Duolingo (DUOL) – GenAI Risk – February 10, 2024
Duolingo has masterfully executed in every sense of the imagination since going public. The effect of this perfection? 2024 revenue estimates are a full 45% higher than they were when it went public; 2023 (wraps up when it reports Q4) EPS estimates now call for $2.29 in EPS vs. -$0.20 originally. That’s not a typo. It has been absolutely perfect.
But what’s the main risk I see involving this business model? GenAI. Google and many others are working on real-time language translators, which leaves many wondering if Duolingo will be displaced in the years ahead. I think that’s the most pressing bear case surrounding this name. And yet I’m confident in holding shares today. Why? A few reasons.
I see Google’s real-time translator as more of a complement to Duolingo rather than a replacement. Duolingo users are learners. They are trying to master a language to more closely relate to a given culture or a spouse. Ask yourselves… Do you think it’s more attractive for a British husband to talk to his Spanish wife through an AI translator, or on his own? Which husband gets more brownie points? Then ask yourself if a job applicant who is fluent and conversational in English would be more preferred over one who must talk to their boss and customers through a machine. I think the answers to these questions are clear; there’s a natural need for human connection, which AI does not supplant. Google’s product will be popular for pressing, real-time use cases (i.e. business contract translations). Duolingo’s product will remain popular for the learning and entertainment use cases it has always aimed to provide. For evidence, consider Chegg’s putrid results over the last few quarters due to GenAI proliferation, and then see Duolingo’s flawless financial execution during the same period. Chegg is for cheating; Duolingo is for learning.
Finally, while Duolingo is today’s language learning app, tomorrow it will need to simply be the language learning app to drive continued explosive shareholder returns. Morphing from language education to education vastly expands its target market and also insulates it from the GenAI language risk if my opinion turns out to be wrong. Its literature and math courses are off to great starts with its music course coming soon. There will be many, many more subjects on the app. While GenAI is a risk, it’s not one that has held back Duolingo’s results thus far and is one that I view the company as being well-equipped to combat.