| Account | Shares | Price | Value | % of acct | Cost/sh | Gain $ | Gain % | Target |
|---|---|---|---|---|---|---|---|---|
| 401K | 11 | $158.50 | $1,744 | 0.07% | $188.65 | $-332 | -16.0% | — |
| HSA | 2 | $157.85 | $316 | 0.29% | $103.50 | $109 | +52.5% | — |
| ROTH | 27 | $158.50 | $4,280 | 1.67% | $323.16 | $-4,446 | -51.0% | — |
| Total | 40 | $6,339 | 0.14% | $-4,669 | -42.4% | — |
In short: Mixed: "the utility of it is exposed but the motivational aspects are not." An agent that speaks every language reduces the practical need to learn one, but most learners want to communicate and join a culture — a hobby and identity built on streaks, leagues and habit — so "I would not extend that into Duolingo is highly exposed overall."
If an agent can read and shop in any language, you need less language skill to get things done. But most Duolingo users aren't learning to buy things abroad; they want to talk to people and connect with a culture, and they enjoy the streaks and progress. That motivation isn't something an agent replaces, so he doesn't see Duolingo as highly exposed.
31:47A lot of people, in fact, the majority of people learning in Dualingo aren't doing it because they want to be able to just transact easier. They really want to be able to communicate and be part of a different culture. So, I would not extend that into Dualingo is highly exposed overall. Language learning is also a hobby, an identity, a project, a social signal, and a long-term achievement.
In short: $37.4k, −$9.5k, but "my thoughts have not changed": DAU growth re-accelerating (21%→23%, an accidentally shown 27.4% for Aug 17), AI video-call inference cost under $0.01, retention at an all-time-high 84%, and Chess.com copying its UI. ~13% expected return but "a pass-fail situation" — a smaller, non-core position.
Duolingo is his only losing position. Investors worry AI chatbots will replace language apps, but the numbers he tracks are moving the right way: daily users are growing faster again, the cost of its AI conversation practice has fallen below a cent per call, and user retention is at a record 84%. Chess.com even copied its design. He calls it pass/fail — either a very large learning platform or a disappointment — so he keeps it as a smaller position.
In short: Buy target $90 from $146 — "nearly a 38% drop… a staggering drop for most companies, but Duolingo is not like most companies. And it was at $90 per share as recent as April of 2026." Still his worst position ($35k / 2.4%, −$11.2k) after growth decelerated and marketing virality faded. He restates his rejection of the AI bear case explicitly: "the biggest concern for Duolingo today is obviously centered around AI and AI assisted learning. Many people believe that ChatGPT or Claude or Gemini will take over AI learning… I don't share that bear thesis… they're going to continue growing in their users and engagement and learning because that's all they focus on." At 18.5% EPS growth and a 27 multiple, $90 gives 14.2% — with "much more upside if they get to scale."
Duolingo is the language-learning app, and it is his worst position — down $11,200 on a $35,000 stake after growth slowed and its famously viral marketing lost some steam. The bigger overhang is the one everyone names: if ChatGPT, Claude or Gemini can teach you Spanish, why does a dedicated app survive?
He rejects that outright, and his reason is focus rather than technology: "I don't share that bear thesis… they're going to continue growing in their users and engagement and learning because that's all they focus on." The implicit argument is that a general-purpose chatbot and a product engineered around daily habit, streaks and structured progression are not the same product, even when both can technically answer a grammar question.
His buy target is $90, roughly 38% below the current $146 — a drop that would be extraordinary for most companies but is unremarkable here: the stock started the year at $185, traded at $90 as recently as April 2026, and has climbed back since. At $90, assuming 18.5% earnings growth and a 27 multiple, the return works out to 14.2% a year, and he thinks the growth assumption is conservative "if they get to scale."
20:14Nobody should really complain about earning 12% per year for 5 years. It seems low compared to these other stocks, but it's still great. Next up, we have the Green Owl, which is Duolingo at position number 12. It's a $35,000 position, which is a 2.4% weighted holding, and it has $11,200 in the red. So, this one has been the biggest loser by far since buying it.
