Duolingo: Great Business At A Fair Value

Dominant business with great prospects trading at around fair value.

If somebody asked me characteristics of the most promising companies, I would say two things:

• Sticky product.

• Big addressable market.

The first point should be intuitive. Companies that don’t need to create a new customer base every year have an advantage over others.

If you can keep a substantial portion of your current customers next year, and charge them a bit higher prices, you will need to add way fewer new customers to grow your business.

This is why Apple is a great business.

Everybody thinks it’s selling consumer electronics, but it’s beyond that. It’s basically a subscription business.

Android users have too many options. This time, they may use Samsung, and next time they may use Google Pixel. However, IOS users only have one option—Apple.

IOS and the ecosystem around it are so strong that it literally locks in the users. This way, what would otherwise be a simple electronics business turns into a subscription business:

When you retain over 92% of your users, growth is effortless.

This is why I like insurance.

For a company like Apple, it’s a very exceptional case to retain 92% of your users year over year. Apple is an exception because it has built a cult around its brand and products over decades, and beyond that, it has really created the best in-class products and deserves its brand value and cult.

Apple is an exception. Insurance, on the other hand, is an intrinsically sticky business:

• You need insurance every year.

• Sticking with your existing provider is cheaper than switching.

This is how insurance companies can grow at mediocre rates for unusually long times.

Banking is another example. Most people tend to stick with their banks unless something unusual happens.

The second element is a big addressable market.

Apple, again, is an exception here.

This time, the exact opposite dynamic is at play as Apple’s business is intrinsically a global business. Technological devices tend to quickly diffuse to global markets, given that we are living in a highly globalized world. What you can sell in the US, you can also sell in India, Russia, Brazil, Germany, etc..

However, intrinsically sticky businesses tend to stay local.

Given that the delivery of the product or the provision of the service is regularly required, consumers demand easy access and close contact, which is best offered by local businesses.

This is why dominant insurance companies differ country by country, state by state, and even city by city.

Same as the banks. Most people tend to favor their local banks over national banks.

As a result, we haven’t had many sticky businesses whose addressable market was global rather than local.

This has changed with the internet.

In the early days of the internet, it was viable solely as an intermediary. The infrastructure wasn’t developed enough to provide services directly through the web platforms.

However, as the 3G network became global around 2007, the internet infrastructure and wifi speed reached good enough levels to allow the provision of services directly through web-based platforms.

This allowed basic subscription businesses to go global.

When a sticky business meets with a global market, it means a decade of runway for fast growth, as it’ll keep existing users, raise prices for them, and will be able to add new users for years, as the market will be too big to quickly saturate.

Investors woke up to this with Netflix:

The payoff of this model tends to be immense.

Netflix is one of the best-performing stocks in market history as it has returned more than 500x since it fully pivoted to streaming in 2007.

Later, Spotify also confirmed that this model works. It’s also up by 5x since it became public in 2018.

As a result, subscription business addressing a global market has become the dream of investors.

This is the reason for the obsession with Duolingo.

It’s a subscription business in a global market. Education, especially language learning, is a global market, and it requires at least a year of commitment to learn any language to a satisfactory degree. So investors expect it to replicate what Netflix has done over the long term.

The premise is correct, yet such businesses need two things to deliver that strong long-term performance:

• It should be able to take advantage of the global market for at least a decade.

• The purchase price should be attractive.

Remember, Netflix completed its pivot to the streaming model in 2007, and the broader public noticed only around 2020 how incredible its stock had performed over the previous decade.

Second, these businesses are rare, and all the smart investors have been looking for the next Netflix and Spotify since we noticed the success formula there. Thus, they are almost always expensive, shrinking the return you could expect.

As a result, to succeed in such investments, the company should be able to take advantage of the global market for decades, and you should buy it at the right price.

Today, we will try to understand whether Duolingo has what it takes for your investment in the company to enjoy above-average returns for a decade.

