build out your models on bookings. Our expected Q2 bookings growth of about 6% reflects a tough comp. The prior year quarter included the initial rollout of energy, a price increase on our most popular subscription plan and exceptional advertising performance.

We do expect bookings growth to accelerate through the second half with about 3 points of acceleration in Q3 and a further rise in Q4. We expect about 17% growth in Q2 for revenue after which we expect growth to step down in Q3 before stabilizing in Q4.

For gross margin, we expect it to be approximately 71% in Q2, after which it will trend down to roughly 69% by the end of the year as AI-powered feature use in our product expands. Adjusted EBITDA margin in Q2 should be roughly 24%. We expect Q3 adjusted EBITDA margin to be flat to slightly down from Q2 before approaching 27% in Q4. That's a lot to digest. But the overall message is that 2026 is a key strategic investment year for us, and it is playing out as we expected so far, as demonstrated by the point estimates for our financials that we have shared. We entered Q2 with over $1 billion in cash, no debt and expect to generate over $350 million in free cash flow this year. We plan to continue executing on our buyback authorization under which repurchases to date are 514,000 shares or about 1% of our fully diluted shares outstanding.

2026 is a big year for Duolingo. And I am very excited about what we are building. And now I'll turn it back to the operator, and we can take your questions. session.

[Operator Instructions]. We are waiting on moment for Wyatt Swanson with D.A. Davidson to accept.

Awesome. Could you talk to some of the different drivers of DAU growth this quarter and maybe entering just talk to like whether it's performance marketing, word of mouth maybe starting to return or something else? And could you maybe also talk to what regions you're seeing any particular strength or weakness?

Yes. Thanks for the question. So DAU growth, this is very important to us. Of course, this is the most important thing we're trying to do this year. We're growing in every single region as we have been for several years. But of course, some regions are growing faster than others. Asia, in particular, is the fastest-growing region. And in terms of the growth drivers they remain pretty similar. I mean, word of mouth has historically been the main growth driver for us. Most of our users come to Duolingo through word of mouth.

We have some amount of marketing, some amount of performance marketing that we're doing. We've increased that budget a little bit, but it's not massive when compared to other apps our size. But generally, I think it's been pretty similar to the growth drivers. And the last thing that I'll say is historically.

The other place where DAU increases is just a lot of increases in retention to the product, and that's basically the work of just making the product stickier. And that has gone really well. We've been making a lot of changes to the product, some of which are small, some of which are larger that just make retention higher.

The way you would see that is just an increase in our DAU to MAU ratio, which keeps increasing pretty much every quarter, and it increased again this quarter.

Perfect. And then 1 quick follow-up. Can you maybe provide some color on how you expect DAU growth to look in 2Q and whether 20% is still the right way to think about DAU growth going 2026?

Yes. Everything we said in the last call remains. I mean, we expect that it's going to stay at around 20% throughout the year. Now there will be some slight ups and downs depending on the comps from last year, et cetera. But it's around 20% for the rest of the year. So nothing has changed from the last time we spoke.

Your next question will come from Ross Sandler with Barclays.

Okay. Luis, you had mentioned a couple of interesting things 90 days back as kind of part of the plan this year. One was to kind of like revitalize some of the engagement in the free tier kind of like high-engaging free tier. That top 20% of the free tier. So curious to hear any update on that effort. And then you also had mentioned that you wanted to kind of like get inspiration from some of the big like mobile gaming companies out there in terms of new things that you could potentially bring into Duolingo. So yes, curious just to hear what you've learned in any new strategies on that front?

Yes. Thanks for the question. Okay. So in terms of the free tier, I mean, we've done a few things. We basically made it so that the free tier is better than it was 2 months ago. Of course, some of these changes take time. It's only been 2 months since the last earnings call. So it's not like we've done 1,000 changes. But we have made the free tier better. There's just more things that are available to free users. And we're very happy with that. We think that over time, that's going to really increase word of mouth.

In terms of getting inspiration from mobile games, we've always gotten a lot of inspiration from mobile games. And the idea, ultimately, what we're trying to do with Duolingo was make a thing that is as good at teaching as a one-on-one human tutor, but it's also as fun as a mobile game. That's what we're trying to do. And we keep doing that. I mean, if you look over the last quarter, some of the things that we've worked on, very soon, you'll see really cool Avatar costumes that's directly coming from mobile games. I think users are going to love that. We're doing a number of changes in terms of how we show rewards to users. I mean, for example, we're showing them as cards now, and that feels really collectible. So we're doing things like that.

Another thing that is not exactly for mobile game, but that is important and it is also important for -- to mention for the free tier. One of the things that we mentioned last time is that we want to work on monetization tactics that are not at odds with a free tier. And we have been finding some really good ones this quarter and one of them that is good is longer free trials historically, Duolingo has given a 7-day free trial, and that has worked well for us. We're finding that giving longer free trials is really good in that not only does it give us more bookings, that's good, but also, it's good for the user. It's not -- they feel good. They like, oh, I have 1 month free, for example. So that's something else that we're finding that we're pretty happy with.

Our next question will come from Andrew Boone with Citizens JMP.

I would love to talk about now growth and top of funnel at large. Can you just help us understand the deceleration there. I understood the comp and everything from last year. But how do we start to think about what has been the deceleration there and whether that needs to accelerate to support down growth?

Yes. I mean, so in terms of growth, the main thing we work on here is daily active user growth. That's the main thing we have worked on for the longest of time, and monthly active user growth. There's just no team that's looking at that, and we do report it, but it's just not a team that's looking at that.

Now related to MAU growth is top of funnel and that we do work on. We're -- the reality is that top of funnel has been about flat for -- certainly for this quarter, and we would like to accelerate it, and we're working on that. There's a lot of stuff with marketing that I think will be really good, particularly in underpenetrated regions. I think that's one thing.

And then the other one is just making changes to the product to make it teach better and be better for users, that should accelerate word of mouth. Again, the main thing that has been responsible for our top of funnel historically has been word of mouth. And word of mouth is this interesting thing that is beautiful because it's free, but we don't have that much control over it in terms of being able to measure it the same way that we can measure retention. So we're doing things that we think are going to be really good for order mouth, but we don't have the granularity of control that we have for things like retention.

And then, Luis, just a strategic question in terms of keeping users on platform, right? You guys have always focused on fun. It seems to me that there's a change as we think about more of a voice front experience. Just talking about keeping the entertainment value and what has to change as you guys do think about and moving towards more of a voice like experience.

I don't -- I understand why you say that you may perceive a change. But internally, there's no change in terms of fund. I mean we are humongous believers that -- the hardest thing about learning something by ourself is staying motivated, like we have to -- this really is the secret sauce of Duolingo and what has gotten us so far is that we know that we have to motivate our users to learn something. Because there's a very big difference between what people say and what people do. People may say, want to learn something, but ultimately, they'll do what's most fun.

So we spend a lot of effort trying to make it fun. We think that making voice and speaking be more prominent in the app does not decrease fun. And in fact, our metrics suggest that it does not. It can be a pretty fun experience. So we're going to continue doing that.

Now one thing that is important to understand about fun, all the things that get shown on the screen, our beautiful animations, et cetera, it just turns out that humans are very visual creatures. So you will see us continue having a lot of very beautiful animations and kind of more game-like things even though there's voice in there. You're just still going to see all the graphics and everything that -- to keep Duolingo as fun as possible. But yes, we're -- our teams dedicated to making the app more fun are really firing in all cylinders. And you'll see a bunch of stuff really in the next couple of days, for example, you'll see this really awesome feature, which is avatar costumes. I think that it's a lot of fun. My favorite one, I dress up -- if you look at my account on duolingo, I don't know if you can see it yet, but I'm just up as a hotdog, and I love it.

Your next question will come from Eric Sheridan with Goldman.

Great to see everybody. Maybe -- if I can. At least for you, what have been the key lessons so far in terms of scaling AI, both in terms of the user experience as well as the scale of content for the platform over the last couple of months. And Julian, as AI scales on both sides of that equation, how should we think about what that means for margins longer?

Yes, great question. I mean we're very excited about AI. In general, if you go to the highest level, what we're trying to do here is we're trying to make a thing that is as good as a teaching as a one-on-one human tutor and as fun as a mobile game. The teaching part, AI is what's going to get us there. We are really doing, for example, like we said, our video call feature, the practices conversation has gotten significantly better over the last year. The conversations are way more fluid, and that's making it so that users are basically saying twice as many words on average as they were a year ago.

So that's a pretty major improvement. Similarly, like you mentioned content what the amount of content that we were able to put out, learning content that we were able to put out in the last quarter has dwarfed basically everything that we've ever done in the past. We put out 20,500 units of content. And that's in 1 quarter, and that's more that we were putting out like in an entire -- that's about what we put out the entire year last year. And by the way, last year, we were already using AI. So we're just getting better and better at using it.

And the other thing that I think we're working on a lot of things, for example, models really picking what exercise to give to you. We've always had a model that picks what exercise to give to each user, but we're working on significantly more personalization. Because that's exactly what a one-on-one human tutor does. It basically personalizes very close to you. So we're very excited about that.

And then from a cost perspective, I think there's 2 things to think about. One, the adoption of AI, both in terms of customer-facing, and you'll see our gross margin guidance has us landing at about 69% in the fourth quarter and that assumes we're going to put a lot more of that ingredient in our product. And then our operating expenses, we have -- we started to see some pretty big increases in AI costs internally, and our guidance would reflect that. But so we are increasing everywhere.

But there's another thing going on. There are always these waves of efficiency that come with AI. So you might have AI costs come up and then the team optimizes and then you move forward. So if you look, for example, at the Q1 gross margin, it was better than we would have expected and pretty good on a year-over-year basis. And yet there's still a lot of new AI content in our product. And that's because -- on a per unit basis, the costs have come down a lot.

So it goes in sort of these waves. The costs come down, we adopt more, and we manage that. But I think as you think about the overall margins, I would expect us to be in that 69% range. on the gross margin and then we'll manage the operating expenses accordingly.

Your next question will come from Brian Smilek with JPMorgan.

Luis, just going back to last earnings as well. Obviously, very good to see just overall voice being infused across the ecosystem. Can you just discuss the impact on max overall -- like are you seeing mat subscribers cross-trade down to super? Or how should we think about the product market, go-to-market approach on max now that AI is becoming more, not commoditized, but more available across the broader ecosystem?

Yes. Thanks for that question because it helps us clarify. So what we said last time was that we wanted to add video call to our medium-tier super, super Duolingo. It's important for us to do that because video call is such a good feature in terms of teaching, and we just want significantly more people to have access to it because if they do, they'll learn better, they'll tell their friends, et cetera. So we really believe this is the right thing to do.

We started doing that. So at the moment, we have a number of experiments giving video call to super subscribers, particularly to new users, new super subscribers are getting video call. We have not scaled this to all our existing user base, et cetera. So at the moment, there's no change for MAX. I don't know what's going to happen with MAX. There's a lot of possibilities, and it remains Again, there's just only been 2 months since we last spoke. So we just haven't run all the different experiments, but there's some possibilities of what it could be. It could be that we lowered the price of MAX. It could be that we do something where we give MAX subscribers unlimited video call versus a prescriber is not unlimited.

So there's a number of things that it could be -- but at the moment, in terms of metrics, we're not seeing a big difference, except for the fact that new users, a cohort of new users at the moment is not even seeing MAX, they're only seeing super. And that's -- that's just one of the many experiments that we're running.

That's helpful. And I guess for Gillian overall -- the guidance here, obviously, understand the tougher comp on bookings in you just help me think about just puts and takes that drive the back half reacceleration. I know you have mentioned as well to about 20% daily growth ebbs and flows in between each quarter. So with that back half guidance intuitively imply that DAUs would improve from early benefits from these ongoing product initiatives?

Yes. I think as you look at the second half guidance, in general, we are just planning the business based on that 20% growth basis. So really, you're going to maybe see some early returns on the investments we're making, but I wouldn't be banking on a lot of that. We're really trying to take the long view this year. And we really want to allow ourselves to operate in that range of bookings guidance that we gave, so that we can make all the investments we want to make and do what we think is right in terms of the customers. So when you think about the rate of growth Q2 into Q3, Q2 is a really tough comp because of the release of energy, in particular. And a handful of other features that really made that bookings a year ago really strong. And so you'll see us bounce back from that comp. And then you'll continue to see those DAU numbers drive in the bookings.

Your next question will come from Nathan Feather with Morgan Stanley.

The rapid increase in your ability to do content generation is really interesting. On one side, I think you have really a full core set across language learning, at least us the most common languages. Are you starting to see the opportunity to kind of EB test new content for different engagement metrics? And I guess, historically, have you seen an increase in retention payer rates, whatever it may be on a higher quality of content -- and then kind of looking further, how does that impact your thoughts on expanding into additional subject and language learning now that, that kind of cost to entry is lower.

Yes. Thanks, Nick, for the question. So yes, one thing that is exciting about -- we've been working, honestly, for years to try to have the top 9 languages have content all the way to Duolingo Score 120, which we now have. Internally, I kept on going around like, oh, we're almost done, we're almost done. And somebody said something that stuck with me a lot, said, no, no, no. That's just the beginning. Because when we finally have all this content and now we do, we really are in a much better spot to be able to make this content significantly get better based on how the users are performing.

