Home / Transcripts / Duolingo, Inc. (DUOL) · February 26, 2026

Duolingo, Inc. (DUOL) Earnings Call Transcript

February 26, 2026

NASDAQ US Consumer Discretionary Diversified Consumer Services earnings 58 min

Earnings Call Speaker Segments

Debbie Belevan executive
#1

Hello, everybody, and welcome to Duolingo's webcast for earnings in the fourth quarter of 2025. Earlier today, once the market had closed, we put out the shareholder letter for this quarter, and you can access a copy over on our investor relations site, investors.duolingo.com. Joining us on this call are Luis von Ahn, our co-founder and chief executive officer, along with Gilian Munson, whom we're delighted to have on board as our new chief financial officer. They're going to make a few prepared comments first, and after that we'll take your questions. <<operator_script>> Also, please be aware that this call is being recorded. <<operator_script>> Prior to getting started, we want to point out that certain forward-looking statements about upcoming events and financial results will be made. Such statements carry risks and uncertainties that are outlined in the filings we make with the SEC, and they rest on assumptions that, as of today, we consider reasonable; we have no duty to revise them. In addition, we'll cover financial measures on both a GAAP and a non-GAAP basis. You can locate reconciliations of the two within our earnings materials, and we urge you to look at those while assessing how we performed. At this point, I'll hand things over to Luis.

Luis von Ahn Arellano executive
#2

Thank you all for being here. I'd like to begin by recognizing Matt Skaruppa, who as our CFO has led every single one of our earnings calls until today — that's 100% of them. Working alongside Matt has truly been among the greatest privileges of my career. Matt, I'm aware you're tuning in this evening, and I sincerely thank you for all your contributions to our mission. Matt is a tough act to follow, so I honestly can't imagine anyone better suited to step into his role than the very person who was instrumental in bringing him aboard — our longtime Board member and Audit Committee Chair, Gilian Munson. She now serves as our new CFO. She's joining us today, and I'm eagerly anticipating a close working relationship with her. Let's move on to the matters at hand. Duolingo enjoyed another robust year in 2025. We crossed 50 million daily active users for the very first time, which represents over 5x the figure we had at our 2021 IPO. On top of that, we generated over 1 billion in bookings along with more than $300 million in adjusted EBITDA. I take pride in having created a company that is both profitable at scale and delivers tremendous positive impact globally; yet what thrills me most isn't our past accomplishments — it's the road ahead. My belief has only grown stronger that rapid AI advancements will reshape how humanity learns at its core. That opens up a massive opportunity for us. No other education app comes close to our popularity worldwide, and we plan to spearhead this transformation. By embracing this pivotal moment, we're confident we can reshape learning's future while delivering outstanding returns to our shareholders over the long haul. So there's a tremendous opportunity before us. And right now, the single most critical priority is keeping users coming in. We emphasized this during our previous earnings call. Although our DAU expansion across recent years has been absolutely extraordinary, it slowed down over the course of 2025, and that pattern persisted as the new year got underway. We now anticipate roughly 20% year-over-year DAU growth across all of 2026. This strengthened my belief that bolder steps were necessary to get DAU growth accelerating again. Put another way — or rather, viewed differently, long-term worth in this enterprise comes from 2 factors: how large our active learner base is, which is essentially the size of the pie. And how well we convert that base into revenue. Think of that as the slice of the pie that actually pays. Right now, our focus is on expanding the pie itself. Therefore, as I explained in detail within our shareholder letter, we've developed a thoughtfully crafted 2026 plan centered on improved teaching and user expansion. Beyond the thousands of A/B tests we intend to conduct during 2026 to enhance what we offer, we also have bigger efforts that break down into 3 groups: better language instruction, a better experience for users on the free tier, and fueling the growth engines that come next — that is, our math, music and chess subjects. Everything here serves the aim of accelerating DAU growth. Over the medium term, we're targeting 100 million daily active users in 2028. Should we manage to double our DAUs, the reward would be considerable. A brand that's more resilient, a business with substantially greater bookings and profit, and above all, a company that reaches and teaches many more people across the globe; yet for the near term, what this means in the near term is that during this year, bookings growth will be slower and profitability lower, exactly as our guide reflects. I'd like you to understand that making this call wasn't easy for me, and I realize it might catch some investors off guard, yet it is fundamentally consistent with what I told shareholders in my very first letter, which I'll now read aloud: "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 hand things over to Gilian long-term Munson.

