Duolingo (DUOL) is a $1 billion-revenue education technology company headquartered in Pittsburgh, Pennsylvania. The Duolingo app is the most downloaded education app in the world, with 133 million monthly active users and 52.7 million daily active users as of December 31, 2025. The company carries zero debt, holds $1.275 billion in cash and investments, and generated $360 million in free cash flow in 2025. It is also down 82% from its all-time high, trading near $98 per share after management told investors in February 2026 that it would deliberately sacrifice near-term profitability to chase a goal of 100 million daily users by 2028.
How Duolingo Got Here
Luis von Ahn grew up in Guatemala. He went on to become a computer science professor at Carnegie Mellon University in Pittsburgh, where he invented CAPTCHA (those distorted text puzzles that prove you are human) and later sold reCAPTCHA to Google. In 2006, he was awarded a MacArthur Fellowship. His PhD student, Severin Hacker, shared a conviction that high-quality education should be universally accessible and free. In 2011, the two of them incorporated Duolingo, and the app launched publicly on June 19, 2012.
The core insight was simple: language learning was a massive global market where existing solutions were either expensive (Rosetta Stone, private tutoring, classroom instruction) or ineffective. By building a free, gamified app and running thousands of A/B tests to optimize engagement, Duolingo could amass an enormous user base, collect the world’s largest language-learning dataset, and monetize a fraction of those users through paid subscriptions.
For the first five years, the company focused almost entirely on growth. A paid subscription tier (now called Super Duolingo) launched in 2017. The Duolingo English Test (DET) launched in 2016 as a cheaper, online alternative to TOEFL and IELTS. The company went public on Nasdaq on July 28, 2021.
The years after the IPO brought rapid scaling. Revenue grew from $250 million in 2022 to $531 million in 2023 to $748 million in 2024 to $1.038 billion in 2025. The business became durably profitable. In 2025, Duolingo released the valuation allowance on its deferred tax assets, a $256.7 million one-time benefit that formally acknowledged the company no longer expected future losses.
What Duolingo Does
The Core App
Duolingo offers over 250 language courses in 41 languages. Every course is fully accessible for free. The app is built around game mechanics: streaks, leagues, XP points, animated characters, and a heart system that limits mistakes. These mechanics are not decorative. They are the product. Duolingo runs thousands of A/B tests per quarter to refine every element of the learning loop, and the result is stickiness that consumer apps rarely achieve. As of Q4 2025, 43 million users had active streaks of 7 or more days, and 15 million had streaks exceeding 365 days.
The company processes nearly 2 billion exercises per day, which it uses to continuously improve its adaptive learning algorithms. More users generate more data, which makes the courses better, which attracts more users. This flywheel has been running for over a decade.
How Money Gets Made
The revenue model is freemium with four streams:
| Revenue Stream | FY2025 | % of Total | YoY Growth |
|---|---|---|---|
| Subscriptions (Super + Max) | $873.4M | 84% | +44% |
| Advertising | $79.7M | 8% | +45% |
| Duolingo English Test | $42.0M | 4% | -8% |
| In-App Purchases | $40.5M | 4% | +5% |
| Total Revenue | $1,037.6M | 100% | +39% |
Super Duolingo removes ads, adds features like streak freezes, and is priced at various tiers depending on geography and plan length. Duolingo Max is a premium tier launched in 2023 that adds AI-powered features, most notably Video Call (a real-time conversation practice with an AI persona) and Explain My Answer. Max is priced higher than Super and carries higher gross costs due to AI inference.
The DET is accepted by 6,100+ programs worldwide, including all Ivy League schools, MIT, and Stanford. At $49 per test, it undercuts TOEFL and IELTS on both price and convenience. Revenue here declined 8% year-over-year in 2025, likely driven by shifts in U.S. immigration policy that reduced international student testing demand. The filing explicitly calls out U.S. immigration policy as a risk factor for this line of business.
