Duolingo Fixed the Funnel and Killed Max: Users Accelerated to 29%, the Stock Re-Rated 26%, and the Constraint Moved to Monetization

Results vs. Consensus

Duolingo reported after the close on November 3. For three quarters the useful question has been whether management is delivering what it promised when it reset the equity in February, and for three quarters the answer has been yes on the financial lines. This quarter it also delivered on the one operating line that the entire strategy was built to move, and it did so by a wide enough margin that the market repriced the stock 26% in a session.

Year-over-Year Comparison

Sequential Comparison

Revenue: the acceleration is in bookings, not in the reported line

Revenue grew 14% to $309.0M and beat consensus by 1.9%, but the reported growth rate decelerated four points from Q2's 18%. That is mechanical rather than informative. Subscription revenue is recognised ratably across a term that is mostly twelve months, so this quarter's revenue reflects bookings struck over the prior year, including Q2's 8%, the slowest bookings growth of 2026. The forward-looking line is bookings, and bookings did the opposite: 12% growth against 8% in Q2, a four-point sequential acceleration, with subscription bookings up six points to 16%. For the first time in this investment year, the leading indicator and the lagging indicator are pointing in opposite directions for a good reason.

The composition is less comfortable. Subscription revenue grew 18% and carried the entire result; every other line of any size was flat or down. Subscription is now 87.4% of revenue against 84.5% a year ago, so a strategy whose central premise is that a larger free base creates non-subscription monetization has, after three quarters, produced a business that is more subscription-dependent than when it started.

Assessment: The bookings reacceleration is the single most important number in the release and it validates the February plan on its own terms. The revenue deceleration is noise and will reverse as the faster bookings cohorts season. The non-subscription shrinkage is not noise, and it is the reason the rating is now a thinner call than the headline suggests.

Margins: a fourth consecutive beat against a guide built to fall

Gross margin has now beaten its guide in four consecutive quarters, and the size of the misses in the company's own forecasting is becoming the story. In February management framed a path to roughly 69% by the fourth quarter; the year will instead finish above 72%, with Q4 guided to 72.0%. At 72.9% the quarter sits 40 basis points above the prior year and just below the 73.0% printed in the first quarter, so this is a sustained level rather than a new high. The CFO was direct about the cause when asked whether the quarter's gross margin reflected a payments-mix shift.

"No. The main delta in the gross margin was better revenue and less AI costs than we expected going through." — Gillian Munson, Chief Financial Officer

Adjusted EBITDA margin fell 400 basis points year over year to 25.5% and 40 basis points sequentially, which is the investment year behaving as designed, and it still landed 30 basis points above guidance. Below Adjusted EBITDA the picture is uglier and mostly explicable: research and development rose 30% year over year and 17% sequentially, and GAAP operating margin of 7.8% is the lowest of the five quarters we have covered, against 12.9%, 15.4%, 15.3% and 11.4% in the four before it.

Assessment: This year's three gross-margin beats, each credited at least in part to lower AI costs, are no longer a pleasant surprise but a structural feature, and that feature is the mechanism that funds everything else management wants to do. The reported operating margin is the price of the strategy and should be read as temporary only to the extent the bonus is non-recurring at this scale, which management has implied but not stated.

EPS: a reported miss that is not an operating miss

Diluted GAAP EPS of $0.54 came in three cents below the $0.57 consensus. Working back up the statement, the quarter beat on revenue, beat on gross margin, beat on Adjusted EBITDA, and carried a $12.4M interest-income line that grew 5% on a cash balance that reached $1.4 billion. The shortfall sits entirely in operating expense, and the bonus accrual alone is several times its size. Adding the accrual back after tax at the guided 24% midpoint rate is about $0.17 per share, so a Street model that had excluded the bonus, as the company's own guidance did, would have modelled EPS nearer $0.71 on these revenue and gross-margin figures.

The year-over-year EPS comparison is meaningless and should be ignored. Last year's $5.95 included the valuation-allowance release; this year's tax provision of $8.4M against a 23% to 25% full-year assumption is the normalised run rate.

Assessment: Treat the EPS miss as a disclosure artefact rather than a result. The number that matters for a company at this stage is bookings and the cash it converts to, and both were better than guided. That the market moved up 26% on a GAAP EPS miss is evidence it read the release the same way.

Users, and the constraint that just moved

In August we wrote that the funnel had widened a quarter early and that the next thing to grade was the fourth quarter, where management's own internal bonus target for daily-user growth was 25%. The third quarter came in at 29%, management expects the fourth to be similar, and the company has begun accruing the bonus because, in the CFO's words, it is on track to hit the targets. That is a decisive answer to the question that has gated this rating since initiation.

The breadth matters more than the rate. Daily-user growth accelerated in 8 of the 10 largest countries, and subscription bookings grew in double digits in a majority of the regions served. Monthly users grew 13% against 10% last quarter and 5.8% two quarters ago, so the top of the funnel is not merely no longer flat, it is accelerating in its own right. The daily-to-monthly engagement ratio reached 42.4%, up 510 basis points year over year, and has now risen in every quarter we have tracked.

Put the trajectory against the target management set in February. Reaching 100 million daily users by the end of 2028 from 64.9 million today requires roughly 21% annualised growth over nine quarters. The company is running at 29% and guiding the next quarter to the same. On the February framing that goal was a stretch; on this quarter's run rate it is conservative, which is presumably why the CEO has started talking about one billion learners instead.

One user metric moved the wrong way, and it is now the centre of the investment case. Paid subscribers grew 14% to 13.1M, decelerating in every quarter we have covered, from 34% a year ago to 28%, 21%, 17% and now 14%. Expressed against the base it is meant to monetize, paid subscribers fell from 22.8% of daily users a year ago to 20.2% today. The free base is compounding faster than the paying base, by a widening margin.

Segment Performance

Duolingo reports as a single segment. Three decompositions carry information this quarter: the revenues-by-product-type appendix, the bookings split between subscription and everything else, and the subject-level detail that management gave verbally after two quarters of silence.