Gilian Munson executive
#3

Luis, thanks so much for that genuinely kind welcome. It's going to be quite an adventure to be part of this team on a full-time basis. I would be doing a disservice if I failed to express enormous gratitude to my colleague and friend, Matt Skaruppa, whom all of us will miss, and to whom I wish nothing but the best. We wrapped up the year on solid footing and finished with a strong financial position — a terrific base from which to look ahead strategically, as Luis laid out. Now to the plan and guidance going forward. Our shareholder letter contains plenty of granular specifics to assist with your 2026 modeling — probably more than what we'll provide later on. I'll restate a portion of that here and add some texture in places we truly want to make sure get emphasis. At the broadest level, our 2026 guidance calls for bookings growth of 10% to 12%, revenue growth of 15% to 18%, and an adjusted EBITDA margin of approximately 25%. For Q1, our guidance is 11% bookings growth, 25% revenue growth, and a 25.5% adjusted EBITDA margin. Below are a few specifics I'm hoping will assist with your models. On bookings: quarter-to-date, through last Friday, Q1 bookings growth was running ahead of our Q1 guidance. That said, it's crucial to keep in mind that we're entering a notably difficult compare in the first half, given dead Duo, a price increase, especially robust advertising bookings, and the Energy rollout. We are modeling a degree of improvement in bookings growth rates toward the end of the year, yet have only incorporated modest early returns on our investments at this stage, which strikes us as prudent since shifts of the kind we're making can require time to show up. On revenue: we believe the pace of year-over-year revenue growth will step down, following bookings growth rates, as the year progresses, with some stabilization during the second half. Since our Q1 guidance is 25% growth, the math suggests quarters 3 and 4 will land beneath the low end of this guidance range. On cost: when considering our COGS, the crucial point to remember is that we intend to share AI features with a substantially larger share of our user base. That is accounted for in the lower gross margins we set out in our letter. On top of that, we are investing in 2026. Consequently, we anticipate that R&D and sales and marketing spend growth will grow faster than revenue. Taken together, the path of adjusted EBITDA margins in 2026 will probably differ somewhat from prior years. For Q1, we've guided to a 25.5% adjusted EBITDA margin, and we expect that margin to fall by roughly 3 points sequentially in Q2. After that it will strengthen across the back half of the year, with Q4 standing as our highest margin quarter in 2026. One thing I want to flag, though: we do handle annual targets, and when our spending happens can move around, but we'll keep you posted as the year goes on; there's a great deal of specifics we're aware of. Our hope is that alongside the shareholder letter, this genuinely assists you in managing your estimates while we carry out our strategy. You might wonder about how we approach guidance. So I figured I should mention that. Our stance for this year is that, as in Q4, we're hoping to shrink the distance between what we guide and what actually happens, compared with what you've typically observed from Duolingo previously. Last, I want to call attention to the buyback authorization we announced today. A repurchase of as much as $400 million of our shares has been approved by our Board. In our view, this reflects solid discipline in allocating capital, and we anticipate acting on this authorization over the next year. I'm tremendously enthusiastic about becoming part of this company right now. The market opportunity in front of us is enormous in both size and scale. And I absolutely see 2026 as a foundational year for capturing that opportunity. Thanks, Luis. It's exciting to take on this role at such a pivotal time for the company; and now I'll hand things back to the operator, and we can take your questions.

Bryan Smilek analyst
#4

Luis, to kick things off, could you expand a little more on what you're observing in the market when it comes to AI, and also around innovation generally in the tech space, that makes right now the correct moment for this strategic pivot aimed at reaccelerating user growth? And then, I suppose, on a more basic level, to hit that 2028 goal of doubling users, what would you call the 1 to 2 highest-priority items on the product side that will genuinely boost that DAU growth and produce the acceleration implied in getting there?

Luis von Ahn Arellano executive
#5

Bryan, thanks for asking. All right. When it comes to AI, look, we truly are living in a remarkable moment in history. AI is going to let us teach much better. And on top of that, I think the very way that people learn is simply going to transform. The reason I say that is that I believe over the coming few years, teaching quality will reach the level of a one-on-one tutor. That's what we're going to witness. And beyond that — not only will it match a one-on-one tutor, it will also be as entertaining as a mobile game. So this amazing development is on the horizon. What we're observing is that we're simply — we can pinpoint where users, or learners, are struggling and truly focus on that, addressing every question they have, and any time confusion arises for them, we can step in and genuinely resolve it. Also — when it comes to language instruction, we can at last do a better job of practicing conversation. There's a feature we have, video call with Lily. It keeps improving at letting you practice your conversation. So we're able to see gains in how effectively people learn. That holds true for our other subjects too, definitely for math. As I mentioned in our shareholder letter, my belief is that this year, the best tutor app for math is going to be ours. That's Duolingo math. And it will just — I think the market potential here is enormous. Roughly 1 billion people around the world are learning math. So we'll be able to teach as many of them as we can, and teach them really well. So AI in general excites me a great deal. And that's why I believe the moment is now to try to draw in as many users as we possibly can. Now, as for which top priorities will enable us to reach 100 million daily active users by 2028 — it's a range of things. Broadly, the 3 buckets we're going to put investment into are: teaching languages better. Languages is our biggest subject, and for a while it will remain our biggest subject. We're going to teach languages much better. There's plenty of work to do there. A second one is improving the free user experience. The less friction that exists in the user — in the free user experience, the more users we get. We've witnessed this across the last 15 years. Any time we removed friction from the free user experience, DAUs grew, right? So that's another one. And then the third item we're going to do is simply put a lot of investment into our new growth engines. That means our new subjects. Specifically, chess has grown considerably. As I said in our shareholder letter, we have roughly 7 million daily active users in chess, and that's less than a year since we launched it. At this point, we believe we're the second largest chess thing in the world, which is, by the way, unbelievable, because if you search for chess on the App Store, we don't appear in the results. We just haven't even gotten around to doing App Store optimization, yet we already reached 7 million daily active users. So those are the things that I think will help us get there.

Bryan Smilek analyst
#6

Great. And then I suppose, to expand on that a little further, you also brought up video call. Could you discuss your thinking and the product roadmap really for Max as opposed to Super? In the past you've been quite transparent about rotating the features and the relative pricing structure. So how should we think about the Max road map overall as you begin to ship those video calls to Super users too?