Beyond Languages
Duolingo launched Math in 2022, Music in 2023, and Chess in 2025. The thesis is that the same gamification and data-driven pedagogy can work across subjects. In July 2025, Duolingo acquired Music Learning Services Ltd. (the team behind the NextBeat app in the UK) for $34.5 million, primarily for talent and IP to accelerate the Music course.
Where Users Are (and Who Gets Paid)
Revenue splits 38% U.S. and 62% international, a ratio that has been trending more international every year (it was 45% U.S. in 2023). The company carries no foreign currency hedges.
Nearly all of this revenue flows through intermediaries. The 10-K discloses that 62% of revenue comes through Apple’s App Store and 20% through Google Play. Combined, 82% of revenue passes through two companies that extract 15-30% of every in-app transaction. This platform fee is permanently embedded in Duolingo’s cost structure and is already baked into the 72% gross margin. Any change in platform fee policies, up or down, would have an outsized impact on profitability.
The Financial Picture
Three-Year Income Statement Summary
| Metric | FY2023 | FY2024 | FY2025 |
|---|---|---|---|
| Revenue | $531.1M | $748.0M | $1,037.6M |
| Gross Profit | $389.0M | $544.4M | $749.5M |
| Gross Margin | 73.3% | 72.8% | 72.2% |
| Operating Income | ($13.3M) | $62.6M | $135.6M |
| Operating Margin | (2.5%) | 8.4% | 13.1% |
| Adj. EBITDA | ~$82M | $191.9M | $305.9M |
| Adj. EBITDA Margin | ~15% | 25.7% | 29.5% |
| Free Cash Flow | ~$140M | $264.4M | $360.4M |
Duolingo went from operating losses in 2023 to a 13.1% operating margin in 2025. Revenue grew 39%, but total operating expenses only grew 27%, showing clear operating leverage. Gross margin is declining modestly (from 73.3% to 72.2%) as AI inference costs for features like Video Call increase. Management has acknowledged this trade-off: AI features drive higher-tier subscriptions, but they also carry higher per-user costs.
The gap between Adj. EBITDA ($305.9M) and GAAP operating income ($135.6M) is primarily SBC, which was $137.4 million in 2025, or 13.3% of revenue. This is the single biggest adjustment to track over time.
GAAP net income of $414.1M includes a $256.7M one-time tax benefit from releasing the deferred tax asset valuation allowance. Stripping that out, normalized net income was roughly $143 million.
Quarterly Revenue Progression
| Quarter | Revenue | YoY Growth |
|---|---|---|
| Q1 2025 | $230.7M | +38% |
| Q2 2025 | $252.3M | +41% |
| Q3 2025 | $271.7M | +41% |
| Q4 2025 | $282.9M | +35% |
Q4 growth decelerated to 35%, the slowest quarter of 2025. This was the first visible signal of slowing momentum and foreshadowed the guidance that would follow in February 2026.
Key User Metrics
| Metric | Q4 2024 | Q4 2025 | Growth |
|---|---|---|---|
| MAUs | 116.7M | 133.1M | +14% |
| DAUs | 40.5M | 52.7M | +30% |
| DAU/MAU Ratio | 34.7% | 39.6% | +490bps |
| Paid Subscribers | 9.5M | 12.2M | +28% |
The DAU/MAU ratio of 39.6% is strong for any consumer app. It means nearly 4 in 10 monthly users open the app on any given day. But DAU growth of 30% in Q4 was the slowest in four years.
Bookings vs. Revenue
Because most subscribers pay annually upfront, bookings lead revenue by several months. This makes bookings the better leading indicator:
| Metric | FY2024 | FY2025 | Growth |
|---|---|---|---|
| Subscription Bookings | $730.7M | $996.3M | +36% |
| Total Bookings | $870.6M | $1,158.4M | +33% |
Subscription bookings crossed $1 billion in 2025. Total bookings of $1.158 billion will flow into 2026 revenue as deferred revenue unwinds.
