Adobe Inc. ADBE

Key data

The business

Adobe sells the software that professional creative and document work runs on, almost all of it by subscription. Three families: Creative Cloud (Photoshop, Illustrator, Premiere, Lightroom), Document Cloud (Acrobat and the PDF standard Adobe invented), and Experience Cloud, the marketing and analytics suite enterprises use to run their sites and campaigns. Subscriptions are 97% of revenue (Q2 FY26 subscription revenue of $6.42B on total revenue of $6.62B), so this is a recurring, high-margin business, with gross margin at 89%. The moat is switching cost. A studio's files, a marketing team's workflows and a generation of muscle memory all live inside these apps, and the PSD and PDF formats are industry furniture. The live tension is AI. Midjourney, OpenAI's image models and Canva's cheaper all-in-one raise the question of why anyone pays for Photoshop when a prompt makes the picture. Adobe's answer is Firefly, its own generative AI baked into the apps and sold as commercially safe for business use.

The numbers

Revenue compounds in a near-straight line and earnings faster.

The top line climbs steadily, from $5.87B to $6.62B, but the last row breaks: net income fell year over year in Q2, up only 1.2%, while diluted EPS still rose 7.9% because Adobe retired 6% of its shares. The gap is tax and spend, not sales, a higher tax rate and heavier AI investment, and operating margin slipped to 33.8% from 35.9% a year earlier.

Over five years revenue compounded 10.7% a year and EPS 13.6%, the wedge between them the buyback at work. Revenue growth was 12.7% last quarter, steady with recent quarters, so the sharp deceleration some feared has not arrived.

Signed-but-unrecognized revenue reached $22.27B, growing in step with sales, and subscription revenue grew 14%. In software that ordering matters: the backlog leads the revenue line, and a backlog growing with or ahead of sales is the last thing you would see before a slowdown, not the first. Cash generation is real, free cash flow near 42% of revenue, though stock pay worth 8% of revenue is the toll between the cash Adobe reports and the cash an owner keeps. What the market does not believe, and the price says loudly, is that this survives AI: on sales the stock has never been cheaper in twelve years.

Management

Management runs Adobe as a cash-return machine. FY2025 buybacks were $11.3B, on top of $9.5B the year before, and the share count has fallen from about 480 million five years ago to 402 million. The honest mark against it: much of that FY2025 stock was retired at $370 to $500, well above today's $266, so the recent buyback bought high even as the long-run reduction compounds EPS. A fresh repurchase authorization landed in April 2026, and with the stock now the cheapest it has been in years, this is the buyback that should count. Adobe took on debt in 2025 to fund the return and is paying it down, long-term debt falling from $6.2B to $4.8B in two quarters. Open-market insider flow over the past year was immaterial and not relied upon here.

How it fails or surprises you

Seats meet AI. Adobe still sells mostly by the seat. If Firefly and rival AI let one designer do the work of three, seat counts and pricing erode even as usage rises. It would show first not in revenue but in the backlog: signed-backlog growth falling below revenue growth is the earliest print that renewals are shrinking. Watch that spread each quarter.

Firefly becomes a line item (right tail). The market pays nothing today for Adobe's AI. If Firefly and generative credits turn into a disclosed, growing revenue stream and push subscription growth back above 15%, the story flips from disruption victim to AI winner, and four times sales re-rates hard. The print: management breaking out Firefly revenue, or subscription growth re-accelerating.

Margins keep slipping. The fact this read explains least: net income was flat last quarter and operating margin fell two points while revenue grew 13%. If the AI build-out keeps outrunning the revenue it produces, the cheap-compounder thesis weakens into cheap-because-earnings-stall. Two more quarters of falling operating margin would prove the cheapness a trap, not a gift.

Closing thoughts

The market has already voted, and it voted disruption: at 15 times earnings and four times sales, below any multiple in the twelve years this pack covers, Adobe is priced as a business whose best days AI will end. What settles that verdict is not a forecast but the backlog. Adobe's signed future revenue grew 13% last quarter, in step with sales, and subscriptions grew 14%, and nothing in those numbers shows the install base leaving. On the other side are bears who believe the numbers simply have not caught up to the technology yet. The tail I weigh as fatter is the upside, because the price already pays for the bad outcome and the current numbers do not confirm it, but AI is a real structural risk, not a passing scare, so the downside is genuine. What is at risk if the moat cracks is years of stalled growth, against a stock that could re-rate if Firefly monetizes.

The bet is still that professional designers, marketers and businesses keep paying Adobe every month for Photoshop, Illustrator, and Acrobat instead of switching to cheaper AI-powered tools. What breaks it is the pair to watch: the signed backlog growing slower than revenue, alongside an operating margin that keeps sliding. Hold those two lines and the cheapest Adobe in twelve years is a gift. Lose them and it is a value trap wearing a moat.