Why Did Duolingo Stock Drop? Inside the Q3 2025 Earnings Call and the Company’s Big AI Pivot

Why Did Duolingo Stock Drop?

Duolingo (DUOL), the company behind the world’s most popular language-learning app, just reported what its CEO called a strong Q3 with solid performance across all metrics. The company is on track for nearly $1.2 billion in bookings for the year and boasts over 50 million people using the app every day.

So why did the market react so negatively, with the stock plunging almost 30% in the past week?

The answer isn’t in a disastrous quarter but a reduction in forward guidance in bookings. This is due to a deliberate, massive, long-term strategic pivot that worried investors . Duolingo’s leadership announced they are intentionally shifting focus away from near-term monetization to chase a much, much bigger prize: using AI to fundamentally rebuild education, aiming for billions of users. Those are lofty goals but the market doesn’t seem to be buying it yet.

Here’s a deep dive into what went right, what (maybe intentionally) went wrong, and what Duolingo’s management is betting the entire company on.

📈 The Good News: A Solid Q3 and Chess Is a Hit

Before the big reveal, the company’s Q3 2025 performance was strong. CEO Luis von Ahn highlighted the bright spots.

• Massive User Base: More than 50 million people now use Duolingo daily with 135M users monthly.
• Strong Growth: The company posted 36% year-over-year growth for daily active users (DAUs) in Q3. That drove 41% revenue growth and 68% adjusted EBITDA growth. Net Income was up substantially but it was driven by a one-time tax benefit.
• Financial Health: The company is guiding to 1.03B in revenue for the year, representing ~37.5% growth, with a hefty 29% adjusted EBITDA margin and bookings are set to be up 32.5% or so. Not many companies that claim that type of growth.
• Global Wins: DAU growth was helped by new partnerships, including one(Luckin Coffee) that helped them grow their brand in Asia quite a bit.

The biggest product win, however, is a complete surprise: Chess.

Analysts on the call were eager to hear about the new chess course, and von Ahn was very excited to talk about it.

“It is the fastest-growing course that we had,” he said. “It’s growing much faster than math and music and faster than the way originally languages grew”.

The new course already has millions of users, and early data shows its retention is slightly higher than language learning. A new Player vs. Player (PVP) mode is currently rolling out, which will soon allow every Duolingo user to play chess against each other. That’s an interesting development and shows that Duolingo doesn’t just have to be a language app but can use its substantial user base to roll out new products and get them going pretty quickly.

📉 The Bad News: Why Guidance Is Decelerating

This is the news that shook investors. The company’s Q4 forecast implies a deceleration in bookings growth. Management was extremely clear: this is not an accident; it’s a choice.

Von Ahn explained that the company is at a unique point in time. “We believe AI will fundamentally transform education,” he stated.

Because of this, Duolingo is making a durable shift in strategy. The company is investing in the long term by shifting more relative investment into two areas: user growth and teaching better.

This investment comes at a direct, short-term cost. As one analyst phrased it, the company is prioritizing user growth instead of monetization.

Von Ahn gave a clear, fictitious example to explain the trade-off:

• Free users currently get 25 “energy units” per day.
• An experiment could lower that to 24 units.
• Von Ahn admitted, “We know that would make us more money” because more users would run out and pay to subscribe.
• However, he also said, “we also know that would decrease daily active users because it would frustrate some of the users”.

In the past, Duolingo might have made that trade. The new Duolingo will not. Still, these various A/B tests that they run have made the app less user friendly. In the past, you could easily complete many lessons with the heart system. Now, the energy system makes it harder to complete more than two or three lessons. That’s better for monetization and driving people to paid versions but less good for the overall user experience.

“What we mean is that… we are going to be prioritizing user growth over monetization in this type of judgment call,” von Ahn declared. “We would not launch that experiment… even if it meant quite a bit of bookings gains, if it has a real hit on daily active users”.

🤖 The Goal: From 100 Million Users to Billions

Why make such a painful trade-off? Because Duolingo’s leadership no longer sees itself as just a language app. They believe AI gives them a line of sight to create a product that teaches better than ever before.

“We have line of sight now to create an app that can teach really, really well, much better than anything that humanity has seen before,” von Ahn said. “[As] good as a human tutor, but one that is also more engaging”. I think this certainly sounds amazing on paper but I’m not so sure it will work well in practice. AI can have a tendency to do things poorly at times and language learning is an item that’s very dependent on accuracy.

If they can pull that off, the upside is astronomical. “If we’re able to do that… we will be talking about billions of users that we have and that’s what we want to shoot for here,” he said.

That’s all well and good but if AI is so great then the other side of the coin is obvious. Why Duolingo where you can potentially get the same and more on one of the many super well funded generative AI models. Does that mean that Duolingo is suddenly competing with much richer players like OpenAI, Google and Microsoft or can it still do well as a standalone app because of its built in user base.