So, let’s cut the intro and dive deep into this exceptional business!

What are you going to read:

1. 🏭 Understanding The Business

2. 🏰 Competitive Analysis

3. 📝 Investment Thesis

4. 📊 Fundamental Analysis

5. 📈 Valuation

6. 🏁 Conclusion

---

🏭 Understanding the Business

I hear hundreds of stock pitches and read hundreds of write-ups every year.

Here is the shocking thing— most of them fail at the first step, which is understanding the business.

It’s interesting because this looks like the first and simplest step. Look at the business of the company, understand what service or product it produces, etc.. Then there are harder parts like competitive analysis, valuation, etc..

Yet, most people don’t understand the business.

Start reading a random write-up on Substack and you’ll often find the writer describing the product that the business creates, how it creates it, etc..

It’s like going into a coffee shop and saying that it sells coffee and croissants.

Is this the business? No, a coffee shop is not in the business of selling coffee; it’s in the hospitality business.

What coffee shops sell is the experience, providing a place for people to socialize, talk to their friends, or read their books while drinking their coffee, without having to be isolated in their homes.

What does this change? It changes everything. It changes how you evaluate the business.

If you thought it was selling coffee and croissants, you would look at their quality and price. If you understood that the business was hospitality, you would look at the locations, designs of the shops, and how much time customers spend there as the things that differentiate the business, not just price and product.

Similarly, if you think that Duolingo is in the business of teaching language, you’ll completely get the business model wrong.

Learning is just a process; the crux of learning is engagement, until the expert level.

Think of it like coffee shops where coffee and bakery are commoditized, and customer preference is largely based on factors like concept, workers’ warmheartedness, etc..

In education, the knowledge itself is a commodity until the expert level. What determines the customer outcome is the provider’s ability to keep the audience engaged.

Duolingo understood this, and from day 1, they based their product on the “hooked model.”

What the hell is the “hooked model?”

It’s the model behind addictive products like Instagram, YouTube, etc. The term was coined by behavioral scientist Nir Eyal.

According to Eyal, the hooked model has four steps:

• Trigger: What prompts the user to take action. It could be external, like a push notification, or internal, like boredom.

• Action: The step the user has to take to reach the reward. The easier it is, the more people do it.

• Variable Reward: User gets some type of reward after the action, so he doesn’t feel like it was for nothing.

• Investment: The user puts in some kind of resource, like time, effort, money, etc. This investment makes the process more valuable for him, increasing the chances of repetition.

Duolingo carefully analyzed the habit-forming products like Instagram and YouTube and engineered their addictive properties into a learning app.

It starts with the trigger.

Users regularly get push notifications that joyfully invite users to open the app and complete a lesson. These aren’t random notifications. Duolingo closely tracks your personal patterns and sends those notifications at times you are most likely to follow.

It makes the action as effortless as possible, so for the user, it’s easier to tap on the notification and complete a lesson rather than carrying the mental burden of inconsistency.

The lessons always start with very simple tasks, so users rarely bounce after the lesson starts. The completion is also effortless. It doesn’t expose you to the totality of the new subject, which takes a high mental effort. It teaches in very small increments, so it feels effortless.

It basically shifts the essence of language learning from a high effort activity to a high consistency activity.

When you complete an action, you get a reward that instantly makes you feel better. It’s not a material reward, but it’s a reward of the self. It gives you a badge, a streak score, or upgrades you to a higher league, so you immediately know that your preference to keep it consistent has a consequence.

When you complete all these steps, the investment element follows automatically. You know that you have put some time and effort to learn something new, and you achieved that even though it was immaterial. If you come back tomorrow, you’ll learn something more, and the more you do this, the more exponential the reward will become. You are now invested; you have more to gain by coming again tomorrow than you gained today.

And just like that, they have created the highest retention app in the market.

As you see, while users of other language learning apps like Babbel have a very high tendency to try other apps, Duolingo users almost never bounce off to try other apps.