So we're starting really to do that. And yes, we have seen the changes in content. Certainly changes in content quality, but also in the type of content that you show have interesting impact on retention, particularly new user retention. So we are doing a lot of experiments to see -- for example, just what we teach you in the first unit matters a lot. Do we teach you greetings -- that is one thing do we teach you just towards for mom and dad. That's another possibility. And it actually -- it's not as simple as it's always better to teach you greetings. It's not as simple as that. It's a little more complicated than that. But generally, these things do have an impact in retention. And so we're very excited about that. And we're also likely going to move to something where -- we have all this content, this is awesome, but we may even start generating content just for you. Based on everything that we know about you, we may just be able to generate content just for you, maybe not the immediate next exercise, but like 2 exercises from now based on everything. We just generated that sentence just for you or that piece of content. So we really are getting to that point, and that's very exciting.

In terms of other subjects. Each subject has its unique challenges and unique things. So for sure, AI is helping us be able to add new subjects faster. And chess is a great example. I mean we were able to add chess in about 9 months. But each subject has its own types of content that you need to add. So for example, adding math was relatively easy if what you're trying to do is add math in the way that like ChatGPT would show it with a wall of text. But if you want to add math with diagrams and user interaction, et cetera. That is harder, but AI is making it -- it's definitely making it easier, but it's still a lot harder. You're not just like querying an AI to be like just give me some content. It's a lot harder to do it with all the diagrams.

So I guess my answer to that is, yes, it is making it easier to add more subjects. No, it is not yet trivial to add new subjects. And at the moment, we're pretty happy with the subjects that we have, particularly, we're very excited about math. You saw we started this call with a video for math. We really finally got to the point where our math course really has all -- pretty much all the content between grades 2 and 12. And you can actually explain things when you get things wrong. So we're very excited about that.

Great. That's really helpful. And then with the balance sheet, you've got a lot of cash on their high free cash flow. I guess, what are your thoughts on what the right level of buybacks may be or what other potential uses of that cash in the going forward?

So as we look at the cash, you saw we returned some level of cash back to shareholders in terms of a buyback in the quarter. And we do have a $400 million authorization. So we are willing to spend that money. I think that, in general, we're largely focused on operating the business. So we are investing in the business as well as we look at the business this year. So really, it's going to be a balance of the two. And of course, on a buyback, you buy more when the stock is lower and less when it's higher. So we will look at that and look at where the stock is. We think it's a great time to buy our stock. It's a great way for us to offset dilution from the last couple of years as we look at the business. So we like doing this.

And as you can tell, with our free cash flow estimates, we're going to generate almost as much cash as that buyback is anyway in the year. So that's the balance there. I think your other product capital allocation is, of course, M&A. We are out there in the market, always looking at things. But as you've seen, a lot of what we've done is fairly small in nature. It's not going to really hit the balance sheet so badly or if you will, until like a big deal, Obviously, we'll always look at everything that's out there. But as you've seen, Duoingo is very focused on growing Duolingo, investing in Duolingo and going from there.

GameStop wants to buy eBay. We may want to do that too. I'm kidding.

Your next question will come from Ryan MacDonald with Needham.

I'm going to leave that last comment along to Luis a little bit.

We're not buying eBay just so you know.

Maybe can we talk about -- you obviously rolled out and sort of had the announcement in late April about now that the advanced content being sort of available across all the top subjects now. Can you talk about from a marketing perspective, how big of an unlock that is in terms of how you're either deploying that incremental performance marketing budget, now that you have the content available and how we should start to think about how that may be sort of maybe can help to sort of replenish the top of the fall as we go through the back half of the year and into next year?

Yes. In terms of -- certainly in terms of performance marketing, I mean, we -- this matters, I think, most for English learners sorry, in terms of trying to find users that are going to come into the platform, English learners are the ones that are most interested in more advanced content. And some of our performance marketing, some of the main places where we use it is in underpenetrated markets, in particular, Asia. And one of the things that is interesting about Asia is in a number of large Asian markets, we can do profitable performance marketing.

Performance marketing at Duolingo has been interesting because our free version is so good, it has not been easy for us to do profitable performance marketing because what happens is we require people and then they're super happy as for users. But we are finding that we can do that, particularly, for example, in China, we're able to acquire profitably.

So -- English learners. So in that respect, these are a bit related. But I would say the main thing with performance marketing for us is that we just historically have underinvested in performance marketing. And we are getting just a lot more professional about it in -- certainly this year, and I think you'll see the results of that in the next few months. We're pretty excited about that just because we're finally building the infrastructure to have the right attribution to send users to the right place after you acquire them, et cetera, that a company our size should have probably built years ago, but we kind of just ignored it. So we're pretty excited about that.

Awesome. And then maybe for a follow-up perspective. As you're testing video call in super Duolingo for sort of a cohort, that cohort of sort of net new paid subs on super, can you just talk about what you're seeing thus far in terms of elasticity on pricing and the potential demand to pay incrementally for that feature at the super level? And Gillian, maybe how is that informing your view on sort of gross margin profile as we move forward?

Yes. I mean the main thing that I'll say there is that we are running some tests on kind of what the price, what the right price should be for super with video call. I cannot tell you all the results because we started this work a couple of months ago. And it takes some time, we have to build the AB test, and you have to run the AB test for a few weeks to get the results, et cetera. So it's just -- it's just really been 8 weeks since we started this one. I just don't have a lot of results. But what I can tell you is that we are willing to pay more for super with video call. That we know. We've learned that how much more I'll be able to tell you more in a quarter or 2.

And Ron, one of the reasons we've been really trying to focus everyone on we're operating within a set of ranges of financials for the business is to allow ourselves to do this kind of work, this kind of testing, check out different ways of approaching the customer on price. So all of that is anticipated in the guidance around -- here are the ranges that we want to operate in 2026. And in any given quarter, it might be a little bit more, a little bit less. But we anticipated that coming into the year and are executing against that. So there's no big surprises in there. And we think that the financials that we've laid out for '26 can accommodate that.

[Operator Instructions]. Your next question will come from Ralph Schackart with William Blair.

Hopefully, this is relevant. I'll try. Historically, if you kind of think about it, you have a little bit of a paradox. You overmonetize historically. Now maybe we're in some sort of duration of time where you're undermonetizing the user base. But maybe just kind of stepping back, can you just give us a sense, the signals that you're looking at today. What are they in terms of informing you're on the right path right now? And maybe a broader sense when would be the right time to start monetizing again? I know it's only been a couple of months since the last call, but just would love to hear you sort of rip on that.

Yes. So it's a great question because it really allows me to emphasize something that I've said before. And you used to work paradox, which is -- it's kind of what we feel. We are at the same time undermonetized and overmonetized. It is a weird thing. The reality is that roughly 12% of our monthly active users are paying subscribers, we think that number should be much higher. I mean if you look at comps for the freemium models, they're much higher than that. Spotify is close to 50%. So we really think there's a lot of room there. We should be able to get more people.

At the same time, I think certain types of monetization, we probably overdid in the sense that we probably were making the free user experience have too much friction. And at the crux of it was that most of our monetization tactics were kind of at odds with DAU growth, not all of them, but many of them were at odds with DAU growth, which were like if you make the free user experience have more friction. What happens is that some more people subscribe, that's good, that makes you money, but also some more people leave. And what we're -- what we need to do and what we are doing, which I'm very excited about, is finding ways to monetize that don't put DAU growth at odds with monetization. Those ways exist, and we're very happy with.

I mentioned one already, which is longer free trials. We've been historically at Duolingo. This is not something we've experimented with a lot, the length of our free trial. But if you look at other subscription businesses that are scaled they have pretty different free trials. You should be much longer than the one we have. You see 1 month. You sometimes even see 3 month free trials. So you're going to see us experiment with that, and we are definitely seeing certainly the 1-month experiment we already see. We are seeing that, that both increases revenue, which is good but also it's not at odds with daily active users, like because when you give somebody -- when you say instead of a 7-day free trial, it's a 1-month free trial, that doesn't drive any user away. They're like, oh, that seems like a good thing.

So we're -- the work that we're doing is finding ways to monetize that are not at odds with DAU growth. They exist. And they're just not as quick as basically adding friction to the free user experience. And that's what we're experimenting with this year. So what Gillian said is exactly right. The idea is that we have this year to be able to experiment with this, will probably experiment with a 3-month free trial and by the way, experimenting with a 3-month free trial, some of that we said before. that is something that we could have never done if we didn't have a year like this one because in a 3-month free trial, what happens is that your bookings get delayed by a whole quarter. And so you're saying, showing up to -- without being able to say we're going to operate with these guidelines, which is what we did for this. Just saying, oh, sorry, all our bookings are going to come next quarter. They didn't come this quarter. it was a little weird. But this is why we're operating this way.

Your next question will come from Mark Mahaney with Evercore.

I want to ask about the gross margins. So it looks like your guidance implies they're going to kind of phase down or whatever to the high 60s in the fourth quarter. Is that -- is there a reason to think that margins whole gross margins hold at that level? Is there a reason to think that they should cover up higher or lower? Just talk about the trajectory of gross margins after how to think about what drives it higher or lower after this year.

Yes, Mark, when we think about any of the margins that have AI content in them, so let's take gross margin what you tend to find is as you start to introduce features, they might be more expensive to us, and then we optimize that cost over time. So when you look at Q1, for example, the margin held up really nicely as it compared to the year before, and that's because our per unit AI costs have come down a lot. Now as we look forward, we really want to put more and more AI as an ingredient in the product -- and so that's why we have the margin guidance that goes down to 69%, which is essentially where we were last quarter 2. And that just implies a lot more AI content, which we think is great for the business long term. I think from a positioning perspective, it's possible we could optimize that more. But we sort of want to be putting that much AI into the product. So I think 69% is a good place to think about exiting the year. And that could -- if that works, that's going to give us nice growth as well. That said, this is a changing environment and some of the optimizations come faster than you expect and you could see both up and down on that one.

Your next question will come from Justin Patterson with KeyBanc.

Think about Duolingo as always, I think, a high pace of product velocity around [ ABTA ], and the coding is certainly -- it easier than ever to do lots of those. So I would love to hear about how you're thinking about just engineer productivity as a whole number of tests being run and how we should think about that perhaps influencing long-term headcount needs?

Yes. That's a great question. So you are right, we bite a lot we concurrently are running hundreds of EBITDAs at all times. And that has been our product philosophy and that is how Duolingo has gotten better over time because we just are able to increase as soon as we set our minds on a given metric. If you just give us a few months, usually, we're able to increase that metric. We just run a lot of AB tests. We are finding that the number of AB tests that we can run is increasing. Some -- we believe that, that is because of AI usage, particularly in our engineering and product organization.

The increase is not humongous, but it's kind of the first time we've seen an increase on a per capita basis. in years. So we do think that, that is helping. I don't know where this is going to end, but it is an interesting thing because if you look last year, for example, like a year ago, and you were to read Twitter. Twitter would have had you think that you can program anything you want in 5 seconds and it's done? And why shouldn't you be running 10,000 AB tests at once with a single engineer -- that is an exaggeration that is just not the case. And up until very recently, we and companies that are scale because have not seen a real increase in velocity like overall. But we're starting to see that increase.

It's still moderate, but we're starting to see that increase. So I'm assuming that you're going to continue seeing that increase. Now I don't think I don't think we're going to be able to run 10x as many -- as per capita like per engineer. But it is increasing, and that's something that we're very happy with.

Your next question will come from John Colantuoni with Jefferies.

So I just want to ask about sort of DAU trends. Can you give us a bit of color on how U.S. DAUs are trending relative to international DAUs? And what that relative geographic growth could mean for bookings over time given U.S. users generally adopt a sub- adopt a subscription at a higher rate than international users.

Yes. So DAUs are growing in the U.S., and they're growing in pretty much every country, but it is true that in the U.S., they're growing less than in many international markets. In particular, Asia is the fastest-growing market that we have that whole region. In terms of how that affects monetization, we're not -- that doesn't seem to affect it that much. It just turns out that yes, the U.S. monetize as well, but it just turns out that a lot of countries monetized relatively well.

A good example is China. China monetizes about as well as Western Europe, so like -- but as well as France, which is not as high as the U.S. but pretty high. And given that the growth rate in China, I don't know the latest numbers, but I mean it's pretty significant. So I just don't think that the fact that the growth in the U.S. is very low just means that our bookings growth implies is implied to be really low. But I would say that growth in the U.S., my hope is that by making the product teach better and also have higher word of mouth that we are -- and also investing some in marketing in the U.S., which historically we had not invested in actual paid marketing in the U.S. My hope is that all of those things combined will allow us to have higher year-over-year growth in the U.S. than what you currently have.

Your next question will come from Shweta Khajuria with Wolfe Research.

So with the AI-driven content creation, there is a meaningful increase in content velocity, I guess, could you please talk to how you're managing quality of content as that continues to grow against volume and engagement?