This is the second consecutive quarter in which this table has contradicted the strategy's stated logic, and the contradiction has deepened. The premise of the investment year is that a deliberately larger and less friction-laden free base creates monetization capacity outside subscriptions. The free base grew 29%. Non-subscription revenue fell 8% in aggregate, with advertising now in outright decline and in-app purchases down 38% for the third consecutive quarter of double-digit declines. On a bookings basis the same picture holds: subscription bookings grew 16% while non-subscription bookings fell 9%.

Subscription

The only line growing at scale, and the acceleration is real: subscription bookings grew 16% against 10% last quarter, which management attributed in large part to the strength in user growth. A crude period-end ratio of subscription bookings to paid subscribers gives roughly $21.25 per subscriber against $20.90 a year ago, up about 1.7%, so almost all of the growth is unit growth rather than price. The company publishes no ARPU, so that ratio is a rough proxy and not a disclosure.

Assessment: Subscription is doing what it should and its reacceleration tracks the user reacceleration with the expected lag. The missing disclosure is price. With Max being folded into a family plan, longer free trials expanding, and an ad-supported tier in test, three separate forces are pushing average realised price in different directions and none of them is visible from outside.

Advertising

Advertising revenue fell 1% to $20.8M, having grown 15% in Q1 and 2% in Q2. In absolute terms the line has been stuck near $21M for three consecutive quarters, at $20.6M, $21.1M and now $20.8M, while monthly users grew from 137.8M to 153.1M over the same span. Management devoted substantial airtime to the opportunity for a second consecutive quarter, described a newly built direct sales team, and framed the current inventory as having been sold through ad networks at "the lowest common denominator of ads."

"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." — Gillian Munson, Chief Financial Officer

The exit-rate framing is the first quantitative-adjacent comment on advertising in three quarters, and it sits awkwardly next to the Q4 guidance rationale, which cites "continued strength in advertising and in-app purchases" as a driver of a revenue line both of those streams shrank in during Q3.

Assessment: A business that generates over a billion organic impressions a quarter and 153 million monthly users, and monetizes them at $20.8M a quarter and falling, is either a large unexploited asset or a structurally weak one. Management's answer is that the transition to higher-quality inventory takes time, which is credible and also unfalsifiable until the line turns. We will give it two more quarters before treating the advertising thesis as broken.

In-app purchases

In-app purchase revenue fell 38% to $6.9M, after a 23% decline in Q2. This is the clearest case of the strategy's cost showing up in a reported line: the energy mechanic that drove in-app purchases was precisely the free-tier friction management has committed to removing. The replacement format arrived this quarter in the form of Junior's Toy Hunt, the first event built on a new Avatar Collections system, which the letter described as "one of our biggest experiment wins for getting learners to complete more lessons."

Assessment: The honest read is that management has found an engagement win and is describing it as a monetization opportunity. Collection events that drive lesson completion are good for retention regardless of whether they sell outfits. The test is whether this line stops falling in the fourth quarter, which the guidance language implies it should.

Duolingo English Test

Revenue of $9.9M grew 2% against an easy comparison, and has fallen sequentially in every quarter of 2026, from $11.3M in Q1 to $10.1M in Q2 to $9.9M now. Neither management nor any analyst mentioned the product on the call, now a fourth consecutive quarter of silence. The credentialing ambition set out at initiation has not been discussed since.

Assessment: Immaterial at 3.2% of revenue and shrinking as a share. It is no longer worth modelling as anything other than a flat line, and the absence of any reference to it in a call that discussed chess, math, advertising, partnerships and avatar outfits is itself the signal.

Subjects beyond languages

Chess is the disclosure of the quarter after languages. It has grown from roughly 7 million daily users in February to 10 million now, which makes it the company's most successful subject expansion by a wide margin, and the CEO volunteered that it is completely undermonetized and that this is about to change.

"Another thing that we're really excited about is, Chess is completely undermonetized. 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." — Luis von Ahn, CEO and Co-Founder

Math got a less flattering answer. Asked about adoption and whether a pricing action was available, the CEO said it is growing at about the same rate as languages, which he framed as a disappointment against chess, and offered an explanation that was characteristically blunt: people prefer playing chess to learning math. Music was not mentioned by management or by any analyst, despite a Q2 commitment to provide an update "in maybe a quarter or 2."

"It is growing about at the same rate as languages. So I'm happy that it is growing, but I would have hoped that it can start growing at the rate that Chess is growing, which is growing faster." — Luis von Ahn, CEO and Co-Founder

Assessment: Chess at 10 million daily users and, in the CEO's words, completely undermonetized is the single largest identified and unpriced monetization lever in the business, and it is the most concrete support for the bull pillar that the engagement moat transfers beyond languages. It is also a reminder of how little the company discloses: a product that has added three million daily users since February received no mention in two consecutive shareholder letters and surfaced only because an analyst asked about paid-subscriber trends.

Key Topics & Management Commentary

Overall Management Tone: The most relaxed call of the five we have covered, and the first in which management was not defending a decision. Prepared remarks ran short and the CFO explicitly said the numbers were straightforward, which they were; the posture was that of a team collecting on a plan rather than explaining one. The least convincing area was monetization in 2027, where management repeatedly described direction and process without producing a single quantified expectation, and where the answer to whether conversion rates can improve was a candid "I don't know."

1. The user constraint released, broadly

Daily users grew 29% to 64.9M, accelerating for a second consecutive quarter from 23% and 21% before it, and the CEO led with the number rather than burying it. The breadth disclosure is the part that resists a one-time explanation.

"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." — Gillian Munson, Chief Financial Officer

Asked to attribute the growth between product work and the quarter's two game partnerships, the CEO pushed firmly toward the former, which is the opposite of what a management team reaching for a flattering explanation would do.