They didn’t create a learning app; they created an engagement engine.

This allows them to monetize their users in multiple ways, not just through subscriptions. For the context, the business generated $271.7 million in revenue last quarter, of which 16% came from sources other than subscriptions:

In sum, Duolingo is not a basic education platform; they know that knowledge is a commodity, and keeping the audience engaged is the real product.

They have designed their whole platform and the experience to maximize engagement and retention. As their retention rate is disproportionately high compared to its competitors, it needs to add way fewer new users to generate growth. If they can do this for a decade, they’ll certainly become a giant business, like Netflix and Spotify.

Thus, the central question is simple:

Can they keep doing this for a decade, given that AI-powered technologies can make language learning redundant or give rise to new ways of learning language?

I am confident that they can.

---

🏰 Competitive Analysis

Most people think that the business must have a competitive advantage because this is what makes it predictable.

If it’s a predictable one, we can value it and compare it with the current market price. If the market price is lower than the fair value we calculated, we invest; otherwise, we avoid. This is why competitive advantage is a must to practice investing dependably.

That’s true, but let me bring some perspective to it.

I think competitive advantage is a must because it makes the business predictable in the short term but unpredictable in the long term.

Well, what does this mean?

Competitive advantage allows investors to more dependably predict the near future. After all, there are two main competitive risks:

• Competition from other players.

• Technological disruption.

The competitive profile of an industry doesn’t change much in short periods. Thus, competition from other players rarely changes the outlook for the dominant players. Same for the technological disruption. The shorter you look into the future, the less likely a technological disruption is to occur.

This creates a significant disadvantage for investors.

Analysts on Wall Street generally use standard 10 or 5-year DCF models to value companies. Given that they are so focused on generating alpha today, most of them use 5-year models and come up with revenue and EPS estimates for the next 5 years.

As a result, businesses that are easy to predict due to their competitive advantages, usually get fairly valued with their cash flows for the next 5 years perfectly priced in.

This means that, unless something like a market crash or recession occurs, they are often perfectly priced for the short term, leaving no room for investors to get above market returns.

However, when we look beyond the next 5-10 years, durable competitive advantages make the businesses extremely unpredictable, which is why investing in exceptional businesses for the long term works so well.

It works because exceptional businesses with a durable competitive advantage tend to surprise on the upside in the long term, which is unduly discounted in the DCF models focused on the short term.

Here is an example.

Warren Buffett bought American Express first time in 1964.

At the time, the stock was trading at $35 per share, which would equate to $2.25 given that the stock has gone through cumulative 1-to-15.53 splits since then. The company had generated $118 million in revenue that year with a 13.4% operating margin and 14% return on capital.

If Buffett did a 10-year DCF model at the time, it would have looked like this:

He would have forecasted a 4.40% terminal growth rate, which was the rate on AAA-corporate bond at the time, and he would get a fair value of $81, or $5.22 split-adjusted, per share, which was already very attractive given that the stock was trading at $2.25 per share on split adjusted basis.

Now, here is the fascinating thing:

Such a DCF would already be slightly off the mark, as American Express grew revenues on average 10.6% from 1964 to 1974. But the more interesting thing is that the company kept growing revenues at a 10.1% CAGR in the next 50 years, up until 2024, while this period was accounted for just 4.4% growth in the DCF model.

Thus, competitive advantage should necessarily come with long-term durability for investors to benefit from it, as most of the value comes from the post-prediction period, where an exceptional business consistently does better than the terminal growth forecast.

I have explained all these because I see that there is already a consensus on the fact that Duolingo is an exceptional business. However, this is not enough. For investors to benefit from a competitive advantage, long-term durability should be there too; otherwise competitive advantage just makes analysts’ jobs easier for the near future and the market more efficient.

So, let’s look at the competitive advantage and durability of Duolingo separately.

When we look at the competitive advantage narrowly, i.e, in terms of how Duolingo is positioned compared to its competitors, it’s obvious that its business dynamics are much stronger and any existing or potential competitor is unlikely to match it.