Yes. We spend a lot of -- I mean, the main reason why our content is not growing even faster. -- is because we're trying to make sure that it is very high quality. We do a number of things. Certainly, we do evaluations of our content, both with AI and with humans to try to make sure that the content is very high quality. And then after that, we try to test it with our own users in small amounts to see if it's high quality. And if it is, we actually increase the number of people that we give it to. So we're trying to be very careful to make sure that the content is high quality. It is a good point because the reality is that as amazing as AI is, if you are not careful about the quality you can get a lot of slump. And we're trying very hard for that not to happen. So the quality of our content, I think certainly over the last couple of quarters has actually increased. And the way we know that is we basically do spot checks and we kind of rate the quality of all the content, and we know that over the last 2 quarters, the quality has actually increased.

Your next question will come from Omar Dessouky with Bank of America.

Well, I just want to get back to performance marketing for a second. Glad to hear that the company is treated that is a lot more seriousness. The last time we spoke, I say you gave me the impression that this product would be leaps and bounds better in the future than it is today and really would change the way that people learn languages. Does the maturity of the product itself is a bottleneck to scaling performance advertising spend, right? Because performance advertising typically tries to optimize specific types of behaviors optimize for specific types of behaviors and users. So, a, is that the case? And then, b, do you have any sense like at what point you might be ready to really put the pedal to the metal assuming your organization has done all these experiments, like when would the product be ready to really go full bore on performance marketing?

Yes. I mean I would say the bottleneck for performance marketing for us has been, first of all, building the infrastructure for it to be again, like a much more serious performance marketing machine. That is something we are doing at the moment. That's one thing, but there's another one. And it is not the quality of the teaching. It is mainly the problem with our -- how good of our free -- how good our free tier is. One of the problems we've had in -- depending on the region and depending on what we advertise for, et cetera, has been being able to acquire a user, and now they're here and actually getting them to subscribe as opposed to acquiring a user now they're here and they're very happy users of our free product. And so that's been the main bottleneck and that's the thing that we need to get over. At the moment, in some geographies, we have profitable performance marketing, but in many geographies, we do not.

The only thing I would add to that is that we're making an investment in marketing this year, and it's not just in performance marketing, right? The team has a multi-tiered approach to marketing. And to stepping up that investment that is really well thought through and it has some diversity to it as well.

I want to make sure that I'm not thinking about performance marketing in the wrong way because I thought it would be difficult to performance market a product that isn't stable or a product that's not like mature and kind of finalize because users -- you don't actually know what you're marketing. -- changed so much. I just want to make sure I'm not thinking about it the wrong way.

I wouldn't say that -- I mean, Duolingo has been around for 15 years. It has never stayed the same and it never will. So I don't think that that's well, that's not going to change, but I don't think that's been the problem.

Your next question will come from Alec Brondolo with Wells Fargo.

You mentioned how fast China is growing. I think there's been 2 really successful brand tie-in deals over the last 12 months, one with Walking Coffee last year, I think another with [ Methane ] in March. Are there any kind of learnings we can take with regard to how successful those brand tie-ins have been in China and we used to kind of extend that success to other markets over the next year.

Yes. I mean you noticed the brand partnerships. We have had incredible brand partnerships in China. I think some of that is just that our IP and our brand in China are very strong, and that commands some of the largest brands wanting to partner with us. For example, we very soon have a partnership with McDonald's in China. And so the very large brands wanting to partner with us. They come to us to do that. I think there's that part -- there's also the part that I think brands in China, brands like Luckin Coffee, for example, are just a lot more opening to partnerships than for example, brands in the Western world. You just don't see, for example, Starbucks changing all their stores every 2 weeks with a new brand, whereas that's kind of what looking us.

So there's some learnings. Certainly, that team is -- our partnerships team in China and our marketing team in China are very high performing. So there are some learnings of things that we can do in other places, particularly in Asia. But I would say some of that also has to do with the fact that it's the China market. I will say -- the other thing that I'll say is that China is not just growing fast because of the great partnerships. I mean the great -- I think it's kind of the other way around. I think the great partnerships are coming in part because we are growing fast and we're seen as a very cool brand. I mean the reality in China is that it's just a humongous appetite for English learning that just keeps growing and growing. And that's the main reason why China is growing.

Your next question will come from Alexander Sklar with Raymond James.

So on the relationship between DAUs and top of funnel growth versus kind of the visibility you've talked to in the shape of the book's inflection this year what early tests or maybe it's tier geo-mix is out there that's providing your visibility in terms of that inflection? Sorry, in terms of an inflection of bookings exiting the year?

Yes. First, on the bookings, I think if you look at the quarterly progression, what we are guiding to, as you go Q2 into Q3, Q3 to Q4, is fairly on par with where the company has been in the last couple of years. So I think one of the things we want to make sure we remind everyone of is this is -- we're playing a long game here, and the investments we're making are -- maybe see some things from this year, but we're really looking out beyond this year. So 2026 is much more about operating around that 20% DAU growth and then growing the business. And then right now, what you're seeing that progression is pretty typical seasonality. There's been an adjustment here in Q1 and Q2 to our new monetization balance. But then in Q3 and Q4, what you're seeing is quite typical for us.

I'm showing no further questions. This concludes the Q&A section of the call. I would now like to turn the call back to the host for closing remarks.

Thank you. Thanks, operator. I'd just like to thank everyone for joining us, and we look forward to seeing you on the next call.

[Presentation]

Welcome, everyone, to Duolingo's Fourth Quarter 2025 Earnings Webcast. Today after market close, we released this quarter's shareholder letter, a copy of which you can find on our IR website at investors.duolingo.com. On today's call, we have Luis von Ahn, our Co-Founder and CEO; and Gilian Munson, who we are pleased to welcome as our new CFO. They will begin with some prepared remarks before we open the call for questions. [Operator Instructions] And please note, this call is being recorded. [Operator Instructions]

Before we begin, please note, we'll make some forward-looking statements regarding future events and financial performance. These statements are subject to risks and uncertainties described in our SEC filings and are based on assumptions we believe to be reasonable as of today. We undertake no obligation to update them.

We'll also discuss both GAAP and non-GAAP financial measures. Reconciliations between the two can be found in our earnings materials, and we encourage you to review them when evaluating our performance.

And now I will turn it over to Luis.

Thank you, everyone, for joining us. I want to start by talking about Matt Skaruppa, who up until today has presided over 100% of our earnings calls as our CFO. It really has been one of the singular honors of my career to work with Matt. Matt, I know you're watching tonight, and thank you really for everything you did for our mission.

Now Matt is a hard act to follow, so I can't think of anybody better to take his place and the person who helped us hire him, our long-term Board member and Chair of our Audit Committee, Gilian Munson. She's our new CFO. She's here with us today, and I'm very much looking forward to working closely with her.

Now let's get to the business at hand. 2025 was another strong year for Duolingo. For the first time, we surpassed 50 million daily active users, and that's more than 5x as many as we had when we IPO-ed in 2021. We also delivered over 1 billion in bookings and more than $300 million in adjusted EBITDA. I'm proud that we have built a business that is profitable at scale and that is having a huge positive impact in the world.

But what excites me most is not what we've already achieved, it's what's ahead. I'm more convinced than ever that the accelerating advances in AI will fundamentally change the way people learn. This creates an enormous opportunity for us. We are the most popular education app in the world by a margin, and we intend to lead this shift. By leaning into this moment, we believe we can redefine the future of learning while generating exceptional value for our shareholders over the long term. So huge opportunity ahead of us. And the most important thing right now is to continue attracting users. We said this in our last earnings call.

Now while our DAU growth over the past several years has been nothing short of phenomenal, it decelerated throughout 2025, and as the new year began, we continued seeing this trend. We now expect DAU growth to be about 20% year-over-year throughout 2026. This reinforced my conviction that we needed to take more decisive actions to reaccelerate DAU growth.

Another way of saying -- another way of seeing it is that long-term value in this business is driven by 2 things: the size of our active learner base, that's like the size of the pie. And how effectively we monetize that base. You can think of that as the piece of the pie that pays. At this moment, we are prioritizing growing the size of the pie. So as I detailed in our shareholder letter, we have a carefully considered plan for 2026 that focuses on teaching better and user growth.

In addition to the thousands of A/B tests we plan to run in 2026 to improve our product, we have larger initiatives that fall into 3 categories: teaching languages better, improving the free user experience and feeding our next growth engines, meaning our subjects of math, music and chess. All of this is in service of growing DAUs faster.

Our medium-term goal is to reach 100 million daily active users in 2028. If we succeed in doubling our DAUs, the payoff would be significant. A more resilient brand, a business with meaningfully higher bookings and profit, and most importantly, a company that reaches and teaches far more people around the world. But in the short term, the short-term implication is that this year, we'll see slower bookings growth and lower profitability as captured in our guide.

I want you to know that I don't take this decision lightly and that I know it may come as a surprise to some investors, but it's fundamentally aligned with what I said to shareholders in my very first letter, which I will now read: "Dear potential investors, the main thing you need to know is that I plan to dedicate my life to building a future in which, through technology, every person on this planet has access to the best quality of education. And not only that, but a future in which people want to spend their time learning. Duolingo is the platform for building that future, and we are just getting started."

And with that, I'll turn it over to Gilian long-term Munson.

Thank you, Luis, for the really warm welcome. Being a full-time member of this team is going to be a great adventure. I'd be remiss if I didn't say a huge thank you to my friend and colleague, Matt Skaruppa, who we are all going to miss and who I wish all the best.

We had a solid finish to the year and ended in a strong financial position, a great platform from which to look forward strategically, as Luis explained.

Turning to the go-forward plan and guidance. In our shareholder letter, we have included a lot of specific details to help you with your 2026 modeling, likely more than we will going forward. I'll repeat a bit here and add some color in areas we really want to be sure we emphasize. At the highest level, our 2026 guidance is bookings growth of 10% to 12%, revenue growth of 15% to 18% and adjusted EBITDA margin around 25%. For our Q1 guidance, it is 11% bookings growth, 25% revenue growth and an adjusted EBITDA margin of 25.5%.

Here are some details that I hope will help you with your models. Looking at bookings. Quarter-to-date, as of last Friday, Q1 bookings growth was tracking above our Q1 guidance. However, it's important to remember that we are heading into a particularly tough compare in the first half with dead Duo, a price increase, particularly strong advertising bookings and the rollout of Energy. We are modeling some improvement to bookings growth rates at the end of the year, but have only factored in modest early returns on our investments at this point, which we think is prudent as shifts like the ones we are making can take some time to materialize.

Turning to revenue. We believe that the rate of year-over-year revenue growth will adjust down following bookings growth rates over the course of the year with some stabilization in the second half. Given our Q1 guidance is 25% growth, the math implies that quarters 3 and 4 will be below the low end of this guidance range. As for cost, as you think about our COGS, the key thing to keep in mind is that we plan to share AI features with a far greater portion of our user base. This is anticipated in the lower gross margins that we outlined in our letter.

Additionally, we are investing in 2026. As a result, we expect that R&D and sales and marketing spend growth will outpace revenue growth. Putting it all together, the trajectory of adjusted EBITDA margins will likely be slightly different in 2026 than in previous years. In Q1, we've guided to 25.5% adjusted EBITDA margin, and we expect adjusted EBITDA margin to decline roughly 3 points sequentially in Q2. Then it will improve through the back half of the year with Q4 being our highest margin quarter in 2026.

I'd note, however, that we do manage to annual targets and the timing of our spending could shift, but we will update you on as we continue the year. That's a lot of detail we know. We hope that in combination with the shareholder letter, this does help you manage your estimates as we execute our strategy.

You may ask about our guidance philosophy. And so I thought I'd note that. This year, our view is that like in Q4, our hope is to narrow the gap between our guidance and actuals versus what you have generally seen from Duolingo in the past.

Finally, I would like to highlight our buyback authorization that was announced today. Our Board has authorized a buyback of up to $400 million in our shares. We believe this represents good capital allocation discipline and expect to execute upon this authorization in the coming year. I couldn't be more excited to join this company at this time. The size and scale of the market opportunity ahead is massive. And I very much view 2026 as a foundational year to achieve that opportunity.

Thank you, Luis. I'm excited to step into this role during this important moment at the company.

And now I'll turn it back to the operator, and we can take your questions.

[Operator Instructions] We'll take our first question from Bryan Smilek with JPMorgan.

Luis to start, can you just elaborate a bit more on what you're seeing in the market across AI and just overall innovation across the technology space that makes this the right time for the strategic shift to reaccelerate user growth? And I guess, more fundamentally, to achieve your 2028 target to double users, what are the, call it, 1 to 2 top priorities across the product that will really enhance that DAU growth and drive the implied acceleration to get there?

Bryan, thank you for the question. Okay. In terms of AI, look, we really are in a unique time in history. AI is going to allow us to teach significantly better. And I also believe that the way people learn is just going to change. And the reason for that is, I think that in the next few years, we are going to see quality of teaching that is as good as a one-on-one tutor. We're going to see that. But not only that, not only is it as good as a one-on-one tutor, but it will be as fun as a mobile game. So it's this incredible thing that's going to happen.

And the things that we're seeing is we're just -- we're able to find where users or learners are having trouble and really specialize in that and answer all their questions and whenever they have any kind of confusion, we can get in there and really fix it. We also -- in the case of teaching languages, we can finally practice conversation better. We have this feature video call with Lily. It keeps getting better and better at being able to practice your conversation. And so we can see improvements in how well people are learning.