"So first of all, the majority of our growth for this year comes from product improvements. ... And that's -- the vast majority of our improvement in active users comes from improvements in retention because of that. The product's getting better." — Luis von Ahn, CEO and Co-Founder

Assessment: Last quarter's acceleration carried an asterisk, because management named a one-time streak campaign as one of three drivers and declined to size it. This quarter there is no comparable asterisk: the named one-time items are two game partnerships the CEO explicitly downplayed, and the growth accelerated in eight of ten major markets alongside a three-point improvement in monthly users. The bear point that has gated this rating since initiation is resolved.

2. Bookings reaccelerated four points, and the lag is now the whole argument

Bookings grew 12% against 8% in Q2, with subscription bookings up six points to 16%, and total bookings growth identical to reported on a constant-currency basis. The CFO tied the subscription acceleration directly to users. Revenue, which lags bookings by the length of a subscription term, decelerated to 14% from 18% in the same quarter, and the letter was explicit about why.

"Revenue growth moderated from Q2 because subscription revenue is recognized ratably over the subscription term, so it reflects the slower prior-period bookings growth. As bookings growth reaccelerates, we expect revenue growth to follow." — Gillian Munson, Chief Financial Officer

The CEO's framing of the lag went further out than one quarter, and it is the clearest statement yet of what shareholders are actually underwriting.

"For bookings growth, our users monetize 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." — Luis von Ahn, CEO and Co-Founder

Assessment: A company whose leading indicator accelerates while its reported revenue decelerates is the textbook setup for a multiple re-rating, and that is what happened. The honest caveat is in the CEO's own words: if new users take years to monetize, the 29% daily-user growth of 2026 is a 2028 and 2029 revenue event, not a 2027 one. Buying that requires conviction in a cohort curve the company does not publish.

3. Duolingo Max is being sunset

The largest open question from the Q2 call is closed. New Max sales stop this month, existing individual Max subscribers move to a Super family plan at the same price, and legacy Max family subscribers will be addressed in 2027. Management said the bookings effect is minimal and already inside the Q4 guide, and for the first time it sized the tier.

"So the vast majority of our bookings now is not Max. It's single digits, both in terms of the number of paid subscribers and the bookings there." — Gillian Munson, Chief Financial Officer

The reasoning is a direct consequence of the AI cost curve. Max existed because Video Call was expensive to serve; once it stopped being expensive, the tier stopped having a purpose.

"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." — Luis von Ahn, CEO and Co-Founder

Assessment: This resolves an overhang we flagged as unquantifiable from outside for two consecutive quarters, and it resolves it in the least damaging available way: a same-price migration to a product with more features, which avoids the one-time bookings step-down a simple cancellation would have caused. The strategic logic is also sound, since a differentiating feature reaching ten times as many subscribers is worth more as a retention and acquisition asset than as a price gate. The residual question is the one management deferred: what happens to legacy Max family subscribers in 2027, and the sizing given, single digits, covers the whole tier rather than just that remainder.

4. Chess reaches 10 million daily users, and monetization is coming

After two quarters in which chess went undisclosed, the CEO volunteered that it has reached 10 million daily users, up from roughly 7 million in February, and that it is completely undermonetized. He named it in the same breath as the new monetization levers the company has been developing, which is the first time a subject outside languages has been put forward as a near-term revenue opportunity.

"Chess is completely undermonetized. We're going to be monetizing it. ... 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." — Luis von Ahn, CEO and Co-Founder

On the pace of adding subjects, the CEO noted that chess went from idea to shipping in about nine months with AI assistance, and that the company can now do it faster. He declined to announce new courses.

Assessment: A chess audience of ten million daily users, roughly 15% of the daily base, that the CEO calls completely undermonetized is the most concrete evidence yet for the bull pillar that the engagement moat transfers beyond languages. It is also the single cheapest monetization lever available, because the audience is already acquired and already engaged. Grade this on whether a chess monetization product actually ships in 2027 rather than on the intent.

5. The 2027 posture: monetization returns, with a guardrail

The CEO stated the 2027 priority shift in prepared remarks and repeated it under questioning, and he was equally clear about the constraint he has placed on it.

"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." — Luis von Ahn, CEO and Co-Founder

Asked how the company avoids repeating the 2025 mistake that cost it a year of user growth, the answer was unusually direct about the error.

"But yes, I mean, this is a lesson learned. We do not want to mess up the free user experience anymore, and I don't believe that's going to happen." — Luis von Ahn, CEO and Co-Founder

Operationally, "more monetization" means more experiments pointed at it rather than new mechanisms: the company told the whole organisation that user growth mattered in 2026 and will tell it in 2027 that monetization matters too. Pressed on whether conversion rates can therefore improve without added friction, the CEO said he did not know, then said he believes in "some increase in conversion rates" from steps such as longer free trials, without knowing by how much.

Assessment: The guardrail is the right one and the candour is worth something, but this is the weakest-supported part of the call. The company has committed publicly to a monetization year without quantifying a single element of it, three months before it must guide 2027. That combination sets up February as a binary event: either the guide carries visible monetization upside, or the market concludes that 29% user growth converts at a rate no better than 14% paid-subscriber growth.

6. Gross margin beat again, and the AI cost curve kept falling

Gross margin reached 72.9% against a roughly 71% guide, the fourth consecutive quarter in which this line has beaten its own forecast. The CFO quantified how wrong the February plan was in the direction that helps.

"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." — Gillian Munson, Chief Financial Officer

The unit economics improved even as AI usage expanded, which is the detail that matters for the competitive risk rather than for the margin.

"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 models." — Luis von Ahn, CEO and Co-Founder

Management also sized, for the first time, what it will spend giving AI features to non-paying users: single-digit millions of dollars of live inference cost in 2027. Against a revenue base guided to $1,221M this year, that is immaterial.

Assessment: Four consecutive quarters of falling AI unit costs against rising AI usage is no longer a surprise, it is the structural advantage of running an education product on open-weight models at consumer scale. It funds the Max sunset, it funds Video Call for every subscriber, and it is why the company can contemplate giving AI features to a 153-million-user free base for single-digit millions. It is also the strongest live rebuttal to the bear case that general-purpose AI compresses this moat.