To start with, it has reached a scale unprecedented in learning apps.

Duolingo currently has over 135 million monthly active users (MAUs). This is unprecedented.

For reference, Spotify has 713 million monthly active users, even though basically the entire planet is in its addressable market. Education is a far smaller market than music, yet Duolingo has still managed to reach a user base that’s one-fifth the size of Spotify’s.

This scale creates massive data-network effects:

The more users Duolingo has, the more data it accumulates. It uses this proprietary data to improve the product, which attracts even more users.

When this is combined with its amazing retention rate, the contestable market for the competitors shrinks substantially. Even if they improve their services and copy Duolingo’s properties, they’ll have to actively take users from Duolingo, which is really hard to achieve given Duolingo’s high retention. Thus, it’s unlikely that Duolingo will suffer from industry competition anytime soon.

What is more consequential for me is the durability of the business.

Since its industry position is rock solid, I think the stock is very well assessed and priced for the short term. Thus, I believe most of the gains will come from beyond the prediction period, just as you would experience if you have held American Express since 1964.

Don’t get me wrong, you don’t need to wait for 5 years or 10 years to get above-average returns. As we go forward, and the business’s growth proves persistent, a portion of the slow growth stage in the terminal period shifts to the prediction period as a higher growth era, and the market bumps up the price.

The example is that if you revisited the American Express DCF above in 1970, you would still start from projecting 10% growth for 1971, converging to 4.5% in 1980. So, compared to your DCF in 1964, 6 years of a low growth period in the terminal stage would have shifted to your new DCF in 1970 as a high growth period. The market would have gradually bumped up the price as this happened, so you would have gotten above-average returns.

So, for Duolingo, the real question is whether it’ll be disrupted in the next, let’s say, two decades, or it’ll prove durable.

I think it’ll prove durable.

The biggest threat to Duolingo is obviously AI. There are three apparent threats here:

• People will use AI models to learn language.

• New AI-based applications will emerge and take on Duolingo.

• Language learning will be redundant as AI will enable instant cross-language communication.

The first one is the most obvious threat on the horizon and also the one I believe is least likely.

As I said when describing the business, the value isn’t the knowledge, it’s the process. I don’t think a significant number of people will be able to consistently engage with AI without any roadmap and learn a language. The key is retention, not the ability to access knowledge.

This is nothing new. I think we have already tested this theory and saw that it doesn’t work. Take my personal background as an example:

• I went to a law school.

• Worked as an equity analyst.

• Currently doing PHD in Competition Economics.

I assure you, anybody can learn law, security analysis, and economics solely by reading. It literally doesn’t take anything else. How many of the interested people do you think are currently achieving this? I would say not too many.

Think about it, this is not the first time the education market has faced such disruption. I am 100% sure people thought tutors would be redundant when the printing press became mainstream. I am sure some people thought the same when the Internet happened. I think people thought print books would die when e-books emerged. None happened. Because what matters is the process. Reading a print book is a completely different process and pleasure than reading an e-book.

Source: Amazon Sales Statistics

That applies to Duolingo itself, too. Do you think enrollments in offline language courses declined because of Duolingo? Hell no. QS estimates that this market has grown from $49 billion in 2018 to $68 billion now.

In short, I don’t believe AI will replace the process that Duolingo offers.

Second, if AI enables us to create new platforms, incumbents will be the best-positioned companies to take advantage.

If there ever will be a more effective AI-powered way of learning a language, Duolingo is the best-positioned company to exploit it, as it can leverage its existing customer base to dominate the new product market.

This is similar to Google’s comeback in AI after they were baffled by the rapid accession of the market. Since they started to take it seriously, they have been leveraging their existing users and taking market share from other players.

Just as Google is the best-positioned company to win in AI, Duolingo is the best-positioned player to win if there ever is a new product market for language learning with AI.

The third scenario of disruption is actually ridiculous.

