EdTech Stocks: Duolingo, Coursera, and the $400B Education Market

The $400B Market Nobody on Wall Street Covers Well

Education technology sits in an unusual spot in the public markets. The TAM is enormous — global spending on education exceeds $6 trillion annually, and digital penetration remains below 5% in most categories. Research firms project the EdTech market specifically will reach $400–450 billion by 2028, growing at a 13–16% CAGR from roughly $250 billion in 2024. Yet sell-side coverage is thin, institutional ownership is concentrated, and most generalist investors struggle to differentiate between a Duolingo (category-defining consumer product) and a Chegg (category-destroyed casualty of AI disruption).

That coverage gap is exactly why the sector is interesting. The market has painted EdTech with a broad “AI risk” brush since ChatGPT launched in late 2022, compressing multiples across the board. Chegg deserved the punishment. But the collateral damage created opportunities in names where AI is an accelerant, not a replacement. Understanding the difference between these two categories is the single most important analytical exercise in EdTech investing today.

The sector breaks into several investable verticals: consumer language learning (Duolingo), online higher education and professional credentialing (Coursera, 2U), K–12 infrastructure (Powerschool, Instructure), corporate learning and development (Coursera for Business, LinkedIn Learning), and test preparation (Chegg, formerly). Each vertical has different competitive dynamics, different AI exposure profiles, and different unit economics. Let's break them down.

Duolingo: The Engagement Machine Wall Street Underestimated

Gamification Is the Moat

Duolingo is not really a language learning app. It is a behavioral design masterpiece that happens to teach languages. That distinction matters enormously for the investment thesis. CEO Luis von Ahn, the computer scientist who invented CAPTCHA and reCAPTCHA, designed Duolingo around a simple insight: the biggest problem in education is not access to content — it is getting people to actually show up and practice consistently. Content is abundant. Discipline is scarce.

The results speak for themselves. Duolingo reported 113 million monthly active users and 37.2 million daily active users in Q4 2025. The DAU/MAU ratio of roughly 33% is extraordinary for a non-social-media app — comparable to Instagram and well above most consumer subscription products. Streaks, the app's signature mechanic, create a psychological commitment that compounds over time. Roughly 15 million users maintain streaks of 30+ days. The leaderboard system, borrowing from mobile gaming, triggers competitive instincts that keep users returning. Hearts create scarcity and urgency. The owl mascot's passive-aggressive push notifications have become a cultural meme — and they work.

Financially, the engagement translates into a machine. Revenue reached approximately $740 million in 2025, up 41% year-over-year. Paid subscribers grew to 8.6 million, with ARPU expanding as more users upgraded from Super Duolingo ($6.99/month) to Duolingo Max ($13.99/month). The free-to-paid conversion rate of 8–8.5% is nearly double Spotify's and continues to creep higher as Max penetration increases. Operating margins reached approximately 17%, up from 3% just two years prior, demonstrating the operating leverage inherent in a software product with near-zero marginal cost per user.

AI as Accelerant: Duolingo Max and Birdbrain

Where Chegg got destroyed by AI, Duolingo has embraced it as a core product differentiator. Duolingo Max, launched in partnership with OpenAI, adds two AI-native features: Roleplay, where users practice conversations with AI characters in contextual scenarios (ordering food in Paris, negotiating at a market in Mexico City), and Explain My Answer, which provides detailed grammatical explanations of mistakes. These features are available only on the highest-priced tier, creating a natural upsell path.

Behind the scenes, Duolingo's internal AI capabilities are even more significant. The Birdbrain machine learning engine personalizes lesson difficulty, content ordering, and review scheduling for each individual learner. The company reported that AI-assisted content creation has reduced the cost of building new courses by 40–50%, enabling expansion into music education (Duolingo Music) and mathematics (Duolingo Math) without proportional headcount increases. This is the key: AI is reducing Duolingo's cost to enter adjacent verticals while simultaneously improving the core product. Few companies in any sector can make that claim.

Coursera: The Enterprise and Credential Play

Coursera occupies a fundamentally different position in the EdTech landscape. While Duolingo is a consumer engagement product, Coursera is a credentialing and enterprise platform. It partners with 325+ universities (including Stanford, Duke, University of Michigan) and 100+ industry partners (Google, IBM, Meta) to offer professional certificates, specializations, and full online degree programs. Revenue was approximately $720 million in 2025, growing at a more modest 8–10% annually.

The bull case for Coursera centers on the enterprise segment. Coursera for Business serves over 5,200 corporate clients, including three-quarters of the Fortune 500. Enterprise revenue now represents roughly 30% of total revenue and is growing at 15–18% annually, meaningfully faster than the consumer segment. The value proposition is straightforward: companies need to reskill their workforces for AI, data science, cloud computing, and cybersecurity. Building internal training programs is expensive and slow. Buying off-the-shelf content from universities through Coursera is faster and carries the credibility of institutional brand names.

