When a stock falls from $544 to $154 in eight months, you're either looking at a company that completely fell apart, or Mr. Market having one of his signature meltdowns over something that won't matter in three years.

With Duolingo (DUOL), I honestly can't tell which one it is yet. And that bugs me.

Here's what I do know: 50 million people use the green owl app every single day. Revenue's up 41%. The business prints cash. And Wall Street just watched management say "we're prioritizing long-term user growth over short-term profits" and proceeded to nuke the stock 25% in a single session.

So yeah, this one requires some thinking.

Duolingo reported Q3 2025 earnings on November 5th, and if you just read the headline numbers, you'd wonder what everyone's so upset about.

Revenue hit $272 million, up 41% year-over-year. Daily active users crossed 50 million, growing 36%. Paid subscribers reached 11.5 million. The company's guiding to over $1 billion in revenue for 2025, not bad for an app that teaches you Spanish using a cartoon owl and passive-aggressive notifications.

Earnings per share? Try $5.95. That crushed the $0.76 estimate by 683%.

Unfortunately, that the EPS number is complete nonsense.

Duolingo got a one-time tax benefit worth $222 million. Strip that out and you're looking at adjusted EBITDA of $221 million through nine months, which is genuinely impressive, but the "beat" everyone's celebrating was basically an accounting gift. Real earnings were more like $0.60-0.70 per share (below estimate).

Obviously, I'm not saying they're cooking the books. I'm saying if you bought this stock because you saw "EPS beat by 683%!" on CNBC, you got played.

So why did the stock crater despite "beating" estimates? Because CEO Luis von Ahn stood up on the earnings call and essentially told Wall Street: "We're done optimizing for your quarterly numbers. We're building for the next decade."

Specifically, Duolingo said they're shifting resources away from near-term monetization experiments and toward long-term product development. They want to build an app that "teaches better than a human tutor" using AI. Noble goal, but couldn’t be more terrible timing to announce it.

Q4 guidance came in light, with bookings between $329.5M and $335.5M versus the $344.3M analysts wanted. And just like that, KeyBanc downgraded the stock, Morgan Stanley cut their price target from $300 to $275, and investors who'd been holding through a 50% drawdown finally threw in the towel.

I mean, I get it. You don't tank your stock to make a philosophical point about education unless you're either a visionary or an idiot. Von Ahn might be both, obviously the jury's still out.

But he’s not wrong about the opportunity.

Everyone's shoving "AI-powered" into their investor decks like it's 2021 and we're calling everything "blockchain-based." Most of it's vaporware.

I believe that Duolingo's actually using AI in ways that matter.

Their premium tier, called Duolingo Max, lets you have video call conversations with an AI character named Lily. You practice speaking French or Mandarin with a chatbot that doesn't judge you for sounding like an idiot. There's also "Roleplay" for situational practice and "Explain My Answer" which gives you personalized feedback when you screw up a lesson.

Max subscribers are only 9% of the paid user base right now, but those bookings doubled year-over-year. And the kicker is that management admitted Max is "underperforming our lofty expectations."

Translation: They thought this would be way bigger by now.

That's either a red flag that nobody wants AI tutoring, or it's a sign they haven't figured out how to sell it yet. I'm leaning toward the latter, because the product actually works - I've tried it myself. The issue is most people learning Spanish on the subway don't want to pay $30/month when the free version gets the job done.

But if they crack the conversion problem? Different ballgame entirely.

The other AI win for Duolingo is in the speed of content creation. Duolingo launched 148 new language courses in 2025 using AI, versus 100 courses over the previous 12 years combined. That's an 18x acceleration. They went from being limited by how fast humans could build curriculum to basically unlimited content generation. This is what makes AI incredibly significant as a general theme for the coming decades - it is the modern day equivalent of the invention printing press for businesses developing content then spreading it far and wide with speed.

Whether that content is real quality remains to be seen. But the leverage is undeniable.

This is where things get interesting for the growth story.

Duolingo's fastest-growing course right now isn't Spanish or Mandarin or even Ukrainian (which spiked after Russia's invasion). It's chess.

Chess. They added a chess course, and it's now bigger than math and music combined, with millions of users and retention rates higher than language learning.

What the hell is happening here?

I think what's happening is Duolingo accidentally discovered they're not actually a language-learning company. They're an engagement-based education platform that happens to do languages really well. The gamification, with streaks, leaderboards, that owl shaming you when you skip a day, it all works for anything you can turn into bite-sized lessons.

So now they're teaching chess. And math. And music. And who knows what else is coming.

This either means the TAM (total addressable market) is way bigger than anyone thought, or they're about to dilute their brand by becoming "that app that teaches everything poorly." Duolingo for coding? Duolingo for investing? At some point you're just Khan Academy with better notifications.

But if I'm being honest, the fact that chess is working makes me more bullish, not less. It validates that the engagement model transcends the original use case.