It’s like arguing that people will stop learning languages because there are professional interpreters now.

No, people learn language primarily for two reasons:

• Hobby

• Professional or social necessity

Hobbies are disruption-proof. We don’t even need to discuss it. Machines are capable of creating every furniture now, but still, there are people learning how to work the wood.

Professional communication is also disruption-proof. If some business or society has a lingua franca, you are more valuable if you can communicate directly in that language. The value isn’t whether you can communicate or not, but it’s whether you can do it personally.

Imagine going into a meeting with people where nobody knows the lingua franca. They all rely on AirPods 3 with immediate interpretation. Can you imagine somebody saying in Chinese, “Hey, sorry, I am off, my Airpods just died, I don’t understand a dime.” And nobody understands him as well because everybody is relying on AirPods?

It’ll be ridiculous. I assure you, if society ever gets to that point, investing in anything won’t work. There’ll probably be nothing like investing as well.

Summing up, I think Duolingo is an amazing business that is dominating its market, and it isn’t subject to serious competition from other players. Most of the concern is with the long-term durability. I think it’s there too.

I am not saying that Duolingo will be able to stay largely the same for decades as American Express did, but I simply cannot see a future where something new comes up that Duolingo can’t capitalize on and lose the market. I don’t think it’ll happen.

---

📝 Investment Thesis

My Duolingo investment thesis relies on three pillars:

1️⃣ There is a huge runway for growth.

Language learning is a giant market that currently sits at $115 billion.

The market has grown, on average, 11% annually since 2019. However, this growth has been largely driven by online learning. While offline language learning grew around 6% annually since 2019, the online learning market size roughly quadrupled:

QS Education

QS Education estimates that 1 out of every 4 people in the world is actively trying to learn new languages, which means roughly 1.8 billion people.

This is a huge market.

Though it’s the leader in online language learning, Duolingo has just scratched the surface of this demand. The company is on track to generate around $1 billion in revenue this year, which means it owns just around 2% of the market.

Given that the younger generations are increasingly mobile as a result of elevated cultural mobility, this demand will keep increasing, and Duolingo is well-positioned to expand its market share. Increasing demand, combined with growing market share, will create dual tailwinds of growth for Duolingo.

2️⃣ There are ample opportunities to expand the addressable market.

Language learning is uniquely suited to Duolingo’s business model, as it can be easily fragmented into dozens of milestones and key learnings, yet it’s not the only thing that can fit in this model. There are dozens of other subjects that can be taught in the Duolingo way.

It has already expanded into a few of these markets, like math, chess, and music. This model of expansion has already proved itself as chess became the fastest-growing vertical last year, passing all the new language courses.

The opportunity here is literally unlimited:

They can easily enter verticals like board games other than chess, brain training, and expand into classical subjects like physics and chemistry. I think all these subjects can be divided into hundreds of small milestones to be taught in the Duolingo way.

It’s literally impossible to quantify this broader market opportunity, but it would suffice to say that Duolingo hasn’t even scratched the surface of this opportunity.

3️⃣ AI will massively boost efficiency.

In 2012, Duolingo made its public launch with just 5 courses. In the next 10 years, they added on average 15 courses a year, and completed 2024 with around 155 courses.

Here is the mind-blowing thing—they literally doubled the number of courses this year:

They added a mind-blowing number of 148 courses this year. They were able to achieve this as they integrated AI into their curriculum development process, which made it 10x faster.

This isn’t the only contribution of AI to their business. They have been able to rapidly improve the weakest part of their business using AI: Speaking lessons.

Up until AI voice models emerged, Duolingo relied on ML-driven voice recognition models that were programmed to recognize particular words or sentences. This was a significant limit on the speaking exercises that could be done on the app. AI voice models completely removed this limitation.

They added an AI video call feature that lets you have longer, more natural conversations where you can go beyond repeating the words and sentences you are expected to say. Users praised this as the biggest improvement ever in Duolingo.