And the same is true for our other subjects, certainly with math. Like I said in our shareholder letter, I believe that this year, we're going to have the best tutor app for math. This is going to be Duolingo math. And it will just -- I think this is a huge market potential. There's about 1 billion people learning math in the world. So we'll be able to teach them as many as we can really well. So I'm very excited about AI in general. And this is why I think now is the time to try to attract as many users as possible.

Now in terms of what are the top priorities that will allow us to get to 100 million daily active users by 2028. It's a number of things. In general, the 3 buckets that we're going to be investing in are teaching languages better. Languages is our largest subject, and it will continue being our largest subject for a while. We're going to teach languages significantly better. And there's a lot of work there.

Another one is in making the free user experience better. The less friction that there is in the user -- in the free user experience, the more users we have. We have seen this over the last 15 years. Whenever we'd remove friction from the free user experience, DAUs grow, right? So that is another one.

And then the third thing that we're going to be doing is just investing a lot in our new growth engines. That's our new subjects. So in particular, chess has grown quite a bit. As I said in our shareholder letter, we have about 7 million daily active users in chess and this is less than a year after we launched. At this point, we believe we're the second largest chess thing in the world, which is, by the way, incredible because if you search for chess on the App Store, we don't appear in the results. We just haven't even gotten there to doing App Store optimization, but we already got to 7 million daily active users. So those are the things that I think are going to help us get there.

Great. And then I guess, elaborating a bit more, you mentioned video call as well, too. Can you talk about the philosophy and the product roadmap really across Max versus Super? You've been very transparent in the past in terms of oscillating the features and the relative pricing set. So how should we think about just overall the Max road map as you start to ship those video calls to Super users as well?

Yes. When we started Max a few years ago, we introduced some AI features. We were super excited about all the AI features because they really can help us teach better. At the time, cost for AI inference was way higher than it is now. And so we decided to put our AI features behind a much higher subscription tier, which is Duolingo Max. That's what we have done. But we said back then that as AI costs came down, we would experiment putting features in the different tiers because ultimately, our goal is to be able to teach as many people as possible.

At this point, the cost of a video call has gone down very significantly over the last couple of years since we launched video call. And we feel pretty comfortable being able to put it inside Super Duolingo. Now like everything we do, this is going to be an A/B test. We're going to A/B test what happens if we put some video call in Super Duolingo. I'm not sure what that A/B test will do. It may be that this increases revenue because more people sign up to Super Duolingo or because the retention of Super Duolingo is significantly higher because video call is there, or it may be that it decreases bookings because we're not making as much money from selling Max.

I don't know what it will do. But whatever it does, we're going to take the appropriate options. So for example, if the decrease in bookings is too high, we may do something like we meter video cost. So for example, on Super Duolingo, you only get access to 1 video call a day versus on Max, you get unlimited. Now I'm saying all of this very much in advance because I don't actually know what will win in these A/B tests. But what I will tell you is that, for sure, we will put some form of video call in our Super tier because we just think that it will help us teach a lot better to a significantly larger number of people. There's about 10x as many subscribers on Super that they are on Max. So if we put video call on Super, 10x as many people have access to conversational practice.

And the other thing that I think is we're going to do this on our own terms. I mean my sense is that over time, it will not make sense to have a feature like video call on their plan that is so expensive as Max. So we're doing it from a position of strength right now where our finances are very strong, and we're going to just do that as opposed to if we had to do it defensively 3 years from now because a competitor popped up. I mean, at the moment, we really are not worried about competition. So we're getting ahead of it.

And Bryan, we designed the guidance to give the company the room to go do all of this and experiment with this and do the right thing for our customers. And that's precisely how the guidance and the way we think about the financials this year are designed.

Our next question comes from Ralph Schackart with William Blair.

Luis, last call, when you talked about the reprioritization for 2026, and I think it was sort of contemplated it was going to have some impact on the financials for 2026. But maybe not to the level guided that was presented today. At least that was my interpretation. I could certainly have that wrong. But I guess, did something change since last quarter in terms of how you thought about 2026? So the user growth, is this larger than perhaps you originally anticipated?

And then maybe a question for Gilian on the 20% DAU growth. When these friction points or monetization friction points are removed, should we expect some reacceleration in the DAU metrics? And I guess, you sort of suggested that you're tracking above in bookings. Maybe just some kind of early reads on, as you're removing some of these friction points, how the user growth is responding?

Thank you, Ralph, for your question. So yes, in the last call, as you said, we said that we will be concentrating on user growth. And for similar reasons as now, one, we had seen deceleration of DAU growth throughout 2025, some deceleration of user growth. And in addition to that, we think this is just a unique opportunity where instead of decelerating user growth, we should be accelerating user growth because we should -- just capture this opportunity. So that was what we said in the last earnings call.

What happened between then and now are two things. The first is that we kept on seeing this trend of user deceleration. And by now, we expect about 20% year-over-year growth throughout 2026. And that -- and the other thing is that my conviction of AI just increased even more. And that just made it so that I have even more conviction that we need to spend really the vast majority of our efforts on DAU growth. And as we built the plan for 2026, this is what came out. I mean the guidance that we're giving.

And in terms of the 20%, I guess, Gilian, you can go ahead and answer.

Yes. So as Luis said, we really do expect that 20% through the year. In terms of the Q1, what we're seeing so far, it is tracking ahead of the guidance. And we think that shows everyone that there is a lot of health in the business. But you do have some really, really tough comps coming. And so the overall guide reflects that.

I think it's really important to put in context that Luis is talking about is the confidence we've gotten in going for the bigger price. And so when you think about that 20% and the 100 million target, that would imply an acceleration to much higher growth rates in the next couple of years. And if you kind of play that out in terms of what the financials could look at, that is, as Luis said, a much bigger business.

So as we look at the business, we think to ourselves, okay, we could have grown the business, probably we would be at about a mid-teens CAGR over the next couple of years. If we could keep our margins where they are, we're probably looking at a $1.5 billion business, $400-plus million in adjusted EBITDA. But we, at the company, are really motivated to go for the bigger prize, which is much more aligned with what Luis and Severin were thinking about when they founded the company.

So then if you sort of play out the DAU, this was all back of the envelope. But if we can get to that DAU and even reasonable monetization assumptions from where we are today, and we believe we can scale our expenses over time, you're looking at a business in a couple of years, it could be $2.5 billion with over $700 million in adjusted EBITDA. And I think around this entire company, we're really motivated to go for the bigger price. And so that's what you're seeing and kind of how we're thinking about it.

In terms of when you're going to see it in the model, we've really modeled it late in the year. So it's going to take a little while. What we're working on is going to take a little while to see through it. But we do have a little bit in the end of the year. We think it's prudent to keep that small for now until we kind of get going. You'll start to see more and more as we move into '27 and '28.

The next question comes from Mark Mahaney with Evercore.

I want to ask a growth question and a product question. The growth question is, you mentioned these factors that caused this deceleration, law of large numbers, increased maybe over focus on monetization. There are 2 other potential reasons, too. So just address those, which is market saturation and competitive intensity. Maybe there's -- so just -- are there new data points that just remind you or open your eyes even more to the long-term growth opportunity here?

And then in the products, the video call is a wonderful product. These other things that you're talking about, the speaking adventures, is that on the come? Like how soon should we expect that? And getting to this more advanced content with all of your 9 largest languages up to that 129 score. Is that something that happens now? Or is that something that's a multiyear build? Like how long does it take to get to those 2 products that will -- that we could potentially see the products and then the impact that they would have?

Thank you, Mark. Okay. So your first question is about market saturation and competition. We're worried about neither one of those. In terms of competition, if you look at language learning apps, we have about 85% of the daily active users of language learning apps in the world. And that number has remained pretty flat over a while. Basically, we're just not particularly concerned about that.

In terms of market saturation, we're -- for example, we look at things like the percentage of daily active users in a given country compared to the size of the Internet base of that country. So for example, in the United States, 2% of all Internet users on a given day use Duolingo. So 2% penetration in the U.S. for daily users. In the U.K., that's 3%. In Germany, that's 4%. And Germany is not the highest country we have. We actually think we can get much higher than that. But even if we only assume that every country got to 2%, which is the U.S. penetration, we would more than double our daily active users. So I just don't think we're near saturation at all.

And in general, we are not seeing anything different. For example, we survey our churn users. And the answers to the surveys of our churn users haven't really changed in years. And the most common answer, by the way, when people stop using Duolingo about where they go is that they stop learning. And that has been true for a while. So that's just not something we're particularly concerned about.

In terms of the features, we mentioned in the shareholder letter, we mentioned speaking adventures. We're very excited about that. What that feature is like it's basically -- it's a little game where you have to do something, for example, you may be told, go buy a sandwich and give it to that person and then ask them for money or something, something that you have to do. So in that -- you do a lot of things with just text, but every now and then in there, you have to speak to get the things that what you want. That's going to come out -- that's going to be for all users, free and paid. That's going to be scaled. I mean we're already testing it. That's going to be scaled sometime midyear. So you probably see it in your app sometime between now and the middle of the year. We're very excited about that.

In terms of more advanced content, that's happening relatively soon. In a matter of a month or 2, you will see significantly more advanced content. So basically, all of our courses that teach the top 9 languages that we teach, which, by the way, cover more than 90% of our daily active users will have content up to Duolingo score 129, which is a particularly important score because that's the score that you need in order to get a knowledge job in that language. So we're very excited about that. And that's coming now. Importantly to say, that content is going to go out there, and we're very happy about that. But that's just the beginning. As soon as it goes out there, we can start improving this content the same way we improve everything. And so over the remainder of the year, you're just going to see better and better versions of this more advanced content, and we're pretty excited about that.

We'll take our next question from Justin Patterson with KeyBanc.

Luis, I was hoping you can talk a little bit more about just the social strategy. We haven't really seen the same degree of virality through the -- through TikTok that you used to be getting. So would love to hear about how you're refining that and thinking of that as just a vector to grow daily usage. And then with respect to video call, I know you've made a lot of improvements within there designed to increase more words being spoken, to make it easier to use. So would love to hear about just how that's been going versus expectations and how you think moving this into the Super tier can really drive more engagement and eventually help contribute to that bookings reacceleration.

Thank you, Justin. In terms of our social marketing strategy, I mean we have an excellent marketing team, really, I think, best in the business. We -- our videos are going viral and they're getting millions of views. But I have to say the comparison point, particularly a year ago, a year ago, there were many weeks where we had the single most watched video in all of TikTok. That was -- that's incredible. So while we're still seeing virality in our videos, we're not in a position where it's the most watched video in all of TikTok. There's a number of reasons for that. One of them is just algorithms have changed enough that it's harder to do that consistently. So I mean, that's one reason. And in general, we're going to continue with the strategy of really trying to get as much virality as possible. And that will -- I'm pretty happy with the results so far. But you are right that compared to a year ago, it's not quite the same.

In terms of video call improvements, I'm very happy with the results for video call improvements. Some -- basically, if you look at the graph of words spoken per user on video call, that is just a monotonically increasing graph over time. It just keeps getting better and better and better. And basically, we're just a lot better at getting Lily that you're talking to, to elicit more conversation. Right now, we're trying a new thing, we're actually starting to tell you, hey, in your next response, ask a question or in your next response, use the word because or something. And that actually really gets people going more and more.

So I'm very happy with all these improvements. And I really do think that putting it -- giving it to more users, what that will do is it will basically get more users to be promoters of Duolingo because we see that people that engage with video call post online and say things like, "My god, I was skeptical that I could learn a language on Duolingo, but I've just been using Duolingo and using video call and somehow, out of the blue, I am now able to have a conversation." So we're very happy with that.

Our next question comes from Andrew Boone with Citizens JMP.

I wanted to go back to the free user experience and improving it. And really, the question is, how does that evolve into a monetization strategy over time, right? So it sounds like one of the lessons was that you guys essentially overmonetized your user base. So Luis, as we think about going through this transition, what happens on the other side in terms of the lesson that you guys now have as you guys come back and then start to monetize users later on?

Andrew, this is an excellent question. Look, we got ourselves into an interesting situation where both Duolingo is undermonetized and overmonetized at the same time. It's undermonetized in that, look, only about 10% of our monthly active users pay us. We wholeheartedly believe we can do much better than that. If you look at comps, if you look at -- if you look at things like Spotify, half of their users are paying them, give or take. And a lot of these freemium businesses, we really think we can get much higher than 10%.

However, the way we were increasing monetization was we found that the quickest way to increase monetization was basically by adding friction. The more we added friction, the more we got people to subscribe. And that's okay, but I think that we got it to a point where it really became at odds with DAU growth. There are other ways to get people to subscribe. For example, you can add more features or you can have features that are not exactly friction, but things like selling customizations for avatars and things like that, that have worked really well for games, and we are kind of half a game.

So what we're doing now is for the next year, we're using this year to also find ways to monetize that get more people to subscribe. That is true, but that may not be doing so by adding friction. And we're really confident that we can do that. It's just slower because monetizing by adding friction, like if you told me next week, I needed to make, whatever, $50,000 more a day, it was actually quite easy to do. You just double the ad load or whatever it was. But these other ways of monetizing are just going to take a little longer to do, but we're very confident that we're going to have some because only 10% of our active users pay us to subscribe.