Luis von Ahn Arellano executive
#7

Yes. A few years back, when we launched Max, we rolled out some AI features. We were really thrilled about every one of those AI features since they genuinely can help us teach more effectively. Back then, the cost of AI inference was far greater than what it is today. So we made the choice to place our AI features behind a substantially more expensive subscription tier, which is Duolingo Max. That is what we've done. But we stated at that time that as AI costs declined, we would try out putting features into various tiers, because in the end, our aim is to teach the largest number of people we can. Right now, the expense of a video call has dropped very considerably over the past couple of years, ever since we introduced video call. And we feel fairly at ease about being able to include it within Super Duolingo; now, like everything we do, this will be an A/B test. We're going to A/B test what occurs if we place some video call in Super Duolingo. I don't know what that A/B test will produce. It could be that this boosts revenue because more people subscribe to Super Duolingo, or because retention for Super Duolingo is much higher since video call is available there, or it could be that it lowers bookings because we aren't earning as much from selling Max. I have no idea what it will do. But regardless of what it does, we're going to take the suitable options. So for instance, if the drop in bookings is too large, we might do something like metering video cost. So for instance, on Super Duolingo, you would only get access to 1 video call per day, whereas on Max, you get unlimited. Now I'm saying all of this well ahead of time because I truly don't know what will win in these A/B tests; but what I will tell you is that, certainly, we will put some form of video call in our Super tier because we simply think it will help us teach much better to a considerably larger number of people. There are roughly 10x as many subscribers on Super as there 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. My take is that eventually, it won't be logical for a plan as pricey as Max to include something like video calling. So right now we're making this move from a place of strength, since our financial position is really solid, and we'll simply go ahead with it — rather than being forced into it defensively 3 years down the road because some rival suddenly appeared. Honestly, at this point, competition isn't something that concerns us much at all. So we're acting ahead of the curve.

Gilian Munson executive
#8

And Bryan, the way we set up this guidance was to give the company room to pursue all of that, to experiment with it, and to do what's right for our customers. That is exactly how the guidance, and how we're thinking about the financials this year, were designed.

Ralph Schackart analyst
#9

Luis, on this final call, you mentioned reprioritizing for 2026, and my understanding was that some effect on the 2026 financials was expected—though perhaps not as big as what was guided today. That's at least how I read it. I might be mistaken, of course. But did anything shift after last quarter regarding your thinking on 2026? Is this user growth bigger than what you'd initially expected? Then maybe a question for Gilian regarding that 20% DAU growth. Once those friction points—or monetization friction points—are taken away, should we anticipate some reacceleration in the DAU metrics? And you kind of hinted that bookings are tracking above. Maybe just some early observations on how user growth is responding as you remove some of those friction points?

Luis von Ahn Arellano executive
#10

Thanks, Ralph, for your question. So yes, on the previous call, like you mentioned, we stated that we would be focusing on user growth. For reasons similar to now: first, we had witnessed DAU growth decelerating across 2025, along with some slowing of user growth; and beyond that, we believe this is simply a rare chance where rather than slowing user growth, we ought to be speeding it up, since we should—just seize this opportunity. That's what we communicated during the prior earnings call. Two things have happened from then until now. First, we continued to observe that deceleration trend in users. And at this point, we anticipate roughly 20% year-over-year growth across 2026. And that—the second thing is that my confidence in AI only grew further. That made it so I hold even stronger conviction that we must direct really the vast majority of our efforts toward DAU growth. And when we put together the plan for 2026, this is what resulted. I mean the guidance we're providing; and regarding the 20%, I suppose, Gilian, you can go ahead and answer.

Gilian Munson executive
#11

Right. So like Luis mentioned, we genuinely anticipate that 20% across the year. When it comes to Q1, based on what we've observed up to now, it's running above the guidance. And we believe that demonstrates to everybody that the business is quite healthy. However, there are some genuinely, genuinely difficult comps on the horizon; therefore the overall guide takes that into account. I believe it's genuinely crucial to frame what Luis is referring to, which is the confidence we've built in pursuing the larger price. And so when you consider that 20% alongside the 100 million target, that would suggest an acceleration toward significantly higher growth rates over the next couple of years. And if you sort of work that through regarding what the financials might look like, that is, as Luis stated, a considerably larger business. So when we examine the business, we tell ourselves, alright, we might have expanded the business, and we'd likely be sitting at roughly a mid-teens CAGR across the next couple of years. If we were able to hold our margins at their current level, we'd probably be looking at a $1.5 billion business, $400-plus million in adjusted EBITDA. But we, at the company, are genuinely driven to pursue the larger prize, which aligns far more closely with what Luis and Severin had in mind when they started the company; so then if you sort of work through the DAU, this was all rough estimation. But if we're able to reach that DAU and even sensible monetization assumptions compared to where we stand today, and we're confident we can scale our expenses over time, you're looking at a business in a couple of years that could be $2.5 billion with over $700 million in adjusted EBITDA; and I think throughout this entire company, we're genuinely driven to pursue the larger price. And so that's what you're observing and sort of how we're approaching it. Regarding when you'll see it reflected in the model, we've genuinely modeled it late in the year; so it's going to require a little while. What we're working on is going to require a little while to come through. But we do have a little bit at the end of the year; we believe it's prudent to keep that small for now until we sort of get going. You'll begin to see more and more as we move into '27 and '28.