In sum, I think Duolingo has massive opportunities going forward. The language learning market is already huge and growing. Duolingo is expanding beyond this market, adding subjects like chess and math, etc. AI allows Duolingo not just to improve the current business, but also to get faster in exploiting new opportunities.

Regardless of whatever angle you take, there is an abundance of opportunities for Duolingo.

---

📊 Fundamental Analysis

➡️ Business Performance

Duolingo’s performance chart is the dream of investors.

It grew revenues by 5x from $161 million in 2020 to $964 million in the last twelve months and turned profitable doing this:

The business literally annihilated all the key metrics in this five-year period:

• Paid subscribers grew from 1.6 million to 11.5 million.

• Paid users as a % of MAUs grew from 4% to 8.5%.

• Free cash flow margin expanded from 8% to 36%.

That’s literally a kind of performance I haven’t seen in a long time from any business. What generally happens is that you get some nice elements at the expense of some other desirable elements, like profitability increases, but the growth slows down.

That’s not the case for Duolingo. All its key metrics have decisively gone up.

Overall, I don’t think there is much to talk about the performance of this business. I can only say that it’s rare to see a business that is both growing and obviously improving its quality and value for the users decisively at the same time.

I think Duolingo has achieved this.

➡️ Financial Position

Duolingo has what I call the “golden standard” when it comes to its balance sheet.

To start with, it’s simple.

No special purpose vehicles, hidden debt, risk of markdowns, or you don’t need to disentangle the cash flow statement to find out its real burn rate and compare it to cash in the balance sheet, etc. If you remember, we have had to navigate complex financial structures in the last two weeks as I have analyzed Pagaya and Soluna. Luckily, we don’t have that complexity here.

It simply has mountains of equity, minimal debt, and EBITDA that can pay the whole debt under a year. This is the golden standard of the balance sheet:

It currently has over $1.3 billion in equity against just $93 million of debt. Its annual EBITDA of $116 million can pay the debt in under a year. It’s a simple, rock-solid balance sheet built to weather any storm in the future.

Beyond the numbers, the simplicity of the balance sheet also signals confidence in business operations. When the balance sheet governance is too complex, it often signals to me that the management relies on pulling rats out of a hat to survive and thrive. When the balance sheet is simple, the message is clear—the business is so strong that it doesn’t need to complicate the finances.

Duolingo’s business is so strong that it doesn’t need to complicate finances. A+ grade.

➡️ Profitability & Capital Allocation

• Gross and Net Margins

The margin profile of the business signals nothing but strength and dominance.

Just look at this:

Its gross margin is almost like a flat line, which signals that it hasn’t felt any pricing pressure in the last five years, meaning its competitive position is rock solid. No competitor has been able to meaningfully challenge it, and I don’t think this will change anytime soon.

Net margins have also been consistently expanding since 2021, signaling operating efficiency. It looks like the net margin has spiked to 40% but this is due to $227 million one-time tax benefit it reported this year. If it wasn’t for that, net margin would be around 15% which is still a meaningful expansion from 2024 levels.

In short, its margins are signaling two things: A giant competitive moat and increasing efficiency.

• Return on Invested Capital (ROIC)

This is one of the things that makes me very bullish on Duolingo.

Its ROIC has consistently improved in the last three years and finally went above the average for all mature American companies, which is 12%:

Here is the bullish thing: ROIC tends to increase while the business goes from a young growth company to a mature company and peaks at maturity. From there on, it tends to decline as reinvestment opportunities dry up.

Duol has just reached the average ROIC for American companies, but it’s still relatively early in its growth phase. Thus, from now on, ROIC will likely increase until maturity. Thus, shareholders will retain more of the value created as it grows the business and captures new opportunities.

In sum, Duolingo’s fundamentals deserve the word “perfect.” It’s growing amazingly fast with expanding margins and rock rock-solid balance sheet. Its return on invested capital is poised to increase from here, allowing shareholders to capture more of the value going forward.