And then I wanted to ask about the chess disclosure, and this is really about utilizing multiple apps at the same time. So if I think about the 7 million users, and I think about the core trend of core Duolingo through 2025, how do I think about multiple users using multiple apps, right? Do I just take the 53 million and I subtract 7 million to be able to think about where you are? What's the right way to think about that?

Another great question. Yes. So we said chess has gotten up to about 7 million daily active users in less than a year. We're very proud of that. It's an incredible growth. Now the majority of these users, whenever we put a new subject out on Duolingo, it is a lot easier to get to our own users rather than to get to users outside of Duolingo because we have a direct line of communication to our own users. So the majority of these users, we're not basically saying the exact number, but the majority of those users are users that were Duolingo users and now are using both languages and chess. That's basically what's the common thing that's happening.

So we're -- the way to see it is not that languages have not been growing. In fact, languages are growing at about the same pace as Duolingo because it's the largest course. It's just that we are getting some new users for chess, but a lot of them are just doubling up from languages and chess. And what we are seeing is that over time, more and more people are finding out that Duolingo has chess and then chess is starting to attract its own user base. And I think that's what's going to happen over time. And that just takes a little bit of time to do that, but that's what's happening.

Our next question comes from Alexander Sklar with Raymond James.

Luis or Gilian, I don't know who wants to take this one. But on the bookings outlook, just given some of the unknowns you spoke to with Bryan's question earlier on how video call tests are going to play out on Super. What is the bookings outlook factor in terms of conversion rate of Super relative to what you've experienced in the past? And how much time or data, Luis, do you think you need to know if the decision is a successful one based on some of the historical testing?

So if you look at the bookings, the main thing to think about right now is if you take -- start from Q4 bookings growth of about 24% and then look over to the guide, we think about half of that is the user deceleration continuation that Luis talked about. And then the other half of that is us monetizing more carefully and thinking about taking away some friction for our users. So that's how we're thinking about positioning that.

In terms of exactly how all the monetization is going to play out, as you know, we A/B test like crazy around here. And so there's a lot to be learned, as Luis explained, over the course of this year. What we're really trying to do with the guidance is give ourselves the room from a position of really good financial strength to go do that work because we think it gives us the opportunity to reach for the higher price.

Yes. And in terms of how long it will take until we know this is working, it's going to take a little while. I mean, we need to make a lot of product changes, et cetera, each of which take a while and then we also need to see user reactions, et cetera. So it's going to take a little while. As Gilian said, we're putting in something modest towards the end of the year to see that some things are working there.

And then maybe just a quick follow-up. Anything factored in terms of pricing changes within the different packages in the outlook?

I mean, what we're going to do this year is we're going to be experimenting, like pretty much every year, although probably more so with prices quite a bit. That means higher prices, lower prices. We're always A/B testing. So we're going to be doing that. But we're not -- this is not something that we're really putting in -- we're putting in the results of all the combined experiments in the outlook. But I don't know if I can tell you for a fact that the price of Super will be higher because we have video call or not. I don't know the results of the A/B test.

Our next question comes from Wyatt Swanson with D.A. Davidson.

Thanks for the question. I've just got one thinking about like the medium-term goal of reaching the 100 million DAUs. Does that assume like marketing stays elevated and monetization stays on the back burner for the entire duration? I think you've kind of answered it, but just sort of thinking through as you sort of ramped monetization over the past year, and you've started seeing DAUs start to decel. So like how do you think about that balance over the medium term? Should we see monetization slowly start to ramp up in these alternate forms that you're kind of exploring?

Yes. I mean I can answer and I can also let Gilian have at it. But in general, we -- our marketing is still extremely efficient. I mean it's really still a small fraction of our revenue that we spend in marketing. I assume that every year, the total number of marketing dollars that we spend will continue going up. However, I'm hoping that past this year, we are going to get some efficiencies. It's hard to say exactly what will happen past this year. It's not like we're guiding, but I assume that we're going to get some efficiencies past this year.

And then in terms of monetization, like Gilian was saying before, just almost any way you look at it, with 100 million daily active users, this is just a much larger business. I do believe -- again, hard to put in numbers exactly to give you a model or anything for past 2026, but I do believe that you're going to see other things really start ramping up. I'm very excited about direct ad sales. We're seeing just good uptake on that. And I'm also very excited about in-app purchases, in particular with our avatar. There's a lot of love for our avatars and for our playful brand, and we think we can do quite a bit there.

Yes. And as you think about the margins of the business, one of the reasons I'm so excited about where we're at is that we've shown that this business can have tremendous scale even at the size we're at today, and we think we can be a lot bigger. And there's nothing that we can see in the business that says we can't scale it going forward. We're making an investment right now and descaling some to make that happen. But really, we don't think we need to hire double the number of people, twice amount of marketing. We think that this business can scale really beautifully. It's one of the nicest things about the business model that we have the opportunity to do what we're doing from a position of strength to profitability. We will continue to generate tremendous EBITDA and free cash flow even as we invest, and we see no reason why we couldn't scale up as we reenter monetization and a bit more of a smarter way from here.

Got it. Got it. That's fair. And could you maybe answer us like what the ads business will look like going forward? You've talked about like doubling the ad load is maybe causing additional friction and users potentially churning off. So like our ads -- it almost made it sound like ads were almost coming off of the platform, and it wasn't going to be as big of a monetization channel.

What we're unlikely to do is increase ad load. At the moment, you see an ad at the end of a lesson. That is unlikely to go up, certainly not this year because that really goes counter to everything that I'm saying. However, we are doing a lot better of a job at displaying better ads. I mean, historically, we just have not spent a lot of effort on ads historically. So for example, we use just these network ads that are kind of the lowest common denominator network ads. We are now doing direct deals.

What that does is it does two things for our business. Number one, it significantly increases the quality of the ad. So rather than getting ads for like games that you've never heard of, you're going to get ads from like Disney or something, significantly higher-quality ads. And then the amount of money we make per ad is significantly higher when we do that. So I think that's kind of how it's going to be. We're experimenting with some stuff that is really cool, but I don't know if we'll make it work.

I'll tell you, my dream has always been to be able to give you ads in the language that you're learning. And we are trying that. So if you're learning Spanish, part of the ad is going to be in Spanish. What I love about that is that the users want that. And so -- and that's also good in the end for the advertiser because the user is actually paying a lot of attention. So I'm hoping we'll be able to try that. I don't know which of these things will work. But in general, I just think ad load will not be higher, but we'll make a lot more money per ad.

Our next question comes from Nathaniel Feather with Morgan Stanley.

Two on my end. First, there's been a lot of concern in the market about the improvement in AI translation tools and the general rise of all-purpose chat bots. I guess what gives you conviction that the user deceleration we saw in 2025 was not due to those factors? And then second, just of the changes you've made so far to improve the user experience, are you seeing anything in the data that gives you conviction that these changes are going to stabilize user growth through the year? And are you already seeing that? Or when might you expect to reach that stabilization point?

Thank you for the question. So AI translation isn't -- people talk about it, but it is not something that we are worried internally about. AI translation has been essentially perfect among the large languages for like more than 10 years. I mean between Spanish and English, it's essentially been perfect. Our users use Duolingo for 2 main reasons: one, as a hobby, they actually want to learn. And those people, whether there's AI translation, it doesn't matter, it's a hobby for them. And then the other group of people is learning English. They want to learn English, and they actually want to learn English.

We know also that we ask users why they leave Duolingo. We have this churn user surveys. This does not come up. I mean, like I said, the main reason that people leave Duolingo is because they got busy. We know what that means. It just means they are doing more social media, that's basically what that means. So we're just not -- that's not something we're particularly concerned about.

In terms of data about -- are we seeing any data that is making us believe that user growth is accelerating, et cetera. So far, we have done a few things to make the app -- the free experience better. For example, we gave a feature that was in Duolingo Max called Explain My Answer. It's not free for everybody. So we've done stuff like that. And we are seeing quite a bit of uptick on the features, and we're seeing that users love it. But we are -- we have not yet seen an acceleration in DAUs that we think is like, "Oh my god, we're done." Now to be fair, we did not expect this. I think this is why we're saying that you should expect seeing some modest results towards the end of the year.

Our next question comes from Ryan MacDonald with Needham.

Luis, I wanted to ask about sort of Duolingo Max to start. Fourth quarter, I think, was your sort of large cohort or lapping the first large cohort of Max subscribers. And I was curious what you were seeing from a retention and renewal perspective and how much that's guiding sort of the strategy to bring video call back into Super. And then if we're going to -- when this sort of transition happens, Gilian, when should we start to see, I guess, the impact of subscribers moving from Max down to Super with the video call being the future? And what assumptions are you making for that in the guidance?

Yes. For -- in terms of Max, none of the Max numbers are making us do anything with video call. The Max numbers are actually quite good. We're happy with the Max numbers. We are taking, what I call, an offensive move with Max. I just believe that the cost of video call has gone down enough. It just kind of doesn't make sense for us to just have it in Max when we can offer it in much cheaper packages. We're not quite there where we can offer it for free, but at some point, we may be able to. And I want to offer it for free because I actually think that it is in our best interest to have the largest possible user base that we can figure out how to monetize in any way rather than just keeping this for a tiny little bit of the user base. So that's just our philosophy. So the entire decision for trying to put video call in Super comes more from we think that this is the right thing for our users rather than any numbers with Max. We're pretty happy with the numbers with Max actually.

And then Gilian, I don't know if you wanted to answer that second part. You're -- we cannot hear you. You're muted, or I cannot hear you. Well, sorry, your question, I kind of forgot what your question was because I was like I'll ask for Gilian. I forget it. What was your question? I'll again try to answer that.

Yes. As you think about the guidance for fiscal '26, I guess what assumptions, if any, are you making sort of building in for when video call rolls out to Super for sort of, I guess, you could call it sort of ARPU trend changes of maybe downsells from Max down to Super from that?

Yes. You know what, I'll tell you probably what -- I know what Gilian would have said in here. Look, we ourselves don't really know exactly what is going to happen with this A/B test, where we start giving video call to Super. We don't know that ourselves. But we're going to A/B test it. Everything -- we put in our guidance some bounds, what we think that will actually -- this is what we think in all, and there are some bounds. We'll stay within those bounds. And we have different ways of staying within those bounds. So for example, if we see that putting video call in Super really destroys our Max business or something, which is like really terrible, then we'll start doing something like metering video call in Super, like you only get access to 1 per day versus in Max unlimited, something like that.

Again, I'm talking about things in the future because -- so I just don't know exactly what's going to happen, but we're going to stay within the bounds of what we have in the guide.

I don't know, Gilian, if you are -- if we can hear you now or not. No unfortunately.

Not quite. Okay. Maybe Luis, just a more quick one for you. You talked about sort of the math and thinking that, that's an opportunity to build and sort of have one of the math -- best math tutoring apps in the world. As that gets developed, would you look at sort of new channels in terms of how to sort of distribute that? Like I think in the past, Duolingo for schools as a way to sort of open up that population on the language side. Is that a potential viable channel as you think about areas where you could get this DAU growth from something like math?

Yes. Math is an interesting subject. On the one side, they're really -- I don't know of another subject other than languages that has as much of a market. I mean math -- there's 1 billion people learning math. So that's awesome. Now they are different than languages in the sense that these 1 billion people generally don't want to learn math, whereas the language people actually want to do that. And these Duolingo people are mostly in schools. In fact, they're mostly in K-12 schools. That's where they are. So that is definitely where we're going to get to now. There's multiple ways to imagine how to do that.

Our first step is to just make this product. We're almost there. I mean if you look at the internal versions of our math course, they are incredibly good. Once that happens, there's a big question about go-to-market. At the moment, our hypothesis is that the easier way to get started. That may not be the ultimate thing we do, but the easier way to get started is to think of it as a supplemental thing for school. So rather than get deals with school districts and stuff like that, think of it more as a competitor to Kumon. That is, we believe, the better way to get started because there's a very large number of people that do -- have some sort of help outside, et cetera.

By the way, what we like about this, and this is not -- you're not going to see this, this year, but what we like about this over the long term is that the propensity to pay, not from the kids, but from the parents for math is very high. So we'll probably be able to do something there.

Do you hear me now, Luis?

Yes.

Okay. Good.

Gilian is back. I'm glad. Otherwise, they would have started asking questions about acronyms that I don't know, and we are in trouble.

I know. Some nice people just snuck around my computer. We're better now.

We'll take our next question from Shweta Khajuria with Wolfe Research.

I was going to ask Gilian very specific three questions while she was muted just to test her mic. No, actually, I have two questions. One is on retention and the other is on AI. So on retention rates as we think about you investing in math and music and chess and perhaps there are some cross-platform benefits. How do we think about churn and/or retention rates over time as you invest in the product? My understanding is that it has remained largely stable. But is there a reason to think that it actually improves markedly and what that can do to your platform?

And then the second is on AI. So you addressed this in terms of competition where you have a market lead versus anyone else in the market right now. But what AI could potentially do is accelerate the path to creating new competitors that don't even exist today. So could you talk to potential risk of a competitor emerging because AI is making it easier to create certain products?