Here's a stat that should terrify you or excite you, depending on your risk tolerance: China is now Duolingo's second-largest market by daily active users and growing faster than anywhere else.

Luis von Ahn mentioned on the earnings call they're barely spending money marketing there, and it's organic growth has been driven by the Luckin Coffee partnership where they sold 10 million Duolingo-branded drinks in two weeks across 26,000 stores.

So China's 5-6% of the business today with minimal spend. That's the good news.

The bad news: China is China. Regulatory risk, geopolitical risk, the possibility that Beijing decides foreign education apps are a threat to "common prosperity" and bans them overnight. We've seen this movie before with Didi, with Alibaba's Ant Group, with every tech company that got too big.

If China becomes 20% of Duolingo's business in three years and then gets shut down, this stock goes to $80. But if China becomes 20% of the business and stays there, this thing could be worth $400.

It's a coin flip dressed up as a growth story.

Okay, so the stock's at $154. Down 72% from the May 2025 high of $544. Forward P/E is 38.6x, which sounds expensive until you remember this company's growing revenue at 40%+ and just turned consistently profitable.

The PEG ratio, P/E divided by growth rate, is 0.78. That's Peter Lynch territory. Anything under 1.0 is supposed to be a bargain. That certainly fits in with our GARP (Growth at a Reasonable Price) Investing Framework.

Now, PEG ratios are imperfect. They assume growth continues linearly (it won't). They ignore profitability quality (important when you're burning cash). But as a rough screen for "growth at a reasonable price," this passes.

Compare that to, say, Duolingo's edtech peers:

At $154, you're paying for a company growing DAUs at 36%, revenue at 41%, with 72.5% gross margins and an adjusted EBITDA margin of 29%. That's software economics with consumer reach. Is it cheap? No. But it's cheaper than it's been in three years.

There's a scenario where Duolingo spent the last two years building a bunch of AI features that users don't care about, sacrificed profitability to chase vanity metrics like DAUs, and the Max tier never scales because people are perfectly happy with the free version.

In that case, you're buying a company that trades at 2.5x sales with decelerating growth and a CEO who thinks he's Steve Jobs when he's really Adam Neumann.

There's also a scenario where the strategy works. Where Max conversion improves, where AI features become table stakes, where new subjects (chess, math, music, whatever's next) expand the TAM by 10x, and where this becomes the default way people learn anything online.

In that world, $154 is stupid cheap and we'll look back at May 2025 the same way we look at Meta at $90 in 2022 - an obvious bottom that felt terrifying in real time.

I think the truth is somewhere in the middle. Duolingo's a really good business run by smart people who just made a strategic bet that pissed off short-term traders. That doesn't make them wrong. It makes them contrarian at exactly the time when contrarian bets are painful.

The stock might sit here at $150 for six months while management proves the thesis. Or it might bounce to $250 if Q4 shows Max subscriptions inflecting. Or it drops to $100 if the AI features flop and they're forced to flip back to aggressive monetization.

I don't know. And honestly, anyone telling you they know is lying or in Congress. But I can’t deny that there is potential.

I'm genuinely considering building a position here although I haven’t yet, and these are the things I'm tracking:

Max conversion: It's at 9% of paid subs now. If that hits 15% by Q2 2026, the thesis is working. If it's still stuck at 9-10%, they've got a problem.

Gross margins: Currently 72.5%, down slightly due to AI costs. If margins keep compressing, the AI features might be profitable but not profitable enough. Watch for stabilization above 70%.

China: If they cross 10% of DAUs from China without getting shut down, that's a green flag. If Beijing starts making noise, you need to be prepared to cut the position immediately.

Chess and non-language growth: If chess retention stays higher than languages and they successfully launch 2-3 more subjects, the platform thesis is real. If chess plateaus or fails, it was a gimmick.

Q4 2025 results: Coming in February. If they beat their conservative guidance and DAUs stay strong, the selloff was overdone. If they miss or guide even softer, this goes lower before it goes higher.

Duolingo at $154 is either a generational buying opportunity or a value trap masquerading as a growth stock. The business fundamentals are strong - thats 50 million daily users, 40%+ growth, cash-generative, massive brand moat. The valuation's reasonable if growth holds. The risk is that management chose to optimize for 2030 when Wall Street wanted 2026.

I'm not saying its a clear buy. I'm saying it deserves your attention.

Because when a stock drops 72% and the actual business is still growing 40%, something interesting is happening. Either the market's right and I'm missing something, or the market's wrong and there's money to be made.

Only time will tell, and I’ll be diving deep into DUOL with a follow-up article after earnings.

The Earnout Investor provides analysis and research but DOES NOT provide individual financial advice. Jamie Dejter may have a position in some of the stocks mentioned. All content is for informational purposes only. The Earnout Investor is not a registered investment, legal, or tax advisor, or a broker/dealer. Trading any asset involves risk and could result in significant capital losses. Please, do your own research before acquiring stocks.