The Balance Sheet
| Item | Dec 31, 2025 |
|---|---|
| Cash & Equivalents | $1,036.4M |
| Short-term Investments | $104.1M |
| Long-term Investments | $135.1M |
| Total Cash/Investments | $1,275.6M |
| Total Debt | $0 |
| Deferred Revenue (current) | $496.2M |
| Goodwill | $35.3M |
| Total Stockholders’ Equity | $1,347.0M |
Zero debt. $1.275 billion in liquid assets. The $496 million in deferred revenue is money already collected from subscribers for future service periods. It converts to revenue over time rather than representing a financial obligation. This is one of the cleanest balance sheets in consumer technology.
Corporate Governance
Duolingo has a permanent dual-class share structure. Class A shares (publicly traded) carry 1 vote each. Class B shares carry 20 votes each. There is no time-based sunset provision. Class B shares convert to Class A automatically upon transfer to a third party, with limited exceptions for estate planning. As long as Luis von Ahn and Severin Hacker hold their shares, they retain 20:1 voting power indefinitely.
As of December 31, 2025, directors, officers, and 5%+ holders collectively controlled 76.3% of total voting power. No acquisition, merger, or major corporate action can happen without founder approval. For investors, this means aligning with management’s long-term vision is not optional. If you disagree with the pivot to 100 million DAUs, there is no activist path to change it.
The February 2026 Pivot
How It Started
On April 28, 2025, CEO Luis von Ahn posted an all-hands email on LinkedIn declaring that Duolingo would become an “AI-first” company. The memo stated that headcount would only increase “if a team cannot automate more of their work,” and that the company would begin applying AI in hiring and performance reviews. Von Ahn also wrote that Duolingo could not “wait until the technology is 100% perfect” and needed to “move with urgency and take occasional small hits on quality.”
The response was immediate and hostile. Subscription cancellation threats flooded the comments. Users accused the company of putting “people last.” The backlash spread across social media and was covered widely in tech press. Duolingo temporarily paused its signature edgy social media presence to manage the fallout.
Von Ahn walked it back over the following months. In interviews, he said the memo had been misunderstood and that he had not given enough context. He clarified that Duolingo has never laid off full-time employees and had no intention of doing so. The contractor situation was more nuanced: Duolingo had already reduced its contractor workforce by 10% in early 2024, and von Ahn acknowledged that the contractor headcount had always fluctuated with business needs. By August 2025, he told the New York Times the backlash had slightly dampened user growth, though he characterized overall performance as still strong.
Duolingo itself acknowledged the damage in its 10-K, disclosing that the memo “may have contributed to unfavorable publicity, adverse impacts on the Company’s brand and social media presence, and a deceleration in user growth.” MAU growth decelerated from 20% in Q3 2025 to 14% in Q4 2025. Seasonal patterns contributed to that deceleration, but the timing is consistent with the memo’s lingering effects.
How It’s Going
On February 26, 2026, alongside Q4 results, von Ahn published a shareholder letter announcing a fundamental shift in priorities. Duolingo would deliberately slow monetization growth to re-accelerate user growth.
Over the past two years, Duolingo had been optimizing heavily for monetization. Von Ahn acknowledged that the company had been “adding friction” to push free users toward paid tiers, and that this friction had started to drag on DAU growth. Only 10% of MAUs pay for a subscription. Von Ahn described the company as “under-monetized and over-monetized at the same time.”
The 2026 plan centers on three initiatives: teaching languages better, improving the free user experience, and investing in math, music, and chess. AI features like Video Call, previously restricted to the highest-priced Max tier, will be tested in the lower-cost Super tier as inference costs decline. The company also plans to explore non-friction monetization like avatar customizations and other in-app purchases.
The medium-term target is 100 million DAUs by 2028, roughly doubling from today’s 53 million. Management believes that reaching this scale creates a larger and more defensible business with higher total revenue and profits, even if per-user monetization is temporarily lower.