Mark Mahaney analyst
#12

I'd like to ask about growth and also about product. On growth, you listed several reasons behind the slowdown — the law of large numbers, and perhaps too much emphasis on monetization. Two additional possibilities exist as well. So please speak to those: market saturation and how intense competition is. Maybe there are — so just — do you have fresh data points that remind you, or make you even more aware, of the long-term opportunity to grow here? Then on products, the video call is a fantastic product; the other items you're referring to, the speaking adventures, is that coming along? How soon can we expect it? And reaching that more advanced content with all of your 9 largest languages up to that 129 score. Does that happen now? Or does it require building over multiple years? How long until those 2 products arrive that will — that we might be able to see the products and then the effect they would have?

Luis von Ahn Arellano executive
#13

Thanks, Mark. Okay; so your first question concerns market saturation and competition. Neither of those worries us. When it comes to competition, if you examine language learning apps, roughly 85% of the daily active users of language learning apps worldwide are ours. And that figure has stayed fairly level for some time. Basically, that's just not something we're especially worried about. As for market saturation, we're — for instance, we look at things such as the share of daily active users in a particular country relative to that country's Internet base size. So for instance, within the United States, 2% of all Internet users on any given day use Duolingo. So that's 2% penetration in the U.S. for daily users; in the U.K., it's 3%; in Germany, it's 4%. And Germany isn't the top country we have. We actually believe we can reach far higher than that. But even if we only suppose that every country reached 2%, which is the U.S. penetration, we would more than double our daily active users; so I simply don't think we're anywhere near saturation. And broadly, we aren't observing anything different. For instance, we survey our churn users. And what our churn users say in those surveys hasn't really shifted in years. And the most common response, incidentally, when people quit using Duolingo regarding where they go is that they quit learning. And that has held true for some time. So that's just not something we're especially worried about; regarding the features, we noted in the shareholder letter, we noted speaking adventures. We're very thrilled about that. So what does this feature look like? Essentially, it's a small game in which you're required to carry out a task. Say, for instance, someone instructs you to purchase a sandwich, hand it over to a particular individual, and then request payment from them, or some similar assignment you must complete. Throughout it, you perform plenty of actions using only written words, yet occasionally within it, you're required to talk in order to obtain whatever it is you're after. That's going to launch — it will be available to everybody, both free and paid users; it's going to scale up. I mean, testing is already underway on our end. It will scale up at some point around the middle of the year. So you'll likely notice it inside your app at some point from now until the midpoint of the year; we're really thrilled about it; regarding content that's more advanced, that's arriving fairly soon. Within 1 or 2 months, you'll notice far more advanced content. So essentially, every one of our courses that instructs in the top 9 languages we offer, which, incidentally, reach over 90% of our daily active users, will include content reaching up to Duolingo score 129, a score that matters especially since it's the one required to land a knowledge job in that language. So we're really thrilled about it. And it's arriving right now. It's worth emphasizing that this content will be released out there, and we're really pleased about it. Yet that's only the start. The moment it's released out there, we can begin enhancing this content in the very same manner we enhance everything. And so across the rest of the year, you'll simply notice increasingly improved versions of this more advanced content, and we're quite thrilled about it.

Justin Patterson analyst
#14

Luis, I was hoping you could expand a bit on the social strategy specifically. The level of virality we used to see coming out of TikTok doesn't seem to be there anymore. So I'd love to hear how you're tweaking that approach and viewing it simply as a way to boost daily usage. Then regarding video call, I understand you've introduced quite a few enhancements meant to get more words spoken and make the experience simpler. So I'd love to hear how that's performed relative to what you expected, and how you see shifting it into the Super tier driving additional engagement and ultimately supporting that bookings reacceleration.

Luis von Ahn Arellano executive
#15

Thanks, Justin. As far as our social marketing approach goes, we've got a fantastic marketing team — honestly, I'd say the best anywhere. Our videos do go viral and rack up millions of views. Still, I have to mention the comparison point, especially twelve months back — twelve months back, there were plenty of weeks when we held the single most watched video across all of TikTok. That was — that's amazing. So even though our videos continue to go viral, we're no longer in a spot where we hold the most watched video across all of TikTok. There are several reasons behind that. One is that algorithms have shifted enough that achieving it consistently has become tougher. So that's one factor. Broadly, we'll keep pursuing the approach of trying to achieve as much virality as we can. And that will — I'm fairly pleased with what we've seen so far. But you're correct that versus twelve months back, it isn't quite the same. On video call enhancements, I'm quite pleased with what we've gotten from them. Essentially, if you look at the chart of words spoken per user on video call, it's simply a monotonically rising chart as time goes on. It keeps improving and improving and improving; and essentially, we've gotten much better at getting Lily, the one you're speaking with, to draw out more conversation. At present, we're testing something new — we're beginning to tell you, hey, in your next response, pose a question, or in your next response, include the word because or something like that. And that genuinely gets people going more and more. These enhancements have me feeling quite pleased, and I genuinely believe that expanding access — letting additional users try it — will essentially turn more of them into Duolingo advocates. The reason is that we observe individuals who use video call sharing on the internet remarks along the lines of, "Wow, I doubted it was possible to pick up a language through Duolingo, yet after simply using Duolingo along with video call, suddenly, without expecting it, I can now hold a conversation." That leaves us feeling great.

Andrew Boone analyst
#16

I'd like to revisit the free user experience and how we can make it better. The real question here is, how does that eventually turn into a way to make money down the road, correct? From what I'm hearing, one takeaway was that you basically squeezed too much money out of your user base; so Luis, as we consider moving through this shift, what's waiting on the other side when it comes to what you've learned, as you return and then begin charging users at a later point?

