Be

$DUOL CEO Luis von Ahn, on scaling learning content: “We’ve now launched content up to Duolingo Score 129 (CEFR B2) across courses teaching our 9 most-learned languages, so learners can use Duolingo to reach professional proficiency. Given the number of language combinations we teach, expanding to this level required building an enormous amount of content. In Q1 2026 alone, we published 20,500 course units across our language courses, up from 7,100 per quarter in 2025 and 1,800 per quarter in 2024. Automating more of our content creation process also lets us iterate in ways that were not possible before. We can now push changes across many courses at once and improve quality more quickly and consistently. This is already improving engagement among new users.” Generative AI is accelerating $DUOL. Content creation and iteration are going vertical. This might take a few quarters to fully kick in, but over time it should translate into better teaching, stronger user growth, and eventually monetization. (Not investment advice.)

I do have some thoughts on SBC as well. There were some special circumstances this Q, as you point out. And there was the recent expiry of the 2021 incentive plan (granted near the all-time peak), which removed a lot of upside optionality for management. The board may have felt compelled to do something to offset that impact. Also I feel that the senior leadership team is doing a hell of a job and deserve to get paid. It is also worth pointing out that $LMND has continued to leverage SBC (i.e. SBC / revenues has been on a steady downward trend). So as long as SBC continues to scale like opex and not like revenues I don't think there is a problem long-term. In my mind, the main problem here was one of optics. In an environment where investors are asking a lot more questions about SBC their guidance for a higher run rate of SBC invites questions like the ones I was responding to. If $LMND wanted to put these questions to rest permanently they could do what some other companies have done and set clearer targets around continued SBC leverage. But that is up to them. And to reiterate, in substance I don't think there is a problem here.

I exited $SE I just got one very , I think, reasonable observation. They are lending to the buyers on their marketplace. This is a difficult model to pull off when you lend to customers that buy from you. 80% of the Y/Y growth in the ecommerce topline can be attributed to Shopee Pay loans through Monee. Hey, I will loan you $20,000 to buy my jeep and I'll book a profit on the lending and the sale. Look, as long as you get the $20K back it is ok. But what if they want to keep sales momentum and they make loans they shouldn't . I just don't like this setup... You should never be the lender and the seller at the same time imo. But I guess auto dealers are the same. Seems odd to me... Amazon doesn't do this. I just wanted to point these things out. Bye!

Just in: Shopee $SE is adding a 5% technical support fee on all cross-border sellers from China to Singapore, Malaysia, Thailand, and Vietnam, starting 2026.02. Shopee will also give out ad credits in return, although the ad credit amount will be lower than 5% of the overall sales. — Before this fee change, they already announced a massive increase in commission across all 6 countries in SEA + Taiwan (excluding Indonesia).

Today’s reason, because $DUOL is still trading below $180: Duolingo and the Math Nobody Seems to Understand Chart 1 shows the cash Duolingo generates, and that is ultimately why you invest in a company. Last year there were two quarters where cash didn’t increase. That’s when you pay attention. Management explained that this pause came from higher AI expenses during the rollout of Duolingo Max, their premium tier designed to help you learn a language faster and better. Those temporary AI expenses pushed margins lower for a short period. But here’s the key point: the price of using AI always comes down. Duolingo benefits from that trend more than almost any company. Either their costs fall as models get cheaper, or they can buy far more AI capacity for the same dollar. They don’t need to invest billions in chips or datacenters, and they don’t carry the risk of wondering whether those investments will ever pay off. By listening to every earnings call and studying the numbers each quarter, you already know in advance when margins will dip. So when this temporary margin pressure appeared, it wasn’t a surprise. Yet almost no one seems to do their homework, and a single datapoint is enough for people on X to tell me I’m wrong. Chart 2 shows this margin dip. Once Duolingo Max was launched and AI costs started falling, and AI costs always fall quickly, margins recovered. Go back to Chart 1 and you immediately see how much cash the company added again. Margins will dip again, because Duolingo has openly said they can already see the point where the product becomes as good as a private tutor. And reaching that level requires more reinvestment. What do you think margins will look like once Duolingo is as good as a private tutor? Doing your homework is the common thread in investing. While Duolingo is racing toward a level comparable to a private tutor, people still reply under my posts claiming that the app teaches only simple sentences you can’t use in daily life. That view is years out of date and clearly shows they haven’t looked closely. Chart 3 is Spotify. The only reason this chart is here is to show what happens when a company grows 50 percent per year and then sees its percentage growth slow sharply into single digits. Cash flow tends to explode during that transition. The reason is simple: margins improve, and absolute revenue keeps rising. Many people told me Duolingo’s growth is “slowing,” but what they mean is percentage growth. And percentage growth is irrelevant without context. Fifty percent growth on 100 million adds 50 million. Ten percent growth on one billion adds 100 million. So nominally the growth isn’t 80 percent lower, it’s actually 100 percent higher. 😉