The Google Professional Certificates program is a good example of the model working. Google offers certificates in data analytics, IT support, UX design, project management, and cybersecurity through Coursera. Over 2 million people have enrolled. Google and 150+ employers recognize these certificates as equivalent to a four-year degree for hiring purposes. This “credential as degree alternative” trend plays directly into Coursera's strengths and is nearly impossible for an AI chatbot to replicate — because the value is in the recognized credential, not the underlying knowledge.

Where Coursera Struggles

The bear case is equally straightforward. Coursera's consumer segment, which includes individual learners purchasing single courses or specializations, is under pressure from free AI alternatives. Why pay $49–79 per month for a course on Python when Claude or ChatGPT can teach you interactively for free? Consumer revenue growth has decelerated to low single digits, and completion rates on self-paced courses remain stubbornly low (estimated at 5–15%). Margin pressure from university revenue-sharing agreements — Coursera typically retains only 40–60% of course revenue, remitting the rest to the university partner — limits profitability. GAAP operating margins remain negative, though adjusted EBITDA turned modestly positive in 2025.

EdTech Public Market Landscape: Key Metrics

The table makes the bifurcation unmistakable. Duolingo's 15x revenue multiple versus Chegg's 0.4x is not a market inefficiency — it reflects a fundamental difference in AI positioning. The market is paying for growth, engagement moats, and AI-native product development at Duolingo, while pricing in terminal decline at Chegg. Coursera sits in the middle: cheap enough to be interesting, but lacking the growth profile or margin structure to command a premium multiple. For investors looking to understand how AI is reshaping other sectors similarly, our analysis of AI agent stocks and autonomous software covers parallel dynamics.

The 2U Bankruptcy: What Consolidation Means for Survivors

2U's Chapter 11 filing in July 2024 was the most significant event in EdTech outside the Chegg collapse. 2U was once valued at over $6 billion as the leading online program management (OPM) provider, partnering with universities like Harvard, Georgetown, and UNC to deliver online degree programs. The bankruptcy was driven by unsustainable economics: 2U typically took 60% of tuition revenue in exchange for marketing, enrollment, and technology services, leaving universities with razor-thin margins and limited control over their own online programs.

The consolidation signal matters for surviving players. 2U's assets were acquired out of bankruptcy for approximately $110 million by a private equity consortium — pennies on the dollar relative to peak valuation. This capital destruction removed a significant competitor from the market and validated the thesis that asset-heavy, services-dependent EdTech models are structurally challenged. The winners from 2U's exit are platforms with asset-light models: Coursera (which competes for university partnerships but on more favorable revenue-sharing terms), Instructure (whose Canvas LMS is the infrastructure layer that universities cannot easily replace), and Noodle (a private OPM alternative gaining share).

More broadly, the EdTech shakeout of 2023–2025 — 2U bankrupt, Chegg imploding, Byju's collapsing in India, multiple smaller players acquired at distressed valuations — is classic late-cycle consolidation. The parallels to the SaaS shakeout of 2022–2023 are instructive: the survivors that emerged (like Datadog, CrowdStrike, and ServiceNow) gained pricing power and market share from departed competitors. The same dynamic is beginning to play out in EdTech.

Corporate Upskilling: The Most Durable Demand Driver

Why This Cycle Is Different from MOOCs 1.0

Remember when MOOCs (massive open online courses) were supposed to replace universities? That was 2012. It didn't happen. But something arguably more important did: corporations became the primary buyers. Global spending on corporate training exceeded $380 billion in 2025, according to Training Industry estimates. The World Economic Forum projects that 50% of all employees will need reskilling by 2027 as AI transforms job functions. McKinsey estimates that 375 million workers globally will need to switch occupational categories by 2030.

This is not aspirational — it is happening now. Amazon has committed $1.2 billion to upskilling 300,000 employees. JPMorgan Chase allocated $600 million annually for technology training. PwC spent $3 billion on a multi-year digital upskilling initiative for its 328,000 employees. These budgets are flowing to platforms like Coursera for Business, LinkedIn Learning, Udemy Business, and Pluralsight. The tailwinds are structural: AI adoption requires new skills (prompt engineering, data science, machine learning operations), regulatory compliance is expanding (GDPR, CCPA, DORA in financial services), and skills-based hiring is replacing degree requirements at an accelerating pace.

For public market investors, the question is which platforms capture these enterprise budgets most efficiently. Coursera's enterprise segment, with its university-branded content and professional certificate partnerships, is well-positioned. Instructure's Bridge platform targets the same buyer but with a focus on compliance training and skills assessment. Powerschool (PWSC), while primarily K–12 focused, benefits from government education technology budgets that are expanding alongside corporate spending. Understanding how these enterprise dynamics play out is closely related to the broader vertical AI SaaS investment thesis we have covered previously.

Instructure and Powerschool: The Infrastructure Layer

Not every EdTech investment needs to be a consumer bet. Instructure (INST) owns Canvas, the dominant learning management system (LMS) used by universities and K–12 districts across North America. Canvas holds an estimated 35–40% market share in U.S. higher education, ahead of Blackboard (now part of Anthology) and Moodle. The business generates roughly $550 million in revenue with 68% gross margins, 22% operating margins, and 95%+ gross retention rates. Canvas is infrastructure — once a university deploys it, the switching costs are enormous (migrating course content, retraining faculty, re-integrating with student information systems).