Just perfect.

Yet, one question still remains—what about the price?

Let’s dig.

---

📈 Valuation

The way I think about the valuation is that you understand the business and the market and create a future scenario for that business. Then, you find the current value of that scenario.

That’s basically it.

This is why, as opposed to what most people think, the valuation is not an objective construct; it’s subjective. Thus, people who have a better understanding of the business will inevitably do better valuations.

When it comes to practicing valuation, I strongly recommend everybody to keep that future scenario as conservative as possible within the plausibility limits. This will increase the discount you demand today, but will also increase the surprises to the upside if you can buy when the model signals undervaluation despite conservative assumptions.

So, let’s create such a conservative, but plausible scenario for Duolingo, and let’s see what we get:

It has grown revenues by 42% annually in the last 5 years. I’ll be conservative and assume that it’ll grow revenues 25% annually in the next 5 years. This will give us ~$3 billion in revenue in 2030.

Netflix, as the largest and most mature of the global consumer-facing digital subscription models, has a 25% net margin. Duolingo doesn’t have the production and royalty costs of Netflix, so it can easily reach 35% net margin.

This will give us $1.05 billion in net income for FY 2030.

Though subscription models are valued at a higher premium due to the consistency of cash flows and easily get 30x exit multiples, I’ll be very conservative and assume just a 20x exit multiple.

In which case, we’ll get a $21 billion company.

The business has been pretty conservative in issuing new shares for the past two years as shares outstanding grew less than 2% annually. In any case, I’ll again be conservative here and assume around 4% annual dilution. In this case, we’ll have 58 million shares outstanding, implying $362 per share stock price.

Discounting it back to the current day at a 10% annual rate, we’ll get a fair value of $224 per share.

The stock is currently trading at $200 per share, implying 10% undervaluation.

Given that +/-10% could be thought within the fair value range, I think it makes sense to think that the business is trading at fair value.

---

🏁 Conclusion

Duolingo is a great business, plain and simple.

I have seen very few businesses dominating their market to the degree that Duolingo does. It literally has no competition as it has managed to achieve retention and monetization levels that the competitors in the market are nowhere near of. It’s in its own league.

As opposed to what many people see, I don’t think it’ll get disrupted by AI. It’s in the engagement and process business, and I don’t see simple LLMs matching Duolingo's level of engagement and process expertise anytime soon. Further, if AI really creates a breakthrough in language learning, I think Duolingo is the best-positioned business to adapt and exploit the new methods.

Ample growth opportunities also exist, both in its core market and in other products. Plus, AI will likely be a major efficiency driver going forward, as it’s already proved this, enabling Duolingo to add 148 courses this year.

When it comes to valuation, I would say that I acted pretty conservatively, maybe too conservatively. You can expect it to grow more than 25% annually in the next 5 years or have a higher than 20x exit multiple, or less than 4% annual dilution, etc.

Still, I think that level of conservatism is warranted. After all, these scenarios don’t include execution risks, and the future always comes with a degree of uncertainty. It’s hard to incorporate these things into a valuation model. Instead, what you could do is act more conservatively with your assumptions so that such risks, to some degree, are factored in.

After going through all these, it’s obvious that the stock is undervalued; however, it’s not as undervalued as permabulls make it to be. It’s attractive, but it’s not a “no-brainer.” Below $170 levels, your margin of safety will be around 30% and it’ll truly be a no-brainer.

There are two paths you can take here, I think:

• Take a small position here and wait for below $170 to double down.

• Leave it and wait patiently until it gets below $170.

The former is preferred for those who see themselves as net buyers of equities over time and have a long-term orientation. The latter may never happen, but it won’t be an unreasonable move even if you miss the opportunity, given that the current market conditions are increasingly risky.

I hope this helps.

---

That’s all friends!

Thanks for reading Capitalist-Letters!

Please share your thoughts in the comments.

👋🏽👋🏽See you in the next issue!