Luis von Ahn Arellano executive
#17

Andrew, that's a great question. Here's the thing — we ended up in a curious spot where Duolingo simultaneously isn't monetizing enough and is monetizing too much. It isn't monetizing enough in the sense that, well, just roughly 10% of our monthly active users give us money. We firmly believe we're capable of far better than that. When you examine comps, when you examine — when you examine companies such as Spotify, roughly half their users pay them. And with many of these freemium operations, we genuinely believe we can reach well above 10%. But the method we used to grow monetization was that we discovered the fastest route to growing monetization essentially involved inserting friction. The more friction we inserted, the more we persuaded people to subscribe. That's fine, yet I believe we pushed it to a place where it genuinely started conflicting with DAU growth. There are alternative routes to persuade people to subscribe. For instance, you could introduce additional features, or you could have features that aren't precisely friction, but rather items such as selling customizations for avatars and similar things, which have performed very well for games, and we're sort of half a game. So what we're doing right now is, over the coming year, we're spending this year also discovering methods to monetize that persuade more people to subscribe. That's correct, but those methods might not rely on inserting friction. And we're quite confident we're able to do that. It's simply slower, since monetizing through inserting friction — like if you informed me next week that I had to generate, say, $50,000 more each day, it was actually pretty simple to accomplish. You merely double the ad load or whatever it happened to be; but these alternative monetization methods are simply going to require somewhat more time, though we're extremely confident we'll find some, because just 10% of our active users give us money to subscribe.

Andrew Boone analyst
#18

Next, I’d like to ask about the chess disclosure. This really concerns people using several apps simultaneously. So if I consider those 7 million users, and I consider how core Duolingo has trended through 2025, how should I think about multiple users on multiple apps? Do I simply take 53 million and subtract 7 million in order to figure out where you stand? What is the correct way to approach that?

Luis von Ahn Arellano executive
#19

That’s another excellent question. Yes. We stated that chess reached roughly 7 million daily active users in under a year; we’re extremely proud of that. It’s amazing growth. Now, for most of those users, whenever we launch a new subject on Duolingo, reaching our existing users is much simpler than reaching people outside Duolingo, since we can communicate directly with our own users. So for most of those users — we aren’t stating the precise figure — most of them were already Duolingo users and now use both languages and chess. That’s essentially the usual pattern taking place. So the way to view it isn’t that languages have stopped growing. Actually, languages are expanding at roughly the same rate as Duolingo, because it remains the biggest course. It’s simply that we’re gaining some new users for chess, yet many of them are merely doubling up from languages and chess. What we’re observing is that gradually, more people are discovering that Duolingo offers chess, and then chess begins attracting its own user base. I believe that’s what will occur over time. And accomplishing that just requires a bit of time, but that’s what’s happening.

Alex Sklar analyst
#20

Luis or Gilian, I'm not sure which of you wants to handle this one. Regarding the bookings outlook, considering a few of the uncertainties you mentioned in response to Bryan's earlier question about how video call tests will unfold on Super: what does the bookings outlook assume about Super's conversion rate compared with what you've seen before? And Luis, how much time or data do you figure you'll need before you can tell whether that decision turned out well, based on prior testing?

Gilian Munson executive
#21

Looking at bookings, the key point at this moment is that if you begin with Q4 bookings growth of roughly 24%, then move to the guide, we consider about half of it to be the continued user deceleration Luis referred to. The remaining half comes from us monetizing more thoughtfully and considering removing certain friction for our users. That's our approach to framing it; as for precisely how the monetization will ultimately unfold, as you're aware, we run A/B tests constantly here. So, as Luis described, there's plenty to discover across this year. What the guidance is genuinely meant to accomplish is to provide ourselves room, from a place of real financial strength, to carry out that work, since we believe it opens the chance to pursue the higher price.

Luis von Ahn Arellano executive
#22

Yeah; and as for the amount of time before we can tell that this is succeeding, it's going to require a bit of time. I mean, there are a lot of product changes we have to make, and so on, and every one of those takes some time, and after that we also have to observe how users respond, and so forth; so it's going to require a bit of time. Like Gilian mentioned, we're adding something modest near the close of the year to check that a few things are functioning there.

Alex Sklar analyst
#23

Then perhaps one more brief question. Has any consideration been given to pricing adjustments across the various packages in the guidance?

Luis von Ahn Arellano executive
#24

Well, during this year, our plan is to run experiments, much as we do annually, though likely to a greater degree when it comes to prices. That involves raising prices and cutting them; A/B testing happens constantly on our end. So that's what we'll be doing; however, we aren't — that isn't something we truly include — what we include in the outlook are the outcomes from all the experiments taken together. Still, whether I could state as a certainty that Super's price will go up owing to video call, I can't say. The A/B test results remain unknown to me.

Wyatt Swanson analyst
#25

Thanks for asking that. I just have one thing on my mind, which is the medium-term objective of getting to those 100 million DAUs. Does that presume that marketing remains at a high level and that monetization is kept on the back burner throughout the whole period? I believe you've sort of addressed this, but I'm just sort of reasoning it through, since as you've gradually increased monetization over the last year, you've begun to observe DAUs starting to decelerate. So how do you think about that trade-off across the medium term? Should we expect monetization to gradually begin increasing through these alternative formats that you're sort of testing?