7. The retention metric was disclosed once, then withdrawn

Last quarter the company disclosed Current User Retention Rate for the first time, at an all-time high of 84%, and we treated it as a genuine improvement in transparency and the first quantitative handle on the product half of the strategy. Asked for an update, the CEO said the metric keeps improving and then declined to keep publishing it.

"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." — Luis von Ahn, CEO and Co-Founder

Assessment: A metric disclosed in one quarter and retired in the next was never a disclosure, it was a talking point deployed when it supported the argument. The substance of the answer is reassuring, because management is saying the retention engine is still the primary driver and still improving, and the daily-to-monthly ratio corroborates it independently. The practice is not reassuring, and it adds to a pattern: chess, math, music, the English Test and now this metric all surface or vanish according to whether they help.

8. The performance bonus is larger than August's framing

In August the CFO volunteered, unprompted, a bonus plan triggering if Q4 daily-user growth reaches 25%, with a cash payout of roughly $10M in the first quarter of 2027 and potentially more at higher growth. The third quarter alone carries an $11.4M accrual, with a second portion inside the Q4 guide.

"GAAP operating expenses grew $39.2 million, or 24%, over the prior year and included an $11.4 million accrual for a company-wide performance bonus tied to achieving our Q4 DAU growth goal of 25% or greater. This bonus was previously excluded from our outlook." — Gillian Munson, Chief Financial Officer

Management also confirmed on the call that it is on track to hit the targets, which is why the accrual started, and that it expects to pay the full bonus in early 2027. The total has never been stated.

Assessment: The accrual is the right accounting and the bonus is probably the best money the company spends this year, because 29% daily-user growth against a 25% trigger is exactly the outcome shareholders wanted. Two things still deserve naming. First, the cost is running above the August framing and the total remains unsized three months after the plan was disclosed. Second, beginning to accrue is itself a soft confirmation that the Q4 daily-user target will be met, which is forward information the company delivered through an expense line rather than through guidance.

9. Advertising: the commentary against the number

For a second consecutive quarter management spent significant airtime on advertising while the disclosed advertising line decelerated. The strategic description is coherent: historical inventory sold through networks at low rates, a newly built direct sales team, a transition to higher-quality brands that pay more.

"And just to remind you about our ads business, we 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." — Luis von Ahn, CEO and Co-Founder

The reported advertising revenue for the quarter was $20.8M, down 1% year over year, against 2% growth in Q2 and 15% in Q1. The CFO's exit-rate comment is the only quantitative-adjacent detail offered, and the Q4 guidance rationale separately cites "continued strength in advertising and in-app purchases," which is difficult to reconcile with lines that fell 1% and 38% respectively in the quarter just reported.

Assessment: This is the clearest gap between narrative and disclosure in the release. We are not calling it misleading, because exit-rate improvement is a real phenomenon that a quarterly average conceals, and because the Q4 guide is a forward statement about a business being rebuilt. But a 153-million-user platform monetizing advertising at $20.8M a quarter and shrinking has no business being described with the word "strength" until the line turns.

10. Collection events, and what in-app purchases actually did

The quarter's named monetization innovation is Junior's Toy Hunt, the first event built on a new Avatar Collections system in which learners earn puzzle pieces through lessons and redeem them for avatar outfits. The letter made both claims for it, engagement and monetization, in one sentence.

"The event was one of our biggest experiment wins for getting learners to complete more lessons, and it showed that collection events can be a meaningful opportunity to grow in-app purchases over time." — Luis von Ahn, CEO and Co-Founder

In the same quarter, in-app purchase revenue fell 38% to $6.9M. Management plans a regular cadence of collection events, and the CEO separately named avatar suits as a lever that "does really well with IAPs."

Assessment: The engagement claim is specific and credible; the monetization claim is aspirational and the reported line contradicts it. The cleanest interpretation is that collection events are replacing the energy mechanic as the in-app purchase engine and have not yet replaced its revenue. Q4 is the first quarter in which that substitution should be visible, and the guidance language implies management expects it to be.

11. Experiment velocity up 90% on roughly 10% more people

The company quantified the AI productivity effect on its own operating model, which is the mechanism behind everything else in this report.

"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." — Luis von Ahn, CEO and Co-Founder

The CEO described non-engineers, product managers and designers, now shipping small A/B tests end to end through an internal tool, said the latency from idea to live test has shortened materially, and expects 2027 experiment counts to be "some integer multiple" of 2026's. He also named the new bottlenecks honestly: generating enough ideas to test, and who fixes the bugs a non-engineer introduces.

Assessment: This is the most investable disclosure on the call after the user numbers. A business whose growth comes from compounding thousands of small experiments has a direct, mechanical relationship between experiment throughput and growth rate, and throughput is up 90% on 10% more headcount. If the relationship holds even loosely, it argues the user reacceleration is early rather than peaking, and it is the strongest reason to expect monetization experiments to produce results in 2027 when they are pointed at that goal.

12. Game partnerships, a channel that costs nothing

The quarter's two partnerships, with Brawl Stars globally and Eggy Party in China, brought what the letter called millions of new and returning users. The economics are the notable part.

"In terms of what is the incentive for the game publishers, there's no money being exchanged in these partnerships. 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." — Luis von Ahn, CEO and Co-Founder

Management designed the mechanic so that claiming an in-game reward requires several days of Duolingo use, which it says produces good retention from the cohort. It also set expectations on scale: a couple of these a year rather than dozens, with most 2027 events being first-party.

Assessment: A zero-cost acquisition channel that arrives inbound because of brand strength is a real asset and a direct continuation of the funnel diversification we credited in August. Management sized the ceiling honestly, which is what keeps it from being overvalued: two partnerships a year is a contributor, not a strategy.

13. The buyback nearly stopped

The company repurchased $6.4M of stock in the third quarter, 48,419 shares at an average near $132, against roughly $44M in the second quarter. Cumulative repurchases moved from $71.9M at August 1 to $82.0M at October 30, so roughly $10.1M was spent across three months. That is 20.5% of the $400M authorised in February, eight months in, with $318M remaining.