Yes, sure. In terms of retention, I'll let Gilian say more stuff. But generally, our retention, especially for payer retention is pretty stable. It's very healthy and very stable. There's some chance it may go up. I don't think we're modeling that. I'll let Gilian know because I don't actually know if we're modeling that or not, but there's some chance it may go up over time. It's because one thing we do know is that multi-subject users have higher retention. And the other thing that is true is that over time, more and more users are learning multiple subjects just because we offer more subjects. So there's some chance that it may go up, but I don't think -- so far, this is -- you shouldn't really believe me on this because it has remained pretty flat.

I don't know, Gilian, if you want to add to that, then I can go to the new competitors on AI.

I mean retention, it looks pretty flat over a fairly long period of time. And we look at it in aggregate, and we look at it via cohorts. I think the perspective we have is that it can stay where it is. That said, I think you're pointing out something really, really important. All the work we're doing to make the product great, so that word of mouth and top of the funnel and all that stuff is keeping the size of the pool bigger. That also helps current customers, too, right? They get those features, too. So I think it's a possibility, but we haven't modeled anything significant on that front on.

Yes. And in terms of your question about can a new competitor show up because AI allows the building of something faster. So it is true that with AI, you can build stuff faster. And I assume that, that will be more true over time. That is an assumption that we have. The good news is that, that also applies to us. So we can also build stuff faster. So that is good.

The other thing to say is there's a little bit of a misunderstanding, I think, in the market about how easy it is to recreate a very sophisticated app like Duolingo. The reality is that if I were to write down just the spec of Duolingo, just -- and I'm not talking about recreating it in a programming language, like write it down in English, it would take me years because there are 1 million corner cases of things that we do that we're really smart about how -- what we do for the user to get them to be more engaged or teach better, et cetera. So I'm not particularly worried about somebody just magically pressing a button and an app appearing that is just better than Duolingo.

On top of that, it's pretty hard. The reason Duolingo so far has not really gotten competition at a large scale, again, we have like 85% of the daily active users of language learning apps in the world is because it's pretty hard to compete with our free product in the sense that we have really huge distribution. And a new entrant, when they come in, they need to figure out how to grow. What they usually do is they try performance marketing to try to grow fast, but because our free product is so good, they usually cannot make their performance marketing work if they are to have a free product. So what then they do is they put a paywall on it. And then that just really caps them. So it's just been really hard for entrants. I mean we see -- look, the reality is that over the span of Duolingo, we have probably seen, I don't know, 400 apps come up and try to compete with us. And I'm making that number up, its hundreds. And we just don't see them grow again because we have a very large distribution.

We'll take our next question from Eric Sheridan with Goldman Sachs.

Maybe building on Shweta's question and asking it a little bit differently. When you think about the AI landscape, what is the scope by which you think you need to sort of be a leader in AI learning in the 4 walls of your application and your experience as opposed to possibly cooperating with LLM agents and being an application layer to their platform if more consumer behavior goes in that direction from an aggregation standpoint? I'd love to get your sort of philosophical view on that.

And maybe the second question would be you guys were probably one of the leading examples of a mobile-first company with Web 2.0. How should we think about the duration and the need to invest to reposition the company to be AI native in its tech stack over the next couple of years?

Yes. So good question about does it make sense for us to continue in our own app versus something like an LLM. I'll tell you what we do, just historically, there have been a number of new platforms that have happened since we launched Duolingo. For example, smart watches came out. Also smart speakers came out like Alexa. In all cases, we had a lot of people asking us questions. And certainly, the companies that were building these things telling us, this is the next big thing, everybody is going to be in there watch. You should be in the watch. In the past, we made the mistake of early on making an app for the watch or whatever. And only to realize later that apps for watches are not really a thing other than exercise apps for watches.

So the way we see it is if it is the case that consumer behavior changes quite a bit so that nobody has their own app and then everybody is over there, we will definitely do something about it, obviously. What I'll say though is that our app in particular, I mean there are many other apps, but you can think of our app a lot like a game. I mean it really matters that we are not text-based. It just turns out that people really, really like doing things like playing Candy Crush or doing Duolingo in part because of the animations, et cetera. So if you're just asking, are we going to be a thing in a chat bot? I just don't think you can get the engagement that we can get inside like a text-based chat bot.

Now again, I just don't know what the future is going to look like. But this is -- at the moment, we feel pretty confident that the right investments are to continue making a beautiful experience that is very visual because that's just what keeps people engaged.

And in terms of AI native, I mean we've always -- ultimately, we are a technology company. We have been really trying to adopt AI and technology as fast as possible. In fact, we've probably gotten in trouble for adopting AI fast, for example, by writing stupid memos that I wrote one time. So that's -- and also, I think we're just -- you're going to see us always take the next step with things like AI.

This concludes the Q&A section of the call. I would now like to turn the call back to the host for closing remarks.

Well, thank you, everyone. These were excellent questions. And I just want to say that we're very excited about 2026. It is probably not exactly what investors expected. But if you are thinking about Duolingo as a long-term company, which obviously I am because that's kind of what I want, this is really the right investment to do. The second thing that I'll say is that Gilian started her job 3.5 days ago, so it is incredible the amount of knowledge that she has about our finances, which already surpasses mine. So thank you, Gilian, for that. And thank you, everyone.

I'm so delighted to be here. Thank you, everyone.

Good evening, everyone. Welcome to Duolingo's Third Quarter 2025 Earnings Webcast. Today after market closed, we released our Q3 shareholder letter, which you can also find on our website at investors.duolingo.com. Today, we have Luis von Ahn, our Co-Founder and CEO; and Matt Skaruppa, our CFO.

[Operator Instructions] Please note that this event is being recorded. [Operator Instructions] As a reminder, we'll make forward-looking statements regarding future events and financial performance, which are subject to material risks and uncertainties, some of which these risks are set forth in our filings with the SEC.

These forward-looking statements are based on our assumptions that we believe to be reasonable as of today, and we have no obligation to update these statements as a result of new information or future events. Additionally, we'll present both GAAP and non-GAAP financial measures on today's call. These non-GAAP measures are not intended to be considered in isolation from, a substitute for or superior to our GAAP results and we encourage you to consider all measures when analyzing our performance.

And now I will turn it over to Luis.

Hi, everyone, and thanks for joining us today. We had a strong Q3 with solid performance across all metrics, and we're on track for another exceptional year. More than 50 million people now use Duolingo every day. And we're guiding to nearly 1.2 billion in bookings this year with 33% growth and an adjusted EBITDA margin of 29%.

Putting that all together means that we have rapidly scaled our impact while expanding profitability. And yet, we still feel early in our journey. We believe AI will fundamentally transform education, and we have line of sight to building a product that teaches better than ever before. That's what makes this such an exciting moment for Duolingo.

And now we'll take your questions.

[Operator Instructions] Our first question will be coming from Bryan Smilek with JPMorgan.

Luis, just to start, can you just help us better understand the underlying drivers of DAU growth in the 4Q and engagement overall? And how close do you think you are to getting back on track in terms of marketing in the U.S.? I know you mentioned growing impression volume. But is that starting to translate to user growth now? Or how should we think about that into the fourth quarter?

Yes. Great question, Bryan. Thank you. So we just posted 36% year-over-year growth for Q3, and we're happy with that. There's puts and takes of how we got there. I mean, on the positive side, we have things like the locking partnership that we just -- that we talked about in the shareholder letter, which helped us grow users in Asia quite a bit. We also had a number of product improvements that helped with retention. So that was great.

On the other side, just like we said last time, we passed all the unhinged posts in our social media for a bit because we were listening to our community and trying to build brand love. And when we don't post unhinged things that basically our posts were much less likely to go viral, and because of that, that did have an impact on DAU growth.

The good news is that over the last few weeks, we have started the unhinged posts again in our social media accounts. And while it hasn't gotten all the way to the peak where it was, we've seen a lot of recovery. So that's really starting to show up. And we do expect that to affect DAUs positively.

In terms of Q4 for DAUs, we expect some amount of deceleration from Q3 but we're pretty happy with where it has stabilized. What I'll say -- we're not guiding to Q4 DAUs, but what I'll say is that September and October were both at around 30% year-over-year growth for DAUs. And that's comping a pretty strong quarter last quarter -- last year.

Great. That's super helpful. And I guess, Luis, you also mentioned 3 core areas: monetization, user growth and just teaching efficacy overall. Can you just elaborate on really what's driving the decision to shift investments towards long view? Is it the AI opportunity or how should we just think about this impacting bookings growth? I know you've obviously talked to 25% plus as your North Star years ago. So just curious how do we get back there over time?

Yes. I mean, look, we're -- like you said, and like I said in the shareholder letter, there's a huge opportunity right now. We see a huge opportunity -- over the next few years, education and the way people learn, they're going to change fundamentally and it's because of AI.

And also because of AI, we see -- we have line of sight now to create an app that can teach really, really well, much better than anything that humanity has seen before. As good as a human tutor, but that is also more engaging. And if we're able to do that, right now, we have, I don't know, we just posted 135 million monthly active users. If we're able to do an app that teaches just that well, way much better than we have now, we will be talking about billions of users that we have and that's what we want to shoot for here.

So this is why we are investing in the long term. And what that looks like is that we are putting relative -- more relative investment in things like teaching better, which -- teaching better -- if we teach better, what that does is that, that helps user growth, but there's a lag.

Just whenever you improve your courses, users do grow, but it takes a while for that to happen. And then user growth, there's a lag to get to monetization because people take some time to subscribe. So this is kind of a long-term thing, but we're very bullish on this, and this is why we're doing that. And the goal here is because the opportunity is so large, the goal here is to be growing DAUs fast for a very long time.

Our next question will come from Nathan Feather with Morgan Stanley.

Given the scale of the opportunity ahead that you're talking to, is the focus on greater long-term prioritization, a durable shift in strategy that we should expect to continue? Or would you plan to return to the current prioritization mix at some point? And then in connection with that focus, should we expect the pressure on bookings that you're seeing in 4Q to continue into next year with the durability of that shift?

Well, okay. So the -- for -- it's going to be years until we get to a point where we have an app that I think is just the best possible way to learn any major subjects. So we will be investing for a while. And that's important to know. I should say, though, we're not really guiding for example, to next year, but we're very excited about a lot of the initiatives that we're going to put out in the product for next year.

We're going to have much better video calls that are for beginners. It's something we're calling guided video calls. The app is going to be a lot more social. We're going to have all the -- right now, the math course only has a little bit of content. In the next few months, we're going to have all of the common core K-12 content in the math.

Our chess course is going to have player versus player and it's doing super well and growing really fast. We're going to have a full revamp of our music course. And for the top 9 languages that we teach, we're going to be able to teach from 0 to Duolingo score 130, which is where you can get a job in that language. So there's just a lot of things that we're very excited about. And I don't know, Matt, if you have something to add to that.

Yes. I think, Nathan, I want to put it back in the context of kind of what we've said before. So we've talked to you all and our investors for a long time about wanting to maximize platform LTV. And what Luis is describing is a change of small proportion right now that is going to help us grow users for a long time, get users to do more lessons, learn or spend more time on the app and, in general, be more engaged Duolingo users.

And we think that if we do that and we do that effectively, we'll both grow users for a long time and will increase platform LTV for a long time. So I think that will -- we're down to bookings and the financials ultimately as well. I think there's a -- we guided to a really strong 2025. There was a -- so we think that this can exist like it has strong financial performance.

And this is just a balancing act that we've always done, and we're telling you all that we want to make sure that at the present moment, we're balancing it so that we can grow really rapidly for a long time in the future.

Our next question will come from Wyatt Swanson with D.A. Davidson & Company.

I'd love to hear a bit more about how the new chess course is progressing. I think at Duocon, you mentioned millions of daily active users. On that front, do you see any differences in engagement or retention for users in the chess course versus your core language forces? And as it relates to that, kind of curious about your new PVP offering? Can you talk about like when you expect that to be fully rolled out?

Yes. So needless to say, we're very excited about chess. It is the fastest-growing course that we had. It's growing much faster than math and music and faster than the way originally languages grew. It's true we have millions of users. We're not saying exactly how many, but it is -- it's already surpassed math and music.

The retention of chess users, it hasn't been around for all that long. Our chess course has been around for 3, 4 months. But so far, from what we can measure over the last 3 to 4 months is slightly higher than language learning and so we're very happy with that. We also -- as we mentioned, we started rolling out PVP. That means player versus player, so people being able to play with other people.

At the moment, 50% of users on iPhones, on iOS can see it. It's not yet on Android, but it's going to come out on Android pretty soon. So we expect that within the next few weeks and small -- few weeks/small number of months, every person that has the Duolingo app will be able to do PVP chess. And over the next year, we just -- we expect quite a bit of growth from chess, and I'm very happy with it.

Great. And then just one quick follow-up. What does prioritization of user growth instead of monetization look like? And how should I think about the actual changes in the app?

Yes. I'll give you an example. That's just -- okay. We've always had to make trade-offs between -- whenever we run an experiment, some experiments improve all metrics. Great. That's an easy call, just launch it because it improves all metrics and that happens.

But there are times when experiments improve one metric but hurt another. I'll give you a fictitious example. If right now, a free user -- free users get 25 energy units at the beginning of the day and every exercise that they do spends 1 unit. If we were to do an experiment that decreases that from 25 to, say, 24. That's 1 fewer unit of energy per day.