Luis von Ahn Arellano executive
#26

Yes. I mean, I can respond, and I can also let Gilian take a crack at it; but broadly speaking, we — our marketing remains incredibly efficient. I mean, it's truly still a tiny portion of our revenue that we put into marketing. I expect that each year, the overall amount of marketing dollars we spend will keep rising; however, I'm hoping that beyond this year, we are going to achieve some efficiencies; it's tough to say precisely what will occur beyond this year. It's not as if we're guiding, but I expect that we're going to achieve some efficiencies beyond this year; and then regarding monetization, as Gilian said earlier, in almost any manner you examine it, with 100 million daily active users, this is simply a far bigger business. I do believe — again, tough to express in numbers precisely to provide you a model or anything for beyond 2026, but I do believe that you're going to see other things truly begin accelerating. I'm extremely enthusiastic about direct ad sales; we're seeing just strong uptake on that. And I'm also extremely enthusiastic about in-app purchases, especially with our avatar. There's a great deal of love for our avatars and for our playful brand, and we think we can accomplish quite a lot there.

Gilian Munson executive
#27

Absolutely. When you look at the margins in this business, what makes me really enthusiastic about our current position is that we've demonstrated the ability to achieve enormous scale at the size we are now, and we believe there's room to grow much larger. Nothing we observe in the operations suggests we can't continue expanding from here. At present, we're putting money in and deliberately pulling back in certain areas to enable that. But honestly, we don't believe doubling our headcount or spending twice as much on marketing is necessary. This business, in our view, can grow very elegantly. One of the best aspects of our model is that we can pursue these initiatives from a place of strength and profitability; we'll keep producing significant EBITDA and free cash flow even while making investments, and we don't see any obstacle to scaling as we return to monetization in a somewhat more intelligent manner going forward.

Wyatt Swanson analyst
#28

Understood. That makes sense. Could you give us some sense of what the advertising side of things is going to look like down the road? You mentioned that perhaps doubling the number of ads is creating extra friction and possibly driving users away, so it sounded almost as if advertising were being pulled from the platform and wouldn't remain such a major monetization channel.

Luis von Ahn Arellano executive
#29

What we probably won't do is raise the number of ads. Right now, you encounter an ad when a lesson finishes. That probably won't increase, definitely not during this year, since it truly runs against everything I've been telling you. That said, we're getting much better at showing higher-quality ads. In the past, we simply hadn't put much work into advertising. For instance, we rely on those network ads, which are sort of the lowest common denominator; these days we're doing direct deals. This accomplishes two things for our business. First, it greatly improves the ad's quality. So instead of ads for games you've never heard of, you'll see ads from companies like Disney, far better ads; and second, the revenue we earn per ad goes up a lot when we operate that way. So that's roughly how things will go, I believe. We're testing some things that are genuinely exciting, though I'm not sure we'll get them to work. Honestly, my longtime dream has been to show you ads in the tongue you're studying. And we're attempting that. So if Spanish is what you're learning, a portion of the ad will be in Spanish. What I really like about this is that users actually want it. And that ultimately benefits the advertiser too, since the user is genuinely paying close attention. So I hope we can give that a try. Which of these ideas will pan out, I can't say. Broadly, though, my view is that the number of ads won't rise, yet we'll earn much more per ad.

Nathaniel Feather analyst
#30

I've got two things to ask. To start, the market has been quite worried about how much AI translation tools are getting better, plus the broader surge of chat bots built for every purpose. What I'd like to know is what makes you confident that the slowdown in users we witnessed during 2025 wasn't caused by those things? Then next, regarding the adjustments you've implemented up to now to make the user experience better, does anything in the data give you confidence that those adjustments will steady user growth across the year? Are you seeing that happen already? Or when do you think that steady point would arrive?

Luis von Ahn Arellano executive
#31

Thanks for asking. So AI translation — folks bring it up, yet internally it's not a source of worry for us. Among the big languages, AI translation has basically been flawless for over a decade or so. Take Spanish and English — it's basically been flawless. There are 2 primary reasons our users come to Duolingo: first, as a hobby, they genuinely wish to learn; and for those folks, AI translation existing or not makes no difference, since it's a hobby to them. Then there's another set of people who are learning English. They wish to learn English, and they genuinely wish to learn English. We also know that we inquire with users about why they quit Duolingo. We run these churn user surveys. This topic doesn't show up. Like I mentioned, people mostly quit Duolingo because they got busy. We understand what that signifies; it basically signifies they're spending more time on social media, that's essentially what it signifies. So we're simply not — that's not a thing we're especially worried about; as for data regarding — do we see any data convincing us that user growth is speeding up, and so on. Up to now, we've made a few changes to improve the app — to improve the free experience. For instance, we made a feature available that had been in Duolingo Max, called Explain My Answer. It isn't free for everyone. So we've made changes along those lines. And we're noticing a fair amount of uptick on the features, and we're noticing that users love it; but we're — we haven't yet noticed an acceleration in DAUs that makes us think, "Oh my god, we're done." To be fair, though, we didn't anticipate this. I believe that's why we're telling you to expect some modest results near the end of the year.

Ryan MacDonald analyst
#32

Luis, I'd like to begin with Duolingo Max. In Q4, if I recall correctly, you hit that big group — or maybe you were up against the first sizable wave of Max subscribers. So I'm wondering what you're observing on retention and renewals, and how much that's shaping the plan to put video call into Super. Also, once that shift occurs, Gilian, when do we expect to see the effect of subscribers stepping down from Max to Super, given video call is where things are headed? And what are you assuming about that in the guidance?

