Adobe's Beat Is Exactly the Problem:13% Growth, '150%+' AI ARR, and a Stock That Drops 4.5% Anyway

Adobe reported record Q3 FY2026 revenue of $6.76 billion on Sept. 10, beating consensus of $6.69 billion, with non-GAAP EPS of $6.13 ahead of the $6.08 estimate. Customer-group subscription revenue — the metric management now steers the Street toward under its post-Q4 FY25 reporting change — grew 14% year over year to $6.56 billion, and both halves of the book accelerated sequentially: Creative & Marketing Professionals subscription rose 13% to $4.65 billion, while Business Professionals & Consumers grew 16% to $1.91 billion. On the headline numbers, this is the cleanest print Adobe has delivered in three quarters.

The market, which had set the bar at a beat and an inflection, took the other side of that bet. Shares closed at $248.83, down 2.4%, then slid another 2.1% after hours to $243.62. That is a combined drawdown of roughly 4.5% on a day when the S&P 500 closed up 0.6% — a stock-specific move that says the print failed the only test that mattered.

Headline AI metrics held, but the framing weakened. Q3 FY2026 press release puts Adobe AI-first ARR at \"more than 150% year-over-year\" growth — true, but a clear deceleration from the Q2 FY26 language of \"tripled year over year and exceeded $500 million.\" At a $500M base, tripling implied ~$1.5B run-rate; at that larger base, 150% growth implies ~$1.25B annualized, suggesting the AI-first book added less in dollar terms than in the prior quarter. Adobe did not disclose a precise Q3 AI-first ARR number, leaving analysts to model it from the YoY direction.

Adobe raised full-year FY2026 revenue and EPS targets — but the size of the raise is the message, not the direction. New targets: total revenue $26.576B–$26.626B (prior $26.5B–$26.6B), non-GAAP EPS $24.45–$24.50, GAAP EPS $18.12–$18.17. Total Adobe ARR growth target — the metric the market cares most about — was maintained at 10.2% year over year. That is the same number management set in June, when it cut the FY26 ARR target by ~2 percentage points to accommodate the freemium pivot and the Semrush book of business.

Adobe crossed1 billion monthly active users across creativity and productivity products in Q3 — a marketing milestone the company leaned on hard — up from >850M at Q1 and >700M a year earlier. The freemium strategy is producing exactly what management promised: a much larger top of funnel. But the trade-off was telegraphed in June when Adobe cut the FY26 organic ARR growth target by about2 percentage points, citing \"the strategic choice\" to push more users into the free tier. Q3 ARR exited at $27.50B, up sequentially from $27.10B at Q2 (which already included ~$480M of Semrush ARR), but the maintained 10.2% growth target tells you that organic ARR is still running below10% — the freemium-driven step-down is real and continuing.

Q3 FY26 is the first print under interim CFO Steve Day, after Dan Durn's surprise departure in June — the original shock that triggered the post-Q2 selloff. CEO Shantanu Narayen transitions to Executive Chair on Dec. 1, 2026; Anil Chakravarthy, currently president of Adobe's Customer Experience Orchestration business, takes the CEO seat. Three weeks before this print, Adobe closed two strategic deals: a more-than-$4 billion expanded partnership with Saudi Arabia's MCIT and HUMAIN to provide free access to Adobe AI tools to 27 million Saudi citizens (announced Aug. 31), and the acquisition of Indian agentic-AI marketing startup Rilo — a six-person team plus IP licensing, completed Sept. 2. The Rilo deal comes four months after the $1.9 billion Semrush close (April 28, 2026).

Operationally, none of those deals moved Q3's reported numbers materially. Semrush contributed roughly $280M of FY26 Creative & Marketing Professionals subscription revenue per Q2 guidance, and the Saudi/HUMAIN deal is structured as free access rather than a paid contract — meaning it scales the AI-first user funnel (good for the freemium narrative) without contributing to AI-first ARR. Rilo is a tuck-in, not a revenue event. So the print is clean of integration noise, which is also why nothing in the numbers forced a multiple re-rating.

Adobe closed Sept. 10 at $248.83, near the bottom of its 52-week range of $190.12–$370.86, down roughly 29% year to date. On consensus FY27 earnings, the stock trades on roughly 9.3x forward P/E and ~4x EV/sales — multiples associated with a company in structural decline, not one growing 13% with $9.5B of trailing-twelve-month operating cash flow. The gap between the print (clean beat, raised outlook, record revenue) and the price (down 4.5% on the day) is the story: the market is not paying for the next 18 months of growth; it is demanding proof that ARR growth re-accelerates once the freemium base is large enough to convert at scale.

Near-term (days to quarters), the catalysts are tightly defined. The Dec. 1, 2026 CEO handover is the obvious binary — incoming CEO Anil Chakravarthy owns the Digital Experience business, which has been the slower-growing of the two segments; he will need to either accelerate that segment or articulate how GenStudio and AEP convert the AI-first funnel into ARR. Q4 FY26 (reported December) is the next hard test: if FY27 ARR growth guidance points above 10.2%, the multiple re-rates; if it's still10%, the value-trap narrative hardens. The 1B-MAU milestone will need a paid-conversion disclosure at some point in FY27 for the freemium story to stop being a cost and start being a tailwind.

Long-term (1–3 years), the structural question is whether Adobe can keep the Creative Cloud franchise intact while layering an AI-first subscription book that compounds above the corporate average. The competitor set is wide — Microsoft (Copilot), Alphabet (Workspace + Gemini), Figma (design-platform), Salesforce (Customer Experience/CDP) — but Adobe's installed base, Creative Cloud's professional moat, and the $27.50B ARR are real moats. The risk is not displacement; it is multiple compression that lasts until ARR growth re-accelerates, at which point the same cheap entry looks obvious in hindsight.