We know that would make us more money. It just does because more people run out of energy, so more people end up wanting to pay to subscribe. However, we also know that would decrease daily active users because it would frustrate some of the users. We've always had to make decisions about different judgment calls about this. What we mean is that what we -- the change that we are doing is that we are going to be prioritizing user growth over monetization in this type of judgment call. So in the fictitious experiment that I just gave you, we would not launch that experiment going from 25 to 24 energy units even if it meant quite a bit of bookings gains, if it has a real hit on daily active users. That's the type of stuff that we're doing now. And again, just to remind you, the reason we're doing this is because the opportunity ahead is so big that it's just good for us to grow fast for a long period of time.

Your next question will come from Ralph Schackart with William Blair.

Luis and Matt, kind of going back to the line of questioning here. I guess maybe the question is like why now, what signals are you seeing on the shift or maybe the semi shift focus more on growth over near-term profitability as AI advancements, kind of what's prompting this? And then can you give us a sense of the duration of this pivot or shift? Is this something it's going to take -- I don't know all through 2026? Is it more short term in nature? Anything you can add there would be great.

Great. As to why now is -- I mean, it's a great question. The reality is that over the last couple of years, it has just become progressively clear and clear that we are in a unique point in time, particularly with education in terms of how education is going to happen in the world and also how well we can teach at Duolingo. We just see it in our own metrics in how fast we can put out content with things like video call. We just see how much it is improving every month.

And so that just kind of has been coming for a while. And what has happened is that over the last month or 2, I've really rallied the company towards this shift. And really, it's like, okay, it's not like it was one day where I woke up and decided let's do that. It's just -- we really rallied the company to say look, opportunity is huge for us, let's prioritize, making sure that we can grow for a long period of time and also making an app that can teach really better than anything that we've seen before.

As to how long this is going to take. This is -- it's an interesting question. I mean, I think you're asking something to the effect of like, well, is this going to hurt bookings and is this going to hurt bookings forever? I don't think that's the case. It's just -- it's going to take some time for us to see results -- financial results over these long-term investments that we're doing.

But we're going to be acting for a while like there is a humongous opportunity because there is one until we get it. But I think we're going to be seeing good results from this even much sooner than that. So it's not like our -- we're saying, oh, throw away all the bookings or anything like that.

Yes. I think the only thing I'd add to that, Ralph, is that, I mean, you've seen us navigate this trade-off over the past 3 years as well, users have grown 55% per year on average over the past 3 years, and bookings has grown about 45%, and all of that while we were making similar trade-offs. And now we're just slightly focusing a little bit different as we navigate those. So it's not that we haven't been making any of them.

And you've seen that like in the rest of the year guide and the Q4 guide, there was a small impact to this. It's not that big. And so would we expect some of that to persist into 2026? Sure. But again, I think as a general framing of this, it's a relatively small financial impact from this kind of reprioritization. And we think that, that's worth it because, as Luis said, it's a huge opportunity. So the risk reward seems right.

Your next question will come from Alex Sklar with Raymond James.

Just following up on Wyatt's question and maybe, Ralph's, your remarks to Ralph at the end there. Just in terms of framing how meaningful some of these changes might be, is it as simple as maybe focusing a little bit less on paid conversion just to improve the premium experience today? Or are there kind of broader thoughts about maybe moving video call down into some of the lower packages? And then I've got a follow-up.

Okay. So a way to see this is what you said first. It is as simple as in some of the experiments which, by the way, not every experiment, many experiments we run, just improved our metrics. This is good. In some of the experiments where judgment calls are needed, we're going to shift the balance a little bit more towards user growth. It's not a humongous change, but it is a change.

In terms of are we going to move video call to other tiers, et cetera, that is likely to happen, at least we're likely to attempt to do that, but that is unrelated to this. We were anyways -- we're always thinking about moving different features in the different plans and we'll test that.

It may be the case that, I don't know, video call, but it may be the case that some of the Max features are better in super or even in the free tier, and we'll test all of that. And while we test that, we are trying to optimize lifetime value of our like platform LTV. We're trying to optimize for that. So you may see us test stuff like that.

Okay. Alex, I'm glad you said the free-to-pay conversion because that's exactly how like it really manifests in the business. And I just want to make sure we're clear on this. So for example, when Luis said over the past couple of months, he mobilized the company. In September, we saw some of this and what it looked like was slightly lower free-to-pay conversion, but that free-to-pay conversion was still growing year-over-year.

So it's still good free-to-pay conversion. It was just on the margin. It was slightly lower. So I think that's an example as you think about the financial impact. You're right to point out that that's how it would flow through.

All right. That's great color there. And maybe for a follow-up, Luis. Just on the last call, you brought up this idea on video call about average number of words spoken per session. And that was kind of a new metric you were going to start testing towards. What have you learned so far now that you've kind of been optimizing for that metric? And then what's kind of the time line to get some of the changes into the product as a result?

Thank you for asking that question. So yes, this is actually a great metric and we've managed to move it. This is important to know about our company. Whenever we fixate on a metric, we are very good at moving it. And this particular metric, we have been moving at -- it's more than doubled this year in terms of average number of words spoken per Max subscriber and so we're very happy with that.

In terms of changes to video call that you'll see coming soon, one that I'm very excited about is video call at the moment is a monolingual experience in the language that you're learning. So if you're learning Spanish, it's all in Spanish. That's great for practicing Spanish. But for beginner users, it's too hard. If you only know 20 words, it's very hard for you to have a full conversation just in Spanish.

So the thing that we're testing now is these things we're calling guided video calls, which are basically bilingual. So it's -- if you're an English speaker learning Spanish, this would be part in English, part in Spanish, and it's a lot easier for beginner users. We're seeing that when we give that to beginner users, they actually speak more words per call because they're actually able to do something.

And so we think that this is going to really help with Max conversion, by the way, I should say 2 things. Most of our users and certainly most of our Max subscribers are beginner users. And so we really think this will help with mass conversion. The other thing that I'll say is that these guided vehicles, we are not advertising them yet or we're not using them converting users into Max subscription yet. So we've put them out and the next step is kind of to tell nonsubscribers that these exist so that we can get them to subscribe. So we're pretty excited about what that can do to Max.

Your next question will come from Ygal Arounian with Citi.

Just on AI and making education sort of better than ever, the way you're talking about things. Can you just -- does that accelerate your road map in terms of adding new language learning modules. I know within the ones that you currently have, but moving into new subjects and -- what is it about what's changing right now around AI that's letting you do that today?

Yes, the types of things you will see. It definitely accelerates our road map in more coverage of languages. That doesn't mean new languages. The reality is languages, it's very love sided what languages people want to learn. We now teach 40 languages. The rest that we don't teach is very little demand for them. But the top 9 languages that we teach, these are kind of like the Spanish and English and French and German and Italian like the big languages that people want to learn. The top 9 account for the vast majority of our users.

And what you'll see us do is you'll see us go faster in terms of adding content to these top 9 languages. And right now, for most of them, we don't get you to the place where we want to get you, which is the Duolingo score of 130 in which is equivalent to CFR level of B2, which is where you can get a knowledge job in that language.

You will see that over the next few months, we're going to be adding content that can do that for all the top 9 languages. The other thing that you'll see is you'll see us just at a lot more different modules in the way we teach languages that are just a lot smarter at teaching you. I mean they're going to adapt a lot better to you. And you're also going to see us just use a lot more things that use AI in the background to allow for many more free responses, so that it adapts a lot more to you.

In terms of -- we're going to also be using AI for other subjects. We're using it pretty heavily for math for getting a lot more content out there. So we're going to do that. In terms of adding other subjects, at the moment, we're not working on any other subjects. I'm not going to say that we're not going to add other subjects next year. That may be the case. Like you saw with chess, it took us 9 months from idea to actually launching. So it is possible that we'll add other subjects next year, although a little unlikely, but it is possible. But at the moment, we're not working on any other subjects.

Okay. And then maybe sort of another broad one on AI. Can you just talk about what you're seeing around compute costs, gross margins coming in a little better than expected. And is that coming in faster than you think? And how is that impacting? One of the big questions we get is just generally on the competitive landscape and how AI is evolving that? What are you seeing there?

Okay. In terms of cost, look, costs are coming down. They've come down just without us doing anything, costs are coming down. For us, this has not been the top priority of optimizing costs. At the moment, the top priority is just making the best possible experience for our users that teach us the best and that is the most engaging.

Every now and then, if we see low-hanging fruit we will -- in terms of optimizing costs, we will do it. But it's not like we have all of our people trying to optimize cost. And the reason we can do that is because most of our AI features, at least the ones that cost the most money are behind Duolingo Max and because the price of that is high enough that we're -- for us, the usage of AI is anyways profitable. So that's why we're not going -- not trying to optimize the cost on that.

In terms of the competitive landscape, I think people say things like -- the 2 things that people say about the competitive landscape with AI are: number one, why would anybody want to learn a language with Duolingo when you can just learn it with ChatGPT. Okay. We're not particularly worried about that. We've said it before. The main thing that we do really well not only do we teach well, but the main thing that we do really well is keep people engaged.

And in order to learn a language, you need to be engaged for years. It really takes years to learn a language coming every day and we need to keep you engaged actually doing it. And not only that, we also need to have curriculum for years for you to do that. So with ChatGPT, you can go there and you can ask it to teach you a few words here and there, but it's not like you can have really curriculum for years that teach that.

So we're not particularly worried about that aspect. And then the other thing that people have said that they're worried about is, oh, well, nobody is going to want to learn a language because we're going to have simultaneous language translation and okay. Also not worried about that.

I believe in 100% of the Google I/O conferences over the last 10 years, they have showcased simultaneous language translation. They do it every single year, and it's good. It works. But this has been happening for the last 10 years, and we have not seen the desire to learn a language go down at all. In fact, it has come up.

And I think the biggest reason for that is because if you look at our users, they fall into 2 big categories. One big bucket is people who are learning a language as a hobby. It kind of doesn't matter whether our computer can do that because -- they're the same with chess, by the way. Computers are way better than humans at chess, but still we have millions of people wanting to learn chess. So it doesn't matter if it's a hobby.

The other big group of people that are learning a language with us are people who are learning English and they actually want to learn English. Like that is -- for them, being able to have like a phone that they have to hold out, it's just kind of -- that's not what they want to do. So we just -- we're not particularly worried about that. It just so happens that people like to tweet about that. We're not worried about it.

Just a couple of points on that. Since we -- the AI costs and Max, just to make sure everyone's aware, Max is now 9% of our subscribers. It doubled in Q3 year-over-year in terms of bookings. So it's clearly doing well in that regard. It's underperforming our lofty expectations for it, though. We expected a bit more than that. And so that's why Luis is talking about guided calls and all the other things we're going to do to help it achieve what we think it can achieve.

And then finally, because AI costs have come down, though, I don't want folks to take away that we're not willing to invest as Luis is talking about the seminal moment we're in to go attack a very large opportunity. We're -- we've shown that we can grow to this scale incredibly profitably. We're guiding to a 29% adjusted EBITDA margin for the year, which is very, very close to our long-term adjusted EBITDA margin range. And so as we do that, we are not going to be afraid to invest in innovation. And so we're going to make those investments over time.

Your next question will come from Mark Mahaney with Evercore ISI.

Okay. A couple of things I wanted to go through. One, I know you had some price actions or price increases earlier in the year. Have you -- what kind of reactions have you seen to those? Secondly, I think you were going to hold off on doing any other -- I think it was just on like new people coming in on the standard plan, but your thoughts on rolling out other price increases. I don't think you're going to do that now given your prioritization of user growth in the fourth quarter and beyond, but just talk about that.

And then third, just so we're clear on the deceleration in bookings and revenue growth in Q4 is largely due to the fact that you're going to sort of slow down the conversion rate from free to paid and really just focus on user growth. I just want to make sure that that's the main driver, and it's not like you're seeing a reduction in retention amongst paid subs, higher churn amongst paid subs possibly because of the price increase.

I mean I can take some of those. I'll take the last one. Yes, the change in Q4 is pretty much because of the shift to go to longer-term initiatives. And that means user growth and also spending some of our -- not only are we prioritizing user growth, we're also spending relatively more effort, shifting some effort to teaching better which we're taking some of that from our monetization efforts.

So that's basically what you're seeing. And it's not like a humungous thing, but it is a shift. In terms of price increases, you'll see us -- we'll be testing prices. We'll be testing all kinds of things. And we will see us launch the things that we think are good for the whole platform. I don't know exactly what's going to win, but you'll see us test prices. And by that, I mean up and down.

We'll probably test some prices much lower or a package -- and this I'm speaking about things that we may do, but I don't know if we'll end up doing them if we don't like the results. But we'll probably do a package that is like half price, that is like super light. That's the type of stuff that you may see us do.

Yes. And Mark, just to round it out, I think 2 points. One is, we've talked a lot about taking price up as a price point every now and again, we do that. Luis just said, we're going to continue to experiment with that, and that will continue to happen. ARPU has gone up this year, every quarter, kind of mid-single digits. That's also reflected in the guide and that's mainly come more from Max than it has from price point changes.