"Through October 30, 2026, total repurchases under the program were 781,439 shares, for $82.0 million, offsetting all of the dilution from 2024 and 2025." — Gillian Munson, Chief Financial Officer

Estimated diluted shares rose to 50.8M from 50.7M, a 0.2% sequential increase and 2.2% above the Q1 level, against guidance for 3.5% to 4.0% gross dilution before repurchases.

Assessment: The dilution objection was the strongest bear point in our August note and it has not improved, it has simply stopped getting worse because the share count barely moved. What did change is that the company slowed its repurchases to 8.5% of quarterly free cash flow during a quarter in which it bought at an average near $132, which in hindsight was a good price. Management's framing, that cumulative repurchases have offset all 2024 and 2025 dilution, is accurate and also sidesteps the question of 2026 and 2027 dilution. No analyst asked.

Guidance & Outlook

The revision pattern held for a third consecutive quarter, with one change in character. The top-line guides, which had been reaffirmed and nudged in the prior two quarters, were raised outright this time and placed at the high end of the ranges set in February: $1,296M of bookings against a 10% to 12% framework, and $1,221M of revenue against 15% to 18%. Adjusted EBITDA went up for a third time, by $3M to $323M, which is $21M above the midpoint of the original February guidance. Stock compensation guidance improved by a point. The only line to move against shareholders is the one that went unrepeated: the full-year free-cash-flow guide of over $375M, given on the August call, appears in neither this letter nor this call. With $301.8M generated through nine months, the standing figure requires only $73.2M in Q4, which the company has exceeded in each of the last four quarters, so the omission is more likely housekeeping than a warning. It should still have been addressed.

Implied Q-over-Q ramp: The full-year point estimates reconcile cleanly to the stated Q4 guide, which is a useful check. Bookings of $1,296M less the $308.5M, $289.1M and $316.1M already reported leaves $382.3M for Q4 against a $382M guide. Revenue of $1,221M less the three reported quarters leaves $321.6M against a $322M guide. Adjusted EBITDA of $323M less $239.5M reported leaves $83.5M against an $84M guide. Q4 therefore remains the fastest bookings quarter of the second half at 13% growth, just below Q1's 14% as reported but the fastest of the year at 14% in constant currency, and Q4 Adjusted EBITDA is guided roughly flat year over year in dollars because the remaining bonus portion and continued investment absorb the revenue growth.

Street at: Q3 revenue consensus was $303.2M against $309.0M delivered, and EPS $0.57 against $0.54. The single-source Adjusted EBITDA consensus of $76.2M was beaten by 3.3%, and a single-source bookings consensus implied by 9.5% expected growth, roughly $308.7M, was beaten by 2.4%. The forward comparison is better: the Q4 revenue guide of $322M sits 2.3% above the $314.8M consensus, the full-year Adjusted EBITDA guide of $323M sits above a $319.9M consensus, and the Q4 daily-user expectation of roughly 29% sits six points above a 23.1% consensus. For the first time in the quarters we have covered, every forward guided line came in above the Street rather than below it.

Guidance style: Conservative on profit, and the conservatism is finally narrowing. Adjusted EBITDA has now beaten its quarterly guide by 13.3%, 8.9% and 3.6% in the three quarters since the CFO said she intended to close the gap between guidance and actuals, which is a clean downward progression and evidence the intent was real. Gross margin remains the exception: a fourth consecutive beat, this time by 190 basis points, against a path the February plan had falling to roughly 69%. Bookings and revenue guides have landed within a few points every quarter, which is the intended behaviour.

Analyst Q&A Highlights

Why paid subscriber growth keeps slowing while daily users accelerate

The second question of the call went straight to the gap at the centre of the investment case, and asked whether extended free trials explain it. The answer was mechanical rather than strategic: paid subscribers are an average-balance metric that tracks recognised revenue, which in turn tracks bookings struck a year earlier, so the deceleration reflects the 2025 bookings trough working through the base rather than anything happening now. Management put a timeframe on the trough and then pivoted to the new monetization levers it says are DAU-aligned, naming longer free trials, avatar-suit purchases, and chess monetization.

Q: "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?"
— Wyatt Swanson, D.A. Davidson & Co.

A: "Wyatt, the way to think about subs numbers is they 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 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 it, and those subs growth numbers will come as well. So it's just -- that's just like sort of a math thing mostly."
— Gillian Munson, Chief Financial Officer

Assessment: The mechanical explanation is correct as far as it goes, and "bottoms in a couple quarters" is a falsifiable commitment that can be graded in February and May. What it does not address is the ratio: paid subscribers have fallen from 22.8% of daily users to 20.2% over four quarters, and a lag argument explains the growth rate without explaining the level. The useful takeaway is the date stamp. If paid-subscriber growth has not inflected by the second quarter of 2027, the lag explanation expires.

What the premium-tier sunset reveals about the AI cost curve

A question about early Video Call testing in the mass-market tier produced the clearest account on the call of why the premium tier is going away, and it is a cost story rather than a product story. Max existed to ration an expensive feature; a roughly 97% reduction in per-call inference cost removed the reason for the ration. Management said the A/B test of the change was good for revenue and extends the feature to about ten times as many subscribers, and that it then tested the sunset itself before deciding.

Q: "Luis, just wondering if you could kind of provide a little more color on sort of in the early days of the testing of moving Video Call to 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?"
— Ryan MacDonald, Needham & Company

A: "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. ... And because of that, and when we tested just -- we actually just tested sunsetting Max, it was very clear that the right decision was to sunset Max. So we've started doing that."
— Luis von Ahn, CEO and Co-Founder

Assessment: Management answered a product question with a cost answer, which is the correct framing and the one a shareholder should adopt. The decision was tested rather than asserted, which is consistent with how this company makes every other decision. The sentence that matters most for the long-term case is that the A/B test showed the change was good for revenue, because the obvious objection to universalising a purchase driver is that it removes a reason to buy, and management says it measured the opposite.