Thoma Bravo acquired Instructure in 2020 and has since shifted the business toward profitability and add-on acquisitions. Instructure acquired Parchment (academic credentialing) in 2023, giving it a direct play in the micro-credentialing trend. Canvas serves as the operating system for education — AI tools, content providers, and assessment platforms all integrate with it rather than compete against it. This infrastructure positioning makes Instructure largely immune to the “will AI replace this?” question that plagues content-focused EdTech companies.

Powerschool (PWSC) occupies a similar infrastructure position in K–12 education. The company provides student information systems, enrollment management, and classroom technology to over 18,000 school districts globally. Revenue is approximately $700 million with mid-60s gross margins and improving operating leverage. Vista Equity Partners took Powerschool private in late 2024, but the company may return to public markets via an IPO or SPAC in 2026–2027, giving investors another entry point.

AI: Both the Existential Threat and the Biggest Opportunity

The Bear Case: Free AI Tutoring Kills Paid EdTech

The bearish argument is simple and not easily dismissed. ChatGPT, Claude, Gemini, and their successors can already explain complex topics with patience and clarity that rivals the best human tutors. They can generate practice problems. They can provide instant feedback. They can engage in Socratic dialogue. Khan Academy's Khanmigo tutor, built on GPT-4, offers a glimpse of what free AI-powered education looks like at scale. If a student can get a personalized AI tutor for $20/month (or free), why pay $49/month for a Coursera subscription or $13.99/month for Duolingo Max?

The Chegg precedent makes this bear case concrete rather than hypothetical. Chegg's subscriber base was cut in half in 18 months, not because ChatGPT was better at explaining chemistry — it often was not — but because it was free and good enough. The lesson: in education, “good enough for free” beats “excellent for a price” when the user's primary goal is getting an answer rather than building a skill. Any EdTech company whose product can be summarized as “it gives you answers” should be valued accordingly.

The Bull Case: AI as Product Moat Deepener

The bullish rebuttal is more nuanced but, we believe, more correct for the right companies. Duolingo's product is not information — it is a habit. Nobody pays for Duolingo because they cannot find Spanish vocabulary elsewhere. They pay because Duolingo makes them actually practice every day through gamification mechanics that no AI chatbot replicates. ChatGPT cannot give you a 500-day streak, put you on a leaderboard against your friends, or send you a passive-aggressive notification at 9 PM reminding you that your streak is about to die. These behavioral hooks are Duolingo's true moat, and AI makes them stronger by enabling better personalization (Birdbrain) and richer interaction (Roleplay in Max).

For Coursera, the bull case rests on credentialing. A Google Data Analytics Certificate completed on Coursera signals something specific to employers: this candidate completed a structured, assessed program designed by Google and recognized by 150+ hiring partners. ChatGPT can teach the same material, but it cannot issue a credential that an employer recognizes. As long as the labor market values signaling (and it always will), credentialing platforms retain value that pure AI tutoring cannot match. The real risk for Coursera is not AI replacing its content — it is AI reducing the perceived value of credentials over time. If employers stop requiring certificates because they can assess skills directly through AI-proctored assessments, Coursera's moat narrows. We are not there yet, but it is the right risk to monitor.

How to Position: A Framework for EdTech Investing

We think about EdTech investment through a two-by-two matrix. On one axis: is the company's core product information delivery or behavioral/credentialing? On the other axis: is AI a product enhancer or a product substitute? Companies in the “behavioral or credentialing + AI enhancer” quadrant are buys. Companies in the “information delivery + AI substitute” quadrant are sells. Duolingo sits squarely in the buy quadrant. Chegg sits in the sell quadrant. Coursera is in between — its enterprise and credentialing segments are buy-quadrant businesses, but its consumer self-paced course business is increasingly sell-quadrant.

On valuation, Duolingo at 15x NTM revenue is expensive by any traditional standard. But the 40%+ revenue growth, 73% gross margins, expanding operating leverage, and a clear multi-product expansion roadmap (music, math, additional languages) create a path to a $2+ billion revenue business by 2028. If margins continue expanding toward 25%+ operating margins, Duolingo could generate $500+ million in operating income on $2 billion in revenue — justifying a $25–30 billion enterprise value at a 50–60x operating income multiple typical of premium consumer subscription businesses. The risk is that growth decelerates before margins fully expand, leaving the stock trapped at a premium multiple on decelerating growth.

Coursera at 3x NTM revenue is cheap for a reason, but the enterprise segment alone could be worth $2–2.5 billion at 5–6x revenue, implying the consumer business is being valued at essentially zero. That feels like too much pessimism for a platform with 148 million registered learners and partnerships with the world's top universities. A scenario where Coursera pivots more aggressively toward enterprise and credentialing — effectively conceding the consumer self-paced market to AI — could re-rate the stock meaningfully. Investors applying unit economics analysis frameworks should focus on enterprise cohort economics and credential completion-to-employment conversion rates.

Frequently Asked Questions