In short: Streak revival. Q2 revenue +18% Y/Y to $298M (a $3M beat) with GAAP EPS $0.66 ($0.05 beat) — but bookings grew just 8% to $289M, the number that matters for future revenue. DAUs grew 23% to 59 million, accelerating from 21% in Q1, with management expecting DAU growth above 20% for the rest of the year. The asterisk: a one-time Streak Revival campaign brought millions of lapsed users back by letting them recover lost streaks, while current-user retention hit an all-time high — "management says those revived users are retaining unusually well, but Q1's concern isn't fully settled yet. The company still needs sustained new-user acquisition, not just better retention and reactivation." On costs, Video Call has fallen below one cent per session as Duolingo shifts appropriate workloads to open-source models, letting the feature expand from Max into the far larger Super base, with longer free trials as another conversion experiment. FY26 bookings growth held at roughly 11% and revenue 16%, with Q3 revenue of ~$302M slightly below consensus; shares fell more than 10% despite the DAU acceleration. "The real test is whether growth stays above 20% once the reactivation campaign rolls out of the comparison." A disclosed author holding.
In short: "I still hold all of my shares of this company going into this week" — a $33,000 Story Fund position, $13,400 in the red, reporting this week. He frames it as the asymmetry of stocks (his worst-ever pick costs $13k against $500k+ lifetime gains). Not a full endorsement — "I'm not 100% confident of Duolingo," small-cap, volatile, and still at a healthy valuation assuming growth + operating leverage — but fundamentals are improving and AI has fixed the product's biggest historical flaw (you now speak and converse with the characters). "The product improves and then the metrics will follow suit over time."
Duolingo is the language-learning app, and it is his worst position — a $33,000 stake sitting $13,400 in the red. He is holding every share into this week's earnings, and uses it to explain the "asymmetry of stocks": Peter Lynch only had about six winners in ten, but the winners rose far more than the losers fell. Carlson's own version is stark — his single biggest mistake has cost $13,000 against more than $500,000 of lifetime gains. A portfolio should never be able to collapse on one name.
He is unusually candid that this is not a high-confidence call: "I'm not 100% confident of Duolingo." It is a small, volatile company with many bears, and even after the fall it trades at a valuation that assumes solid growth and improving profitability — so there is real downside if the numbers stay weak.
What keeps him in is the product. Learning a language on Duolingo used to be typing and filling in blanks with almost no speaking; AI has turned it into constant spoken conversation with the app's characters, fixing its biggest long-standing flaw. Product improvements show up in the reported numbers only later — "the product improves and then the metrics will follow suit over time."
20:45So, I'm not 100% confident of Duolingo. But what makes me confident enough to continue my investment in it is I do see the fundamentals improving. And I've also noticed something with this company. If you haven't used a Duolingo app in a long time, you probably have an outdated view on the app itself. Duolingo has been infusing their app with AI to make it far more advanced.
In short: His most unpredictable holding (~14% revenue, ~17% EPS, but analyst estimates "all over the place"). The one moat he concedes is weaker / "under test" — AI lowers the barrier to clone personalized language learning. But he still holds every share: brand, the freemium price, streaks, leaderboards, the social layer, the data-flywheel and a hard-to-build structured curriculum are also the moat. "The story remains for me."
Duolingo is the one holding whose moat Carlson openly concedes is "under test." Because AI makes it much easier to build a personalized language-learning app, new competitors are popping up, and the earnings outlook is so uncertain that analyst estimates are "all over the place." But he still owns every share, arguing the real moat is much more than how hard the app is to code: a trusted brand, a free-to-use price, daily streaks and leaderboards, a social layer where friends cheer each other on, a data-flywheel from millions of daily learners, and a structured curriculum that takes years to build well. He thinks it stays prosperous even with the moat being tested.
In short: Portfolio aside — down 70% but a much smaller position (~$15k red) so it stings less than Meta, and it's "actually moving upwards" toward break-even. He still holds.
7:08It is my biggest loser by far. Meta is the biggest one that I'm in the red on. You can even compare it to the Duolingo, right? Duolingo is down 70%. But because that was a much smaller position, that's only down around 15,000. It's actually moving upwards. We're getting closer to break even on that one.