🧐 Analyst Q&A: The Good, The Bad, and The Unhinged

The analyst question-and-answer session was a masterclass in trying to understand this new philosophy. Here are the highlights.

1. On DAU Growth and Unhinged Posts

One analyst asked what drove the 36% DAU growth, but also what hurt it. Von Ahn gave a stunningly candid answer: Duolingo’s social media team got too corporate.

They paused what they call  their unhinged posts(chaotic, meme-based, often nonsensical, etc.) in their social media for a bit to try to build better brand love.  And when we don’t post unhinged things… our posts were much less likely to go viral, and because of that, that did have an impact on DAU growth. It does seem like that type of post resonates better with audience who may download the app although it does tend to make the brand seem less serious which is a trade-off.

The solution? Start them back up again and they’re already seeing positive engagement results from that marketing spend.

2. On the Duolingo Max AI Tier

Analysts were curious about the performance of Duolingo Max, the premium AI-powered subscription. The answer was mixed.

• The Good: Max now accounts for 9% of subscribers, and its bookings doubled in Q3 year-over-year.
• The Bad: “It’s underperforming our lofty expectations for it,” CFO Matt Skaruppa admitted.

• On ChatGPT: He argued Duolingo’s key strength isn’t just content; it’s engagement. “In order to learn a language, you need to be engaged for years,” he said. ChatGPT doesn’t provide the daily, gamified curriculum to keep users motivated for that long, at least it doesn’t yet. It wouldn’t be too hard to replicate what Duolingo is doing given that they have a larger user base already. It’s certainly a real worry and one that likely has driven some of the sell-off in the last year.
• On Translators: Also not worried about that apparently. He pointed out that “simultaneous language translation” has been demoed for 10 years, and during that time, the desire to learn a language has not gone down at all; in fact, it has come up. I do think that until those type of translators can basically be real time(think something like a piece of eye wear or an ear piece that everyone wears) that automatically translates things you hear into whatever language you want with good accuracy. I don’t see that happening anytime soon but generative AI certainly makes it a bit easier than it was before.

• Geographically: “The fastest-growing region as a region is Asia for us,” von Ahn said. China is the fastest-growing country and now Duolingo’s second-largest market in terms of daily users. Duolingo Max is being tested at the moment in China. I do wonder what USA growth looks like because generally as a market, it generates the highest ARPU(average revenue per user) and that number drops off drastically in other countries.
• Content: The new strategy will accelerate content for the top 9 languages. The goal is to take users from zero to a Duolingo score 130, which is where you can get a job in that language.

The Long View: A Risk Reward That Seems Interesting Again

Duolingo is a rare company: highly profitable, growing fast, and yet willing to hit the brakes on its own financial momentum for a long-term vision.

Management isn’t worried. They are guiding to a 29% adjusted EBITDA margin for the year, which is close to their long term margin range and right now they are not afraid to invest in what is essentially user acquisition. While management may paint that as the right move long term, it does gently push the brakes on a growth story that has driven this stock’s price appreciation for a while.

While 20%+ bookings growth for Q4 isn’t terrible, it is a sizable drop from the front of the year and likely spells similar results headed into 2026 which certainly spooks investors. The question then becomes, how long is this investment period, are the users they’re acquiring primarily from lower APRU countries and is AI a real existential threat they’re trying to head off by building as big a user base as possible to test and improve their AI tools upon. On top of that, for subscribers, Duolingo is not a service that screams value and can easily be one of the first things they cut if money is tight. That doesn’t seem to be occurring yet but it’s always a worry.

Maybe investing in the future right now is the right move but this leads to slowing growth which is the real reason for why the stock is down 37% in the past year and 60% off its all-time highs it reached in May of this year. If this continues past this quarter, there could be some more pain ahead as the market is still pricing 20%+ growth for next year.

The story of Duolingo’s Q3 2025 earnings isn’t really one of failure. It’s the story of a company making a big, calculated bet. They are sacrificing near-term bookings to build an AI-powered human tutor for billions of people.

The valuation after this recent drop has certainly gotten more interesting. While a relatively new company this is one that’s already generating free-cash flow and still growing its users at a good clip. They’re sitting on $1B in cash without any debt(so a buyback is possible) and based on 26 estimates, are trading at a 4% free cash flow yield which is pretty cheap for a company growing their top line above 30% even if that number is slowing. Without SBC(stock-based compensation), that FCF yield is closer to 3% but hopefully they start a buyback program soon to at least offset the dilution from their SBC program.

As the CFO put it, “the risk reward seems right”. Investors, for now, are focused on the risk. Duolingo’s management is focused on the reward. I’m not an owner of the stock right now but at this valuation, it might be worth a closer look.

Disclosure : This is not investment advice. Please talk to a qualified financial advisor before making any investment decisions.