Reading time: 4 min
Duolingo shares crashed more than 22% in after-hours trading on Thursday after the company announced a deliberate strategy pivot: sacrifice near-term profitability to chase user growth, with the goal of doubling daily active users to 100 million by 2028. The Q4 numbers were fine — revenue beat estimates, EBITDA beat estimates — but the forward guidance was not. Full-year 2026 bookings came in at $1.27–1.30 billion versus the $1.39 billion Wall Street expected. The stock, already down 38% year-to-date before the print, traded to roughly $92 in the post-market session.
The Quarter That Didn’t Matter
The Q4 CY2025 results were solid by any conventional measure. Revenue came in at $282.9 million, up 35% year-on-year and $7 million above consensus. Adjusted EBITDA hit $84.3 million — a 29.8% margin, nearly five percentage points of expansion versus the year-ago quarter and 7.8% above analyst estimates. Net income was $42 million. Gross margin expanded 90 basis points to 72.8%, driven by rising subscription margins and a favourable revenue mix shift toward paid tiers. Daily active users grew 30% year-on-year in Q4, despite lapping 51% growth in the prior-year period. Monthly active users hit 133.1 million, up 16.4 million. Duolingo crossed two milestones in 2025: 50 million DAUs and $1 billion in annual bookings for the first time.
None of it mattered. The sell-off was entirely about what comes next.
The Pivot
CEO Luis von Ahn’s shareholder letter framed the shift explicitly: Duolingo has spent recent years fine-tuning monetisation through more ads and subscription prompts, which lifted bookings per user but coincided with decelerating user growth. DAU growth slowed through 2025 and is expected to fall to roughly half the pace of prior years. Von Ahn called this a function of both natural scaling and “our increased focus on monetisation in recent years.” He is now reversing course.
The centrepiece is moving “Video Call with Lily” — the AI-powered conversation feature that has been Duolingo’s most commercially successful AI product — from the premium Max tier down into the cheaper Super Duolingo subscription. The company also plans to roll out more AI-driven speaking tools to free users, reducing the friction that previously nudged learners toward paid plans. Von Ahn told Reuters the AI video call feature is now more than ten times cheaper to run than at launch, making broader access economically viable.
The financial trade-off is explicit. Bookings growth is expected at roughly 11% in 2026, compared with the approximately 20% the company said it could have delivered under the previous approach. Adjusted EBITDA margin is forecast to decline to about 25%, down from 29.8% in Q4. Revenue guidance of $1.20–1.22 billion missed the $1.26 billion LSEG consensus. First-quarter bookings guidance of approximately $301.5 million came in 8.5% below the $329.7 million Visible Alpha estimate.
The Math Behind the Bet
Duolingo’s argument is straightforward: it believes that accelerating AI capabilities — frontier models getting cheaper and more capable — are about to fundamentally reshape education, and that the company with the largest engaged user base will capture disproportionate value when that happens. “If we’re seeing faster user growth than we’re expecting, and what we are expecting is about 20%, then that means the strategy is working,” von Ahn told Reuters. The target is 100 million DAUs in the medium term, roughly doubling the current base.
The board simultaneously authorised a $400 million share buyback — a signal that management considers the stock undervalued at current levels and wants to offset the earnings dilution from the strategy shift. At Thursday’s post-market price of roughly $92, the buyback capacity represents close to 9% of the company’s market capitalisation.
The Broader Context
The sell-off does not exist in isolation. Duolingo had already shed 24% in January and another 18% in February before Thursday’s report, driven by a combination of broader SaaS multiple compression and investor anxiety about AI disruption of software business models. The stock is now down more than 80% from its 52-week high of $544.93 reached in May 2025. Prior to the earnings print, Investing.com reported the forward P/E had contracted to under 9× — remarkable for a company growing revenue at 35% year-on-year.
The analyst consensus heading into the report was still a Buy, with a mean price target of $250 — implying 124% upside from the pre-earnings close of $117.45 and even more from the post-market crash level. Morgan Stanley had already cut its target to $245 from $275 ahead of the print, noting it was “tactically cautious” about user growth prioritisation leading to below-consensus guidance. That is precisely what happened.
The question for investors is whether Duolingo’s pivot resembles Netflix’s 2011 decision to sacrifice DVD profits for streaming scale — a move that destroyed the stock in the short term but created extraordinary long-term value — or whether it is a company responding to decelerating growth by giving away what it previously charged for and hoping volume compensates. The AI cost curve is moving in von Ahn’s favour: if running Video Call with Lily keeps getting cheaper, giving it to 100 million users instead of charging a premium for it to 10 million could eventually generate more total revenue through higher engagement and conversion. But “eventually” is doing significant work in that sentence, and markets are not currently in a patient mood.
