Duolingo CEO: AI Lifts Output 4-5x, No Layoffs

Duolingo CEO Luis von Ahn says Duolingo AI makes employees four to five times more productive, as the company lifts its 2025 revenue guidance to $1.02 billion without cutting headcount.

Lead

What Happened

Von Ahn stated the productivity multiple plainly: "With the same number of people, we can make four or five times as much content in the same amount of time." The remarks, made in an interview published September 17, 2025, came five months after Duolingo formally declared an AI-first strategy and signaled that contractors performing content-creation work would be phased out as AI assumed those functions.

The CEO was explicit that the strategy is not a cost-cutting measure. "The goal is not to save money. The goal is not to replace human employees. The goal is to do a lot more with a slightly larger number of employees," von Ahn said. Since April 2025, Duolingo has expanded its full-time headcount rather than reduced it — and has not laid off a single full-time employee in its history. Contractors have been gradually replaced in tasks where AI now performs equivalent work, but the full-time workforce has grown.

The model retains human employees in oversight and directorial roles. "There are still humans that have to direct the computer to do the right thing, but each human is able to do way more," von Ahn said — a framing that positions AI in workplace integration as an amplifier of human decision-making rather than a substitute for it.

AI Products and New Verticals

The AI-first transition has accelerated Duolingo's product expansion. The Lily AI agent, which enables learners to conduct simulated video conversations in their target language, anchors the premium Duolingo Max subscription tier and is a primary driver of subscription revenue acceleration. A chess curriculum launched after an internal "vibe coding" experiment in which a product manager and designer built a working prototype using AI-assisted development tools — illustrating how small teams can now ship new product lines in compressed timelines.

Generative AI costs have risen alongside these launches, representing a growing component of cost of revenues alongside third-party payment processing and hosting fees. Early progress in optimizing those costs at scale has been reported, a condition for preserving margin as the model matures.

Financial Impact

The tech efficiency gains are registering in Duolingo's guidance trajectory. The company raised its full-year 2025 revenue forecast to $1.02 billion in August 2025, up from a prior estimate of $996.6 million. Q3 2025 revenues came in at $271.7 million, a 41% increase year-over-year. Q2 2025 posted 41% total revenue growth and 46% subscription revenue growth, reflecting accelerating Duolingo Max penetration. Full-year 2024 revenues reached $748.0 million, also 41% higher than the prior year — indicating that the growth rate has been sustained across multiple periods rather than inflated by a single-quarter effect.

Limits and Recalibration

Von Ahn was candid about where AI has not yet closed the gap with human output. AI-generated code introduces debugging complexity not present in human-authored code, and AI-written narratives — where Duolingo's storytelling format depends on creative coherence — do not yet match human quality consistently. "It's not yet the case" that AI codes better than humans, von Ahn noted.

Those limits informed a notable course correction. In April 2026, von Ahn removed AI usage as a formal performance-review metric, reversing one of the more contentious elements of the April 2025 AI-first announcement. Employees had questioned whether measurement of AI adoption was pressuring them to use tools regardless of whether outcomes improved. "I'm not going to force you," von Ahn said, adding that performance should be judged on doing the job "as well as possible" — with AI as a means, not a metric. The episode follows a pattern now familiar in enterprise AI adoption: declarative commitment, internal friction, and subsequent calibration toward a more pragmatic integration model.

Outlook

Duolingo enters the final stretch of 2025 on course to cross $1 billion in annual revenue for the first time, sustained by a 41% growth rate consistent across consecutive quarters. The company's AI productivity thesis — four to five times more output from the same workforce — is translating into content scale, faster product launches, and expanding subscription revenue, while full-time employment remains stable. The central variable going forward is whether rising generative AI infrastructure costs erode the margin benefits of higher throughput, or whether ongoing cost optimization preserves the efficiency gains at scale. As AI capabilities advance in code generation and narrative quality — the two areas von Ahn identified as current ceilings — the productivity multiple is likely to shift upward, with compounding implications for competitive positioning across digital education.