Cannibalization risk in moving individual subscribers onto a family plan

The sharpest question of the call noted that migrating individual premium subscribers to a family plan creates sanctioned password sharing, and asked what stops existing individual Super subscribers from cancelling and joining a friend's plan instead. Management's answer was that family plans have materially better retention because a plan survives as long as any one member uses it, that the same-price migration avoids a one-time bookings step-down, and that the A/B testing has not shown material cannibalization. It conceded some will occur.

Q: "And then that creates a potential for more paid sharing, if you will, friends potentially sharing with one another. So I'd 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 friends plan."
— Justin Patterson, KeyBanc

A: "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'll 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."
— Luis von Ahn, CEO and Co-Founder

Assessment: The question identified the real risk and the answer did not fully retire it. "Not material enough" is a judgement about a test population, not a disclosure about the whole base, and the migration has not happened yet. Against that, the retention logic for family plans is genuinely strong and the economic case for tolerating some sharing in exchange for much lower churn is defensible at a company that says it is targeting a billion learners. This is the one element of the Max decision that could still cost money, and it would show up in paid-subscriber counts and subscription bookings per subscriber during the first half of 2027.

Whether 72% gross margin is the new run rate

Asked directly whether the 72% guided for the fourth quarter is a reasonable steady state once Video Call reaches every paid subscriber, management declined to commit and instead disclosed its policy: the appetite to put AI features in front of users is high, the margin will be managed to an overall plan rather than defended as a level, and the company will not guide next year yet. That is a more useful answer than a number, because it tells modellers which variable is the output.

Q: "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?"
— Shweta Khajuria, Wolfe Research

A: "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."
— Gillian Munson, Chief Financial Officer

Assessment: Read this as a soft warning that gross margin is a discretionary variable, not a structural one. Management has beaten its gross-margin guide four quarters running and is telling the market it intends to spend some of that upside on product rather than bank it. For a company whose problem is monetization rather than profitability, that is the right allocation, and it means gross margin should be modelled flat to slightly down from here rather than extrapolated upward.

An update on the retention metric disclosed last quarter

One analyst returned to the retention metric the company published for the first time in August and asked what it has done as daily-user growth accelerated. Management said the metric keeps rising and is responsible for the majority of daily-user growth, then declined to publish the figure or to keep publishing it in future quarters. It was the only question on the call where management narrowed a disclosure rather than widened one.

Q: "Last quarter, we had talked about CURR improvements at all-time highs and new cohorts retaining better. Just curious any updates you have there, especially as the velocity of DAU growth starts to accelerate? What have the overall CURR trends looked like on a cohort basis?"
— Bryan Smilek, JPMorgan

A: "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."
— Luis von Ahn, CEO and Co-Founder

Assessment: The substance supports the thesis and the practice undermines confidence in the disclosure regime. A metric offered in the quarter it looked good and withdrawn in the next one cannot be relied on as a tracking series, so the retention claim now has to be verified indirectly through the daily-to-monthly engagement ratio, which rose again to 42.4%. That corroboration exists, which is why this is a governance complaint rather than a thesis problem.

The ad-supported tier, and the state of the advertising experience

An analyst who had been served the cheaper ad-supported tier repeatedly while testing the app asked where it is going and what is happening to the advertising product. The answer on the tier was unusually unguarded: it is not working well enough to scale, might end up confined to lower-income geographies, and management does not know what will happen to it. The answer on advertising was a plan rather than a result, and the CFO added that the business exited the quarter growing after a flat quarter.

Q: "I think the other change this year to tiering beyond the deprecation of Max is the Lite tier. ... So just any thoughts about your expectations around, kind of, Lite penetration next year and the state of the advertising experience in Duolingo right now would be great."
— Alec Brondolo, Wells Fargo

A: "Lite, we are testing it. We have not pushed it very far. And we may try to push it more. We're mainly driven by the results of our A/B tests. And I think our A/B tests for Lite 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 Lite. 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 per capita."
— Luis von Ahn, CEO and Co-Founder

Assessment: Credit the candour and mark down the lever. In August the ad-supported tier was presented as one of the new monetization vectors that would replace the friction the company had removed; three months later it is a test that has not produced a clear win and may end up regional. Combined with advertising revenue in outright decline, two of the monetization replacements named in August are now weaker than when they were named, which is precisely why the monetization pillar moves to at-risk in our scorecard below.

What a monetization year actually means operationally

A recurring line of questioning on the call asked whether returning to monetization in 2027 means restarting the engine that damaged the business in 2025, or something different. Management described a process change rather than a mechanism change: the same thousands of experiments, pointed at monetization as well as growth, with one category permanently off the table. Pressed immediately afterward on whether conversion rates can improve without reintroducing friction, the CEO said he did not know, then allowed for "some increase in conversion rates" of unknown size.

Q: "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?"
— Andrew Boone, Citizens

A: "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."
— Luis von Ahn, CEO and Co-Founder

Assessment: The strongest and weakest answer on the call at once. Strongest because the guardrail is explicit, permanent and correct, and because the follow-up answer, that the CEO does not know how much conversion rates will improve, is the kind of honesty that makes the rest of his disclosures credible. Weakest because the entire 2027 monetization case rests on an unquantified list, and because the named items are a declining in-app purchase line, a declining advertising line, and a feature the company just made free to existing subscribers. The market has priced the direction; February has to price the magnitude.

Whether AI-native tutoring eventually displaces the gamified product

The most conceptually interesting exchange put the central competitive question directly: one-on-one tutoring is the most effective known way to learn, general-purpose models are improving fast, and a large platform recently copied the company's conversational avatar. Management's answer conceded the capability gap, claimed it closes, and then argued the gap is not the binding constraint because the failure mode in education is attrition rather than instructional quality.

Q: "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?"
— Ross Sandler, Barclays

A: "But we are today not as good as a one-on-one human tutor. Nothing has gotten as good as 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 they'll be, but we're improving."
— Luis von Ahn, CEO and Co-Founder

Assessment: This is the most durable argument in the bull case and it was made without hedging. The claim is that effectiveness is necessary but engagement is sufficient, that a perfectly effective tutor nobody opens is worth nothing, and that the company's actual asset is the retention engine rather than the teaching engine. The supporting disclosure, that nobody holds comparable data on how people learn, is the moat statement. It also implicitly concedes that a general-purpose model can match the teaching and still lose, which is the right place for this debate to sit.