I will say that we did take pricing at various times over the past a little bit and that does influence our ability to discount during our one and only discount of the year. So again, if you have a higher price, you can run different experiments with the level of discounting and one of the things that happens every Q4, when we talk about it, as we talk about the ability in our Q4 bookings guide because Q4 is our most variable quarter because we run this New Year's promotion.

This year, we did energy, which is a core pricing mechanic, and we have never run a New Year's promo with energy. And so we're going to be experimenting with all sorts of things as we get towards the end of the year with the promo on how we run it, how we show it and display it, when we run it, all of these type of things. So that is also baked in here in the guide.

Your next question will come from Ryan MacDonald with Needham & Company.

Luis, maybe stick with me a little bit on this question. I apologize in advance. But how learners -- people learn differs by generation and by demographic. I think we've already seen that with sort of advanced English learners require -- meeting different requirements from Duolingo than maybe the traditional core base.

So can you talk about how -- or if you are going to be targeting certain demographics, Gen Z learns or generations, Gen Z seemingly learning different than millennials with some of these product updates? And then how should we think about the metrics that you will be looking towards to prove out this works?

Because obviously, user growth can be beneficial -- can be benefited by some changes, but that might not always mean that, that sticky user growth where MAUs might not always convert to DAUs. So again, long-winded question, I apologize. But how are you targeting or what are you targeting in terms of these changes? And then how are you measuring success?

Yes. I mean you asked about -- it's true. Some people learn different than others. That is true. But you would be surprised that there are a lot more similarities than differences. The reality is, I mean, this is not just generations, also geographically. I mean we always hear these things about like, "Oh, well, people in that country do that or people in that country do that."

What we have found time and again is that not only a lot of people learn pretty similarly, also the things that get people to use the product more are pretty similar across the geographies. I mean like a streak, it works in every country or it's just -- so there's a lot of similarities. So at the moment, you're going to see us just make a better course for the masses that's what we're going to be spending most of our effort on.

Of course, the courses do adapt to each individual and probably one of the places where there is most adaptation that is needed is the pace of learning really is different. And it just happens that as you get older, you get slower. That is just -- that's not controversial as somebody who's getting older and slower, I can tell you that.

So the pace -- but that's very easy to adapt. We're just -- we really do just adapt to the pace pretty easily, and we've been doing that for a while. Now in terms of the metrics that we're going to be looking for, certainly, user growth is an important metric that we're going to be -- that we're really keyed in on right now, probably the most important metric in the company. So we're going to be looking at that a lot.

Now the thing about improvements in teaching, and this is what I was saying before, they don't translate to user growth immediately because if you improve a course and it's much better, over time, maybe people are starting feeling that they're learning a little better, so there's more retention or maybe there's more word of mouth because the people are saying like, it really works for me. Let me tell you about it.

So it does translate. We know that improvements in teaching do translate to user growth, but it's not immediate. And this is kind of what we mean by long term. What we're going to be looking at that, there are things like just improvements in learning outcomes, we can measure how well people are learning.

And the good news is that really almost every year -- since we started measuring that every year Duolingo is actually teaching better than the year before. We're probably going to see improvements in how well we teach, move faster than in the past because we are taking it -- we're spending more effort on it. So we're probably going to be seeing that and our hypothesis, but it is a hypothesis that I very much believe in, is that, that will translate to user growth. It's just not going to be linear or quick.

Makes sense. Okay. And I'm also older and slower, Luis. So no problems there.

Aren't we all? Aren't we all? But wiser, but wiser.

We hope. We hope. Matt, I know you probably don't want to obviously get into a conversation about 2026 guidance or anything like that. But obviously, a lot of course investments, a lot of content investments. Can you just give us a sense of like the magnitude we should expect here? And like is there an expectation still that you can continue to expand EBITDA margins even through this process as we think over the next couple of years?

Yes. No, I appreciate Ryan. And even though you asked a very [indiscernible], I'm not going to guide to 2026 on this call, we'll do that in February. But I will -- and I mentioned this on an earlier question, we have made incredible progress towards our long-term margin while we've been growing the platform to a scale that has 135 million monthly actives, and we guided to nearly $1.2 billion of bookings this year.

So we've [indiscernible] everyone that we can scale the business, operate with discipline and expand margins. And that's great. And we want to continue to do that. And we're going to continue to operate with discipline and grow the business along the lines of what Luis is talking about, hopefully to a whole other order of magnitude.

While we do that, though, we're not afraid to invest. We think we can invest and still operate very profitably. But we are not going to prioritize linear margin expansion from here when we should be -- we view the opportunity is so big, so we can prioritize investing. That's how we think about it.

Your next question will come from Andrew Boone with Citizens.

I'd love to talk about kind of a topic that we've revisited maybe 2 or 3 years ago, right, in terms of advanced English learners and basically the improvement of efficacy driving those learners to the platform and kind of like whether AI has accelerated that and whether some of that thesis came to fruition and whether any of the slowdown in growth is impacting those types of learners. Or is there anything else you can kind of speak to in terms of that cohort?

Yes. As we mentioned -- I don't know whenever it was, a couple of years ago, we started talking about this. English learning is a major opportunity. I mean this -- 80% of the people who are learning a language in the world are learning English. And so we started investing in teaching better for English learners.

We have done that -- we've done a major launch here throughout this year. Now all of our English courses now cover up until Duolingo score 130, which is the place where you can get a job -- a knowledge job in that language.

So we have that, and we've been improving how well people learn English and we are seeing that in the metrics, our number of English learners and certainly a number of advanced English learners has been growing steadily. The other thing that I'll say is that the regions at the moment that are growing fastest are English learning regions. And so we're seeing that Asia is a really good example of that.

There's -- we are certainly still posting or have never stopped posting unhinged content in Asia kind of from the social media side, but also we're just getting a lot more users there. So that's the parts that are growing the fastest. So I'm pretty happy with the progress there. We did mention that it was going to take a while, and it is taking a while for really the word to get out that Duolingo is very, very good at Advanced English and so while I believe that we've made good progress in that, we're still not there yet as this is a thing that even throughout next year, we're going to be seeing an increase in the number of advanced English learners that we're going to have.

And then can we just get an update in terms of family plan? I know there were a bunch of features that you guys were adding and then just broader adoption, where are we today? And where can that go?

Yes. Yes. In terms of broad adoption, the Family Plan continues to do well. I think it was about 29% of subscriber [indiscernible] in the most recent quarter. So it's grown nicely. And I think there's a bunch of reason to believe that it's going to continue to grow nicely.

For example, last year -- in Q4 of last year, really everything basically went right. which is why we grew so fast last year, I think, 42% year-over-year last year. Part of that was that the family plan during New Year's promo did really well. So there's reasons to believe that we still have room to run on Family Plan, Andrew.

Your next question will come from Justin Patterson with KeyBanc.

I'd love to hear a little bit more about just learnings from energy. There was a lot of optimism on this product, roughly 90 days or so ago. So I would love to hear how just that has evolved and might be playing into some of the trends we're talking about on this quarter. And then when we think about just the arc of reacceleration here, how much of this is really dependent on getting the product right versus really tapping into that cultural relevance that Duolingo had earlier this year and then and encounter that speed bump around the social backlash?

Okay. So energy, we're very happy with energy. It did exactly what we wanted it to do. It increased bookings and also increased DAUs. So that was good. The kind of the way it went is we launched all of energy for our iOS users first, then for Android.

We rolled it out over a span of a couple of months for each one of these platforms. By now, it is done. We have, I don't know if it was a couple of weeks ago, we have every single person who has updated the app in the last, I don't know, 6 to 9 months has energy in there. And we're very happy because it did exactly what we expected it to do, and it's been launched.

Now in terms of -- you asked a question about product versus kind of marketing, look, both are important for us to grow, product and marketing. The longer-term thing is if we have a product that is extremely retentive that also teaches really well, that's the best thing you can do, and that's where we spend most of our efforts, and we're going to continue spending most of our efforts on that.

Yes, the cultural relevance matters. But to me, that's just an accelerant to something that is where the main dish is the product. In terms of kind of being culturally relevant, et cetera. We really are seeing a complete pickup on that. I don't know if you've watched, for example, the things that happened in Halloween.

I mean there were throughout the world, certainly, there were thousands of Halloween costumes that were just Duolingo costumes. So we see that -- we're seeing that quite a bit. And so we're not particularly worried about that.

Yes. And the only thing I'd add to that, Justin, is there is -- we talked about it on the last call, which is we've had some success, particularly in the U.S., spending a little bit more on actual marketing, seeing a nice return in the U.S. in particular.

And so that happened in Q3, and we're going to -- it's relatively small dollars, but we're going to continue that into Q4 because the U.S. growth, as we've talked about before, is slower than the rest of the world, in part because of the rest of the world. is growing really rapidly. But that continues. And so that's just the other element we'd add to the marketing mix.

Our next question will come from Shweta Khajuria with Wolfe.

I actually have 2. First one is on China. If you could please talk about how engagement has trended in China through the quarters, through the year and what your expectations are, what you're seeing there? And even if you could comment on likelihood of Max there this year or next. And then second is on renewal rates at Max in particular. Could you please talk about how that trended in Q3 so far and what your expectations are?

Okay. Let me take the China one, and then I'll let Matt take the renewal rates one.

Okay. China. We are, of course, doing very well in China. I believe it's our fastest-growing country. It is our second largest country in terms of DAUs and growing fast. It's still not that large of a fraction of our business, it's about 5%, 6% of our business at the moment, but it is growing, and it is growing very fast and the engagement is very high, retention is high.

Max in China is being tested at the moment. And so we have the permission to test it in China because of this is an LLM you need to get permission to test this. So it is being tested. And so this is going to launch in some number of weeks or months, it's hard for me to tell you exactly because it does depend on approvals, but it's moved along pretty well.

So what I'll say at the very high level about China is that it's a pretty major opportunity for us. But of course, China comes with a risk. There's geopolitics, et cetera. So the way we're treating it is we're not spending a crazy amount, for example, in marketing in China. We're spending a very modest amount there, but we just happen to be growing quite a bit. So it's nice and it's a really nice opportunity.

But if at any point in time something happens, we just didn't end up investing all that much there. That's kind of how we're treating it. And then I'll let Matt talk about Max renewals.

Yes. So just to put it in context, Max renewals are important, but the broader -- I think there was a question earlier on broader platform retention. That remains strong, no real changes in that. And then on Max, in particular, we mentioned in the last call that we were going to start to see in Q3 and then in Q4 larger cohorts come through the Max renewal cycle.

And what we saw is that Max right now, again, there's still small as they've been ramping up, but was renewing slightly better than Super. But it's early, and Q4 is relatively a relatively sizable cohort. So we'll talk about this again on the next call.

Your next question will come from Hanyi Cai with CITICS.

It's so good to have you connected. And it's really excited to hear the last question from like Max is finally rolling out in China. And I'm going to expand that question a little bit further because geographically, you see -- you saw that China was the fast-growing country in the past quarter.

And for this -- in this quarter, you talked about like growing more users. So what geographically would you think -- like which country do you think like -- or which region do you think will be most potential to grow in the next quarter?

And my another question is related to your Duolingo score you published in this year's score count because like that was related to the efficacy. And so I'm wondering like how do you think that this kind of efficacy measurement relate to the core user performance metrics? And how -- will that be like the key motivation for you to grow the users in the next quarter?

Yes. Thank you for the question. So region-wise, the fastest-growing region as a region is Asia for us and we expect that to be true for a bit. I would expect that's going to be true. I don't know the future, but I would expect that to be true for a bit. And certainly, China is leading that, but it's not just China. It's basically all of Asia that is growing pretty fast.

In terms of the score, we're very excited about the score. Our announcement was that first of all, all of our major language courses now have the Duolingo score. And also you can share the Duolingo score on LinkedIn. We're seeing quite a good number of people sharing their score on LinkedIn. And so that really means they're using it kind of for job purposes and we have that -- the score there.

The score also is in the same range as our Duolingo English test, at least for English learners. And English learners are the one that would care more about a score like that, particularly English learners in Asia are the ones that would care more about a score. So we're very excited about that.

And it is something -- the ultimate goal for it is to become the proficiency standard for at least the major languages and certainly for English, where rather than when people ask you how much French do you know or how much English do you know at the moment where people say it's like, "Oh, I'm intermediate." We want people to say, "I am a Duolingo 60 in French or I'm a Duolingo 80 in English." That is what we want. And we think we're making pretty major progress.

In the case of the score, particularly for English, the Duolingo English test is now accepted by over 6,000 educational institutions in the world, including all, Ivy League universities and also the 99 of the top 100 universities in the United States, accept the Duolingo English test.

So we think the combination of that prestige plus the large scale that we have in the app plus doing things like sharing with -- on LinkedIn will hopefully get us to be the standard for proficiency.

Okay. And one thing is that we are not using LinkedIn that most like in China. So we are really hoping to like connect it to another social media platform [indiscernible].

We are working on that. I can't really give you details on that, but we are working on -- it's not just going to be LinkedIn that we have the score on. We're working on that.

I'm showing no further questions, and this concludes the Q&A section of the call. I would now like to turn the call back to the host for closing remarks.

Thanks, operator. I'd just like to thank everyone for joining us, and we look forward to seeing you on the next call.