What the Guidance Says
| Metric | FY2025 Actual | FY2026 Guidance |
|---|---|---|
| Bookings Growth | +33% | +10-12% |
| Revenue Growth | +39% | +15-18% |
| Adj. EBITDA Margin | 29.5% | ~25% |
| DAU Growth | +30% | ~20% |
| FCF | $360M | $350M+ |
Bookings growth decelerates from 33% to 10-12%. Revenue growth drops from 39% to 15-18%. EBITDA margin compresses from 29.5% to 25%. R&D and sales and marketing spending will outpace revenue growth. Q1 2026 bookings growth is guided to 11%, with Q2 dropping another 3 points sequentially. Q2 will also be the trough for margins before improvement in the back half, with Q4 expected to be the highest-margin quarter.
The company still expects to generate $350 million or more in FCF in 2026, even through the investment year. That level of cash generation while deliberately compressing margins speaks to the strength of the underlying model.
The stock fell 22% in after-hours trading on February 26. From the all-time closing high of $540.68 in May 2025, the stock has declined 82%. Director James H. Shelton purchased 5,000 shares on the open market at $99.76 on March 3, 2026, days after the drop.
Valuation
The valuation below is built entirely from Duolingo’s SEC filings, management guidance, and first-principles assumptions. No analyst estimates or current stock price are used as inputs.
Normalized Earnings Power (FY2025)
| Metric | Amount |
|---|---|
| Operating Income | $135.6M |
| + Interest Income | $45.2M |
| Pre-tax Income | $180.8M |
| Tax (21%) | ($38.0M) |
| Normalized Net Income | $142.8M |
| Adj. EBITDA | $305.9M |
| Free Cash Flow | $360.4M |
GAAP net income of $414.1M includes a $256.7M one-time tax benefit from the deferred tax asset valuation allowance release. The $142.8M normalized figure strips this out.
Growth and Margin Assumptions
The 2026 pivot complicates projections. Management is intentionally compressing growth and margins this year. The base case assumes the pivot succeeds in re-accelerating user growth, with monetization recovering in 2027-2028:
| Year | Revenue | Revenue Growth | Adj. EBITDA Margin | FCF |
|---|---|---|---|---|
| 2025A | $1,038M | 39% | 29.5% | $360M |
| 2026E | $1,200M | 16% | 25% | $350M |
| 2027E | $1,440M | 20% | 27% | $420M |
| 2028E | $1,725M | 20% | 29% | $530M |
| 2029E | $2,000M | 16% | 30% | $640M |
| 2030E | $2,260M | 13% | 31% | $740M |
The base case assumes DAU reaches 85-90 million by 2028 (short of the 100 million target), monetization recovers starting in 2027, and operating leverage gradually returns as the investment year ends. FCF margin stays above 29% due to the deferred revenue float.
DCF Valuation
Discount rate: 10%. Terminal growth rate: 4% (justified by 133 million MAUs against an addressable market of 2+ billion potential learners globally).
| Component | Value |
|---|---|
| PV of FCF (2026-2030) | $1,990M |
| Terminal Value (4% growth, 10% discount) | $12,830M |
| PV of Terminal Value | $7,970M |
| Enterprise Value | $9,960M |
| + Net Cash | $1,275M |
| Equity Value | $11,235M |
| Diluted Shares | 48.3M |
| Per Share Value | $233 |
Multiple-Based Cross-Check
| Method | Multiple | EV Implied | Per Share |
|---|---|---|---|
| EV/2025 FCF | 25x | $9,010M | $213 |
| EV/2025 Adj. EBITDA | 30x | $9,177M | $216 |
| EV/2026E Revenue | 8x | $9,600M | $225 |
| EV/2025 Revenue | 10x | $10,376M | $241 |
A 25-30x multiple on free cash flow and EBITDA is reasonable for a high-growth, zero-debt consumer technology company with strong unit economics and a clear monetization runway. Even at the low end of these multiples, the implied per-share values cluster in the $213-$241 range.