What They're NOT Saying

1. The total size of the performance bonus: disclosed in August at roughly $10M, with a potentially larger payout at higher growth. Q3 alone carries $11.4M and a second portion sits in the Q4 guide, so the total is above the August framing, and three months later it has still never been stated. The cash lands in early 2027, which makes it a first-quarter free-cash-flow item nobody has sized.
2. Any conversion rate, anywhere: the CFO said top-of-funnel users convert "at a multitude" of the core base's rate, which is the single most important quantitative claim on the call, and gave no number for either rate. Paid subscribers grew 14% while daily users grew 29%, and the gap cannot be decomposed into conversion, churn, price or mix from outside.
3. ARPU, for a fourth consecutive quarter: no average revenue or bookings per subscriber, in the quarter that the premium tier is being folded into a family plan, longer free trials are expanding, and an ad-supported tier is in test. Three forces pushing realised price in different directions, and no disclosure against which to judge the net.
4. Why the Q4 guidance cites "continued strength in advertising and in-app purchases": advertising revenue fell 1% and in-app purchases fell 38% in the quarter just reported, and non-subscription bookings fell 9%. The only explanation offered was the CFO's comment on the call, in answer to an analyst's question about the ad experience, that advertising exited the quarter growing. Nothing addressed the in-app purchase decline.
5. The full-year free-cash-flow guide: the "over $375 million" figure raised on the August call was repeated in neither this letter nor this call, in a release that raised bookings, revenue and Adjusted EBITDA. Nine-month free cash flow of $301.8M makes the figure easily reachable, which is why its quiet omission is odd rather than alarming.
6. What happens to legacy premium family subscribers: explicitly deferred to 2027. The "single digits" sizing given covers the whole Max tier, not the remainder left after individual subscribers migrate, so the residual exposure is smaller than single digits by an unstated amount and the resolution is a 2027 event with no stated mechanism.
7. Music, at all: no user count, no revenue, and no mention by management or by any analyst anywhere in the letter or on the call, despite an August commitment to provide an update "in maybe a quarter or 2." Math received only a qualitative answer and no user figure. Chess surfaced at 10 million daily users only because an analyst asked an unrelated question.
8. The Duolingo English Test, for a fourth consecutive quarter: $9.9M in the appendix, down sequentially in every quarter of 2026, and unmentioned on the call by anyone. The credentialing opportunity described at initiation has not been addressed since.
9. Why the buyback nearly stopped: $6.4M repurchased in Q3 and roughly $10.1M across the three months to October 30, against $318M of remaining authorisation and an average purchase price near $132 in the quarter. Management's framing is that cumulative repurchases have offset all 2024 and 2025 dilution, which does not address 2026 or 2027 dilution guided at 3.5% to 4.0%. No analyst asked.
10. Any 2027 number of any kind: management committed publicly to a monetization year, to multiples more experiments, to monetizing chess, and to continued AI feature expansion, and attached a figure to exactly one of those commitments, the single-digit millions of inference cost for free users. Every other element of the 2027 plan was described directionally, three months before the company must guide it.

Market Reaction

• Pre-print setup: DUOL closed at $128.00 on November 3, down 27.1% year to date against the S&P 500's +10.6%, down 51.7% over the trailing twelve months and down 11.3% over the trailing thirty days. The 52-week closing range entering the print was $90.03 to $270.64. Short interest stood at roughly 7.15 million shares, between 18% and 20% of the float depending on the denominator used, with about 6.7 days to cover.
• Into the print: the stock fell 5.4% on November 3 itself, from $135.27 to $128.00, on roughly 2.8 times average volume. Positioning going into the release was negative.
• After hours: shares traded at $141.00 at 4:25 p.m. ET, up 10.2% on the release alone, and the move extended through the 5:00 p.m. call and overnight to roughly +19% in Wednesday pre-market.
• Reaction session: the stock opened at $160.00 on November 4, a 25.0% gap, traded a range of $157.00 to $175.56, and closed at $161.27, up 26.0% or $33.27. It finished 8.1% below its intraday high, at roughly 23% of the day's range.
• Volume: 9.6 million shares against a 30-day average of 0.9 million, or 10.7 times normal, the highest volume multiple of the five reaction sessions we have covered.
• Market context: the S&P 500 rose 1.6% on the session, so the overwhelming majority of the move was idiosyncratic.

Four consecutive reaction sessions in this coverage were negative, falling 25.5%, 14.0%, 5.6% and 9.4% in turn, on a volume multiple that fell from 10.1 times to 3.1 times as the shareholder base turned over. The fifth broke the pattern in both dimensions: the first positive move of the series and the highest volume multiple of it. A base that had stopped producing forced selling on disappointment turned out to still produce forced buying on a surprise.

The trigger is identifiable and it is not the financial beat. Revenue beat by 1.9% and Adjusted EBITDA by 3.3%, both unremarkable, and GAAP EPS missed. What moved was the operating metric: 29% daily-user growth against a Street at roughly 23%, a Q4 expectation of roughly 29% against a Street at 23.1%, and a four-point bookings reacceleration against a Street looking for 9.5% growth. A stock that had fallen 52% in twelve months on the premise that user growth was structurally broken repriced when the premise failed.

Three features of the session argue against reading it as a pure fundamental re-rating. The move built in stages rather than landing at once, from +10% on the numbers to +26% by the close, which is the signature of a narrative being absorbed rather than a model being updated. The 10.7 times volume multiple against a short base equal to three quarters of the day's turnover suggests a meaningful share of the buying was not discretionary. And the stock closed near the bottom of its range after reaching +37% intraday, so the marginal buyer above $170 did not stay.

The cleanest summary is that the market has now paid for the user reacceleration in full and in one session. At $161.27 the stock trades above the Street's average target and above all but three of the raised targets we can find, which is an unusual configuration and one that puts the burden of proof back on the company at the February guide.

