A 9-for-9 management team voluntarily reset its own playbook to chase 100M DAUs by 2028.

• The reset. In the Q4 FY25 letter (Feb 2026), management guided FY26 bookings to +11% (from +33% delivered) and adjusted EBITDA margin to ~25% (from 29.5%) — a self-inflicted ~$50M bookings give-back and 450bps margin cut to widen the free-user funnel.
• The confession. The CEO admitted in writing that "some of this deceleration is a function of our increased focus on monetization in recent years" — the same team that beat the high end of guidance for nine straight quarters had been over-monetizing growth and choking the word-of-mouth flywheel.
• The load-bearing number. Getting from 52.7M DAUs today to 100M by 2028 requires DAU growth to re-accelerate from 30% in FY25 toward the mid-twenties for three years. Q1 FY26 came in at +21%. Everything in the thesis sits on that re-acceleration.

Software-grade economics — and a one-time tax sleight that flatters every screen.

Gross margin held at 72%, FCF hit $370M against $18M of capex, and the balance sheet carries no real debt — software-quality on every operating line. But FY25 net income of $414M includes a one-time $245M deferred-tax-asset release, so the trailing P/E of 13.7× is fictional; on normalized earnings the multiple is closer to 36×. SBC at 13% of revenue means the $400M buyback authorized in February 2026 is sized to neutralize annual dilution, not to compound per-share value yet.

Is this Spotify or Chegg? The 79% drawdown is the market arguing over the answer.

• The Spotify case. S&M is only ~12% of revenue — roughly half the performance-marketing norm and in line with Spotify. DAU/MAU at 39.6% is best-in-class for a consumer app. 15M users carry 365+ day streaks and 43M carry 7+ day streaks — habit infrastructure AI-native challengers cannot bootstrap.
• The Chegg case. The replicable layer — tap-to-translate lessons, AI feedback — is exactly what a free AI chatbot already does. Speak crossed a $1B private valuation in December 2024 attacking conversation, the niche Duolingo is weakest in. Chegg's revenue fell from $776M to $377M on the same logic and the stock now trades below cash.
• The tell. 82% of revenue routes through Apple and Google — both own foundation-model competitors and could bundle a free tutor tomorrow. A 5pp move in store fees swings ~$50M of gross profit. The bear's worked example is one policy change away.

Same team. Opposite playbook.

Before: Duolingo's IPO-through-2024 arc was textbook — gamification drove engagement, engagement drove paid subscribers, and the founder-CEO team beat the high end of bookings, revenue, and EBITDA guidance for nine consecutive quarters with multi-raise discipline. Duolingo Max was the premium AI tier through seven straight letters.

Pivot: In February 2026, the Q4 FY25 letter dismantled the playbook in one print. Max's flagship Video Call feature was demoted to the cheaper Super tier, FY26 bookings growth was guided to a third of FY25's pace, and EBITDA margin was guided down 450bps. The stock fell 14% on the day on 8× normal volume.

Today: Q1 FY26 (May 2026) executed the new playbook on plan but did not yet show DAU re-acceleration. Fidelity cut its position 59% on May 13. The Q3 FY26 print in early November is when the reinvestment trade either earns the lean or breaks the moat narrative.