Luis von Ahn Arellano executive
#33

Sure. Regarding Max, nothing about the Max figures is driving any decision we're making on video call. The Max figures are honestly pretty strong. We're pleased with them; what we're doing with Max is what I'd call an offensive play. My view is simply that video call's cost has fallen enough that it no longer makes sense to keep it only inside Max when we could put it in far cheaper tiers. We aren't yet at a point where we can give it away at no charge, though eventually we might get there. And I do want to give it away at no charge, because honestly I think our interests are best served by having the biggest user base possible that we can find some way to monetize, instead of reserving this for a very small slice of users; that's just how we think about it. So the whole reason for trying to move video call into Super comes from our belief that this is what's right for our users, not from any Max figures; we're quite pleased with the Max figures, in fact; and Gilian, I'm not sure whether you wanted to take that second part. You're — we can't hear you. You're muted, or I can't hear you. Well, apologies, I've sort of lost track of your question since I was going to hand it to Gilian. I've forgotten it. What was it you asked? I'll try once more to respond to it.

Ryan MacDonald analyst
#34

Yes. When you're considering the fiscal '26 guidance, what sort of assumptions, if any, are you incorporating regarding the timing of video call's rollout to Super — and maybe, I suppose, what you might describe as ARPU trend shifts, perhaps from Max downsells to Super because of that?

Luis von Ahn Arellano executive
#35

Yes. You know, I can probably tell you what — I'm aware of what Gilian would have said here. Look, we don't truly know ourselves precisely what will occur with this A/B test, where we begin offering video call to Super. We don't know that ourselves; but we are going to A/B test it. Everything — in our guidance we placed some bounds, what we believe that will actually — this is what we believe overall, and there are some bounds; we'll remain inside those bounds. And we have various methods for remaining inside those bounds; so for instance, if we observe that offering video call in Super genuinely wrecks our Max business or something, which is really awful, then we'll begin doing something such as metering video call in Super, like you only receive access to 1 per day versus in Max unlimited, something like that. Again, I'm discussing things in the future because — so I simply don't know precisely what will occur, but we are going to remain inside 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.

Ryan MacDonald analyst
#36

Ryan, not quite — okay. Luis, here's a quick one for you. You mentioned the math and the thinking behind it, that this represents a chance to build what could be one of the world's best math tutoring apps. As it comes together, would you explore new channels for getting it out there? In the past, I think Duolingo for schools served as a method of opening up that population on the language side. Could that be a workable channel when you consider where DAU growth might come from with something like math?

Luis von Ahn Arellano executive
#37

Yes. Math is a fascinating subject. On one hand, they are really — I can't name another subject besides languages with a market that large. When it comes to math, around a billion individuals are studying it, which is impressive. But there's a key difference from languages: most of those billion people aren't eager to study math, whereas those learning languages truly have the desire. Also, the Duolingo users are largely in educational settings — specifically, K-12 schools make up the majority. That's their primary location. So that's definitely the area we'll be focusing on next. There are numerous approaches to envision how we might achieve this. Our initial move is just to create the product itself. We're almost finished with that. If you check the internal builds of our math course, they're remarkably strong. After we complete that, the major challenge becomes how to bring it to market. At this point, what we're thinking is that the simplest entry point—and it might not be where we end up—would be to treat this as something that adds to what kids do in school. So instead of trying to strike arrangements with school systems and the like, picture it more as something going head-to-head with Kumon. Our sense is that this is the smarter way to begin, given how many folks out there are already getting some form of outside assistance and so on. One thing that appeals to us here — and keep in mind, this won't show up during the current year — is that over the longer haul, parents, as opposed to the children, show a strong willingness to spend on math. That likely opens up an opportunity for us to act on.

Gilian Munson executive
#38

Can you pick me up at this point, Luis?

Luis von Ahn Arellano executive
#39

That's right.

Gilian Munson executive
#40

All right, that's fine.

Luis von Ahn Arellano executive
#41

Gilian has returned. I'm happy about that. If she hadn't, people would have begun posing inquiries regarding abbreviations I'm unfamiliar with, and we'd be in a difficult spot.

Gilian Munson executive
#42

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

Shweta Khajuria analyst
#43

Before you unmuted, I had planned to hit Gilian with three pointed questions, purely as a mic check. Actually, though, it comes down to two. The first concerns retention; the second concerns AI. On retention rates: given that you're putting money into math, music, chess, and there may be certain advantages across platforms, how should we view churn and retention down the road as product investment continues? From what I gather, things have stayed more or less steady. Still, is there any basis for expecting a notable uptick, and what would that mean for your platform? Then, on AI: you spoke to competition, noting you hold a market lead over everybody else at present. Yet what AI might do is speed up how quickly brand-new rivals — ones that don't exist yet — could appear. So could you discuss the risk that a competitor pops up because AI lowers the difficulty of building particular products?

Luis von Ahn Arellano executive
#44

Sure, yes. On retention, I'll hand it to Gilian to elaborate. Broadly, though, our retention — particularly among payers — is quite steady. It's extremely healthy and extremely steady. It could possibly rise. I doubt we're building that into our models. I'll check with Gilian, since I'm honestly unsure whether we model it or not, but it could rise at some point. The reason: one thing we do know is that people who use several subjects retain better. And it's also true that as time passes, a growing share of users study multiple subjects, simply because we keep adding subjects. So it could go up — though I don't think — up to now, this is — you really shouldn't take my word for it, since it's stayed fairly flat. Gilian, I don't know if you'd like to chime in, and after that I can address new competitors and AI.