Street Perspective

Debate: Does 29% user growth convert, or is the free base a cost?

Bull view: The bull case being made is that bookings already answered the question, accelerating four points to 12% and subscription bookings six points to 16% in the same quarter users accelerated, which is the conversion showing up on the leading indicator. Management says new users convert at a multiple of the existing base's rate, the company is pointing its 90%-faster experiment engine at monetization in 2027, and chess sits at 10 million daily users that the CEO calls completely undermonetized.

Bear view: The bear camp contends that the only disclosed evidence of monetization is subscription, that the other large streams are shrinking, and that the paid-subscriber ratio against daily users has fallen from 22.8% to 20.2% over four quarters. On this reading a free base compounding at 29% while the paying base compounds at 14% is a gross-margin and support cost with an option attached, and the option has an unknown strike.

Our take: The bulls have the better of it, but by less than the share price now implies. The bookings reacceleration is real, clean of currency effects, and broad across regions, and it is hard to produce without conversion. What the bears get right is that management has quantified none of it, and that the two monetization vectors named as replacements in August, the ad-supported tier and advertising, are both weaker now than when they were named. We would want the February guide to show mid-teens bookings growth or better before treating the conversion question as settled.

Debate: Was sunsetting the premium tier a gift or a concession?

Bull view: The optimistic reading is that this is the AI cost curve paying a dividend. A feature that cost thirty cents a session had to live behind the most expensive tier; at under a cent it belongs in the mass product, where it reaches ten times as many subscribers and functions as a retention and acquisition asset. Management A/B tested the change and says it was revenue-positive, sized the exposure at single digits, and structured the migration at the same price to avoid a bookings step-down.

Bear view: The skeptics note that the company is deleting its highest-price product in the same quarter it says monetization becomes a priority, that moving individual subscribers to a family plan sanctions sharing and reduces revenue per seat, and that the sizing disclosure arrived only after the decision was irreversible. A tier that was a strategic pillar two years ago has become something management describes as "a very small thing."

Our take: A concession that is also the right decision. The bears are correct that the premium tier failed, and that the company spent two years building a price ladder it has now dismantled. But the alternative was defending a tier whose only differentiating feature costs a cent to deliver, against a mass-market product that could carry it, and that is not a defensible position. The same-price migration is competent execution of a decision that should have been obvious once the cost curve broke, and the residual risk is cannibalization in the first half of 2027.

Debate: Is the stock still cheap after a 26% session?

Bull view: The bull argument is that an enterprise value near $6.8B against guided bookings of $1,296M is 5.3 times, against guided Adjusted EBITDA of $323M is 21 times, and against standing free-cash-flow guidance of over $375M is roughly 18 times, for a debtless business with 17% of its market value in net cash, 29% user growth, and an Adjusted EBITDA margin deliberately held three points below the prior year. Normalise the margin toward the 29.5% achieved in 2025 and the multiple on 2027 earnings power is materially lower than it looks.

Bear view: The bear response is that the enterprise value rose roughly $1.7B in a session on a 1% increase to the full-year guided numbers, that the stock now trades 14% above the Street's average target, that the business is more subscription-concentrated than when the strategy began, and that the next catalyst is a 2027 guide management has deliberately refused to pre-frame. At 5.3 times bookings the valuation discount that justified owning this in February is substantially gone.

Our take: The bears have the better of the near term and the bulls the better of the twelve-month frame. The re-rating was a multiple event, not an earnings event, and anyone buying here is paying 5.3 times bookings for a monetization plan with no numbers attached. That is a different proposition from the 3.0 times available in February or the 3.8 times available in August, and it should be said plainly. Over twelve months the user base, the cash balance, the cost curve and the undermonetized chess audience still argue for a higher valuation than this, which is why the rating holds and the conviction does not.

Model Update Needed

Valuation impact: At the $161.27 reaction close and 50.8M estimated diluted shares the equity is worth roughly $8.19B. Netting $1.37B of cash and short-term investments gives an enterprise value near $6.82B, or approximately 5.3 times the full-year bookings estimate, 5.6 times revenue, 21.1 times Adjusted EBITDA and roughly 18 times the standing free-cash-flow guide. Net cash is 16.8% of market value, down from 21% in August purely on the share price. We raise the target to $185 from $155, which implies +14.7% from the $161.27 reaction close. The $185 applies roughly 20 times our FY27 Adjusted EBITDA estimate of $398M, the same multiple the August target used rolled one year forward. On FY26 figures the same target is 6.2 times bookings and 24.8 times Adjusted EBITDA, which is why it depends on the 2027 estimate rather than on this year's. A 29% margin on the same bookings assumption supports $189 per share; 12% bookings growth at a 27.5% margin supports $179.

Thesis Scorecard Post-Earnings

Overall: The thesis strengthened and simultaneously rotated. Bear #1 has been the binding constraint on this rating since initiation, and this quarter retires it: user growth accelerated for a second consecutive quarter, broadly, on a metric management does not fully control, and the 2028 target now looks conservative against the run rate. In its place Bull #2 becomes the open question, and it moves to at-risk on evidence rather than on sentiment. Three quarters into a year whose premise was that a larger free base creates monetization capacity outside subscriptions, the free base is 29% larger and non-subscription revenue is 8% smaller. The thesis is no longer about whether this company can grow users. It is about whether it can charge them.

Action: Maintain Outperform; raise the target to $185 from $155. In August we said the next thing to grade was the fourth quarter and the 25% internal daily-user target, and the company has effectively pre-announced both by starting to accrue the bonus. That earns the target raise. It does not earn new money at $161.27, where the stock has closed through our prior target, through the Street's average target, and through all but three of the raised sell-side targets we can find, in a single session on ten times normal volume against a short base worth three quarters of the day's turnover. We are holders, not buyers, into February. The February print is now the whole position: a 2027 bookings guide in the mid-teens or better confirms that 29% user growth converts and supports a higher target still, while a guide at or below this year's 12% would say the free base is a cost rather than an asset, and would cost the rating.