Scenario Analysis
Bear Case: The pivot fails. DAU growth stays at 20% but the company cannot re-accelerate monetization. Revenue growth settles at 12-14% annually. Margins stay compressed at 25% as AI costs and marketing spend remain elevated. The company never reaches 80 million DAUs.
Base Case: The pivot partially succeeds. DAU reaches 85-90 million by 2028. Monetization recovers in 2027-2028 with new revenue streams (IAPs, avatar customization, broader Max features). Revenue growth re-accelerates to 18-20% before decelerating naturally. Margins recover to 28-30%.
Bull Case: 100 million DAUs by 2028. Management hits its target. Subject expansion into math, music, and chess broadens the TAM. Non-language revenue becomes a material contributor. Paid subscriber penetration rises from 9% to 12-15% of MAUs. Revenue could approach $2.5 billion by 2030 with 30%+ EBITDA margins.
| Scenario | Revenue 2028E | EBITDA Margin | Per Share |
|---|---|---|---|
| Bear | $1,500M | 25% | $130-$150 |
| Base | $1,725M | 29% | $220-$240 |
| Bull | $2,100M | 30% | $350+ |
The $400 million share buyback authorized post-earnings could retire 8-9% of diluted shares at current prices. If management executes aggressively in 2026, it would meaningfully increase per-share value and more than offset annual SBC dilution. The scenarios above do not assume any buyback.
If you found this helpful, consider supporting my work. This blog has no ads and no paywalled content, and reader support is how it stays that way. You can also subscribe to my mailing list to get new deep dives in your inbox.
What to Watch
| Catalyst | Why It Matters | Timeline |
|---|---|---|
| Q1 2026 earnings | First quarter under new strategy and new CFO. Bookings growth guided to 11%. Any deviation signals credibility. | May 2026 |
| DAU trajectory | Management guided to ~20% YoY DAU growth in 2026. Acceleration above 20% validates the pivot. Deceleration below 20% is a red flag. | Quarterly through 2026 |
| Video Call in Super tier | Testing Max-tier AI features in the lower-priced Super tier is the key monetization experiment. Uptake data will matter. | Mid-2026 |
| Q2 margin trough | Guided to ~3pt sequential margin decline in Q2. If the actual number is worse, expect another sell-off. | August 2026 |
| Buyback execution | $400M authorized with no timeline. Pace of execution signals management conviction. | Ongoing |
| App Store fee developments | Any change in Apple/Google platform fees directly impacts 82% of revenue. EU DMA enforcement is the catalyst to watch. | 2026-2027 |
| 2026 Proxy Statement | Will update share ownership, insider transactions, and governance details. First proxy with Munson as CFO. | Spring 2026 |
| DET recovery | DET revenue declined 8% in 2025. Continued declines signal structural erosion of the testing business. | Quarterly |
| 100M DAU progress | The medium-term goal. Credible progress requires DAUs in the 65-70M range by end of 2026. | Dec 2026 |
| Acquisitions | $1.275B in cash, zero debt, and a history of acqui-hires for design, animation, and music talent. The largest deal to date was $34.5M. The balance sheet can support something much bigger, and subject expansion into math, music, and chess creates a natural pipeline of acquisition targets to accelerate content and pedagogy. | Ongoing |
Sources:
- 10-K filed February 27, 2026
- 10-Q filed November 6, 2025
- 8-K filed February 26, 2026 (Q4 Earnings + Buyback)
- 8-K filed January 12, 2026 (CFO Change)
- 8-K filed December 9, 2024 (Board Addition)
- 8-K filed March 6, 2024 (Lease Expansion)
- “Duolingo’s CEO outlined his plan to become an ‘AI-first’ company. He didn’t expect the human backlash that followed,” Fortune, June 9, 2025
- “Duolingo CEO says controversial AI memo was misunderstood,” TechCrunch, August 17, 2025
- “Duolingo CEO clarifies layoff plans after AI memo controversy,” HR Grapevine, August 19, 2025
Research and analysis conducted with AI assistance using SEC EDGAR filings as primary sources.