Gilian Munson executive
#45

When it comes to retention, it's been pretty level for quite a while; we examine both the aggregate numbers and the cohort data. Our view is that it could remain right where it is. That said, you're raising a point that's genuinely, genuinely important. Everything we're doing to improve the product — so that word of mouth and top of the funnel and so on keep the pool larger — benefits existing customers as well, correct? They receive those features too; so I'd say it's possible, but we haven't modeled anything substantial on that front.

Luis von Ahn Arellano executive
#46

Yeah, and regarding your question about whether a new rival could emerge since AI makes it possible to create something more quickly — sure, it's accurate that AI lets you build things faster, and I'd expect that to become even more the case as time goes on. We're operating under that assumption. The encouraging part is that this same advantage applies to our team too, so we're able to build things faster as well, which is great. Another point worth making: I believe the market slightly misjudges how simple it is to replicate a highly advanced application such as Duolingo. Truthfully, if someone were to simply jot down the specification for Duolingo — and I don't mean coding it in some programming language, I mean just writing it out in plain English — that alone would take me years, since there are 1 million edge cases in what we do, and we're quite clever about how we handle them for the user, whether it's boosting engagement or improving instruction, and so on. So I'm not especially concerned about someone merely hitting a button and suddenly an app shows up that's simply superior to Duolingo. Beyond that, it's quite difficult. The explanation Duolingo thus far hasn't truly faced competition on a big scale — once more, we hold something like 85% of the daily active users among language learning apps globally — is that going up against our free offering is quite difficult, in the sense that our distribution is enormous. And when a newcomer arrives, they must determine how to expand. Typically they turn to performance marketing to attempt rapid growth, yet since our free offering is so strong, they generally can't get their performance marketing to succeed if they also offer something free. So next, they place a paywall on it. And that severely limits them. So it's simply been quite tough for newcomers. I mean, we observe — well, the truth is that throughout Duolingo's existence, we've likely witnessed, I don't know, 400 apps launch and attempt to go up against us. I'm inventing that figure — it's hundreds. And we simply don't watch them expand, again, because our distribution is very large.

Eric Sheridan analyst
#47

Perhaps I can piggyback on what Shweta asked and frame it somewhat differently. Regarding the AI landscape, how broadly do you believe you must lead in AI learning within your application's 4 walls and your experience, versus potentially partnering with LLM agents and serving as an application layer atop their platform, should greater consumer behavior shift that way from an aggregation perspective? I would really like to hear your philosophical take on this. And perhaps the second question: you were likely among the foremost examples of a mobile-first company during Web 2.0. How ought we to consider the timeline and the investment required to reposition this company as AI native in its tech stack across the next couple of years?

Luis von Ahn Arellano executive
#48

Yes. So a good question about whether it makes sense for us to keep going with our own app versus something such as an LLM. I will tell you what we do — just historically, a number of new platforms have emerged since Duolingo launched. For instance, smart watches appeared; smart speakers also appeared, like Alexa. In every instance, many people asked us questions. And certainly, the companies building these things told us this is the next big thing, everyone will be in their watch. The watch is where you'll find us. In the past, we made a misstep when we went ahead and created an app for the watch, or something along those lines, quite early on. Then, only afterward, we came to understand that watch apps aren't really a category that exists, apart from fitness-related ones. So here's how we look at it: should it happen that how consumers behave shifts significantly, to the point where nobody keeps a standalone app and instead everyone gravitates to that other place, then obviously we would certainly take action in response. That said, one thing I'd point out is that our particular app — and of course there are plenty of others out there — is something you can think about in much the same way you'd think about a game. What I'm saying is that not being text-based is genuinely important, since it turns out folks truly enjoy activities such as playing Candy Crush or using Duolingo, partly thanks to the animations and so on. So if the question is whether we'll exist as something within a chatbot, my honest view is that inside a text-based chatbot, you simply can't achieve the level of engagement we're able to achieve. That said, I don't know how things will unfold down the road; but right now, we're fairly confident that the correct place to put our money is into continuing to build a gorgeous, highly visual experience, because that's precisely what holds people's attention. When it comes to being AI native — look, at the end of the day, we're a technology company, and we've been working hard to bring AI and tech on board as quickly as we can. Honestly, we've likely landed ourselves in hot water for how fast we've embraced AI — take, for instance, those foolish memos I penned once; so that's part of it — and beyond that, I believe we're simply — what you'll keep seeing from us is that we'll always move to whatever comes next when it comes to things such as AI.

Luis von Ahn Arellano executive
#49

All right, thank you all. Those were some great questions. I’d like to add that 2026 has us really enthusiastic. It’s likely not quite what investors were expecting. But if you view Duolingo as a company you’d hold for the long term — which clearly I do, since that’s basically what I’m after — then this is genuinely the right investment to make. The second point I’d make is that Gilian began her role 3.5 days ago, so it’s amazing how much she already knows about our finances, which is already more than I do. So thank you, Gilian, for that. And thank you, everyone.

Gilian Munson executive
#50

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

Read the full transcript via the API

You're viewing the first half of this call. Get the complete Duolingo, Inc. transcript - plus 132,000+ transcripts from 7,000+ companies, speaker segments and full-text search - through the EarningsAPI REST API or hosted MCP server.

Get an API key View API docs →

For developers and AI pipelines

Programmatic access to Duolingo, Inc. earnings transcripts and 132,000+ others is available through the EarningsAPI REST API and the hosted MCP server. Quarterly plans from $162 - full transcripts, speaker segments, full-text search, and the /api/v1/transcripts/recent polling endpoint for ETL pipelines.