In short: Very high conviction $1,000 long-term target. FCF/share rising and inflecting while the stock is down ~78% — "a violent re-rating." Not a language app but a personalized superhuman AI tutor; the AI-disruption fear is "totally trivial"; only real bear case is flat MAU top-of-funnel, which he believes management can re-accelerate.
Duolingo is the language-learning app with the green owl. Darnton's whole method rests on one idea: over the long run a stock's price tracks its "free cash flow per share" — the spare cash the business generates each year, divided across all its shares. Right now Duolingo's cash generation is climbing and accelerating while the stock has fallen about 78% from its high. To him that gap is a coiled spring: when the market finally notices, the stock "re-rates" violently higher. His long-term target is $1,000 a share.
Why isn't he scared of AI killing it? Because these apps aren't a few lines of code anyone can copy — they're millions of tiny tweaks layered over years (he calls it the "innovation stack") that make them addictively smooth to use, the same moat that let Spotify beat Apple and Amazon. And Duolingo's real moat is its proprietary learning data: generic chatbots like Claude have to be fine-tuned on exactly that kind of data to teach well, and a casual user can't recreate it. AI actually helps Duolingo — it shipped 10x more course content in a quarter thanks to AI.
The one genuine worry is "top of funnel": the number of monthly users went flat. But he argues there's huge room left (only ~12% of users pay vs ~50% at Spotify; over 3 billion people are learning math and English and Duolingo has barely scratched it), the company is deliberately spending on growth now and monetizing later (the "Bezos algorithm"), and management has a strong record of fixing exactly this kind of problem fast. With ~$400M of cash flow, $1B in the bank, no debt, and ~85% of all language-app daily users, he calls it "a monopoly on sale."
0:00Hey everyone, welcome back to the podcast. Today I wanted to discuss why I believe Duolingo stock is going to a thousand dollars a share long term and why the market is totally and utterly wrong about this business. What we can see, as you can see on the slide that I'll share on the screen, is that free cash flow per share is going up and up.
In short: Still holds every share and adds on dips; 74% off highs, the absolute bottom of its valuation range. "Just an app" — so was Spotify; he wants to give the scaled-education story years. Buy.
Duolingo is the language-learning app. The stock has been crushed — 74% below its high, swinging between $88 and $489 over the year, and now at the absolute bottom of its valuation range. He still owns every share and has been buying more as it falls.
Critics dismiss it as "just an app," but he points out Spotify is "just an app" too and did fine. He believes a scaled, digital education platform is a big long-term market and wants to give the story years to play out, so it stays a high-conviction buy for him.
14:00The company's growing in the mid teens. Their earnings per share is growing strong. They're gaining market share and subscribers. Uh Netflix, I think, is firing on all cylinders. So, I believe this one's a buy today. Now, of course, another stock we can't forget about is Dolingo. I still hold every share of Dolingo.
In short: Also debated, but people use it for the motivation — streaks, scores, social — more than the content, so he expects it to stay differentiated and rerate higher.
18:29I think Uber and DoorDash will be spectacular winners. I think that Shopify is a spectacular winner. I think that Airbnb will do really well. We even have companies that there's a bit more debate about companies like Spotify and Duolingo. They don't own proprietary data per se. Spotify just has access to music. Duolingo has access to AI curriculum and courses.
In short: #6 pick and the only new-economy consumer name in the list. "Users keep coming back because learning feels like a game. That's why twelve million people now have a paid subscription. This creates high-quality recurring revenue." Expanding beyond languages into maths and music; "since 2019, the number of paying subscribers has grown by 55.7%" a year. Table: 40.0% net margin and 37.7% ROIC — the best pair on the page — against 44.5x forward and no long-term EPS growth estimate. Reported, not rated.
Nothing matches this filter.
Verbatim excerpts from the public transcripts (auto-pulled at each mention's timestamp, lightly cleaned). Timestamps link into the video; "source page" opens that commentator's full analysis at this row. Click a mention's header line to expand it (one open at a time). For personal study — not investment advice.