Creative Cloud Moat

Creative Cloud Moat

Adobe's advantage is real and shows up in the numbers: Photoshop, Illustrator, Premiere Pro and PDF are industry standards, 96% of revenue recurs, monthly active users of Acrobat and Express passed 750 million, and Digital Media ARR reached $19.20 billion still compounding at double digits. But durability and momentum are not the same thing. ARR growth has stepped down for five straight years, the FY2025 10-K now names "AI-first" tools as direct competitors, and the genuinely new AI revenue Adobe has built so far is roughly 1% of the book.

The question The Cash Franchise left open — whether generative AI erodes this franchise — turns on the moat's construction and on what the ARR line is already telling you. The moat holds today. The rate of change is where the debate lives.

What the moat is built from

Adobe's switching costs are not a slogan; they sit in file formats and installed habit. Photoshop is described in the filings as "the world's most advanced digital imaging and design app," Illustrator as "our industry-standard vector graphics app used worldwide," and Premiere Pro and After Effects as industry-standard tools for video and motion graphics [1]. The PSD, AI and PDF formats are the interchange layer of the professional content world; a studio, agency or enterprise that standardizes on them inherits a library, a training base, and a set of collaborators who expect the same tools. Leaving is not a software swap — it is a workflow migration.

Three numbers make the lock-in visible. Recurring revenue is 96% of the total (The Cash Franchise). Remaining performance obligations — contracted revenue not yet recognized — stood at $22.52 billion exiting FY2025, up 13% year over year, a backlog larger than a full year of Digital Media revenue [2]. And the top of the funnel is wide: monthly active users of Acrobat and Express surpassed 750 million, growing 20% year over year, feeding a free-to-paid conversion engine that the enterprise ETLA relationships sit on top of [3].

Sources: Digital Media ARR — FY2025 10-K, MD&A [4]; RPO and MAU — Q4 FY2025 earnings call [5].

The moat in the ARR line — and the deceleration

Digital Media ARR is the cleanest gauge of whether the franchise is holding. It has risen every year, from $12.24 billion exiting FY2021 to $19.20 billion exiting FY2025 [6][7]. That is a 12% compound rate over four years, achieved while the price fell two-thirds — the durability side of the case is not in doubt.

Sources: FY2021–FY2025 10-Ks, MD&A Digital Media sections [8][9][10][11][12].

The growth rate tells the other half of the story. Adobe's own reported year-over-year ARR growth has fallen in a straight line: 19% in FY2021, 15% in FY2022, 14% in FY2023, 13% in FY2024, and 11.5% in FY2025 [13][14]. Some of that is arithmetic — a larger base is harder to grow — and Adobe reports these rates largely on a constant-currency basis. But the direction has been unbroken through exactly the years generative AI moved from novelty to product.

Sources: growth rates as reported by Adobe (constant currency), FY2021–FY2025 10-Ks, MD&A [15][16].

Adobe's Digital Media ARR growth decelerated five straight years — 19%, 15%, 14%, 13%, 11.5% — to an ending $19.20 billion, and its standalone AI products reached only ~$250 million of AI Direct ARR (about 1% of that book), even as Figma grows 41% with 136% net dollar retention in the adjacent collaborative-design pool, and Adobe folds its three segments into one from FY2026, removing the standalone Digital Media growth line as the competitive question sharpens.

There is a genuine counterweight, and it belongs in the same breath. In absolute dollars, net new Digital Media ARR set a record in FY2025, and management said net-new growth "reaccelerated year over year," with over 75% of it coming from subscriptions, cross-sell and upsell rather than price increases [17]. A base that adds record dollars while its percentage growth slows is a large business behaving normally, not necessarily a franchise being taken apart. The percentage line is a warning to watch, not yet evidence of erosion.

AI: defending the moat more than widening it

Adobe's AI disclosures reward a careful read, because two very different numbers get discussed as one. The headline is that "total AI-influenced ARR exceeded one-third of our overall book of business" exiting FY2025 — on total Adobe ending ARR of $25.2 billion, that is more than $8 billion [18][19]. But "AI-influenced" counts existing subscriptions that now carry AI features. It is a retention and engagement measure — how much of the installed base touches AI — not a measure of new money.

The new money is a separate, much smaller figure Adobe labels "AI Direct ARR": the standalone AI products — Acrobat AI Assistant, the Firefly app and services, GenStudio for performance marketing. Management guided that book to surpass $250 million exiting FY2025, up from about $125 million [20][21]. At roughly $250 million, direct AI revenue is about 1% of the $19.20 billion Digital Media book. Management itself framed the strategy plainly — "drive the entire book of business with AI-influenced solutions" — which is a statement about defending and upselling the installed base, not yet about a new growth engine [22].

Sources: AI-influenced share — 2026 proxy and Q4 FY2025 call [23][24]; AI Direct ARR — Q2 FY2025 call [25]; AI Direct share derived against $19.20B Digital Media ARR [26].

Two further moves say more about how Adobe reads the threat than any margin figure. First, the FY2025 10-K, for the first time, names "AI-first" tools as direct competition — "AI-first creativity and productivity tools" for business professionals and consumers, and "AI-first creative tools" for creators [27]. Earlier filings framed competitors as software companies and point tools; the language changed because the competitive set did. Second, Adobe stopped trying to win on its own model alone. It opened Firefly to over 25 third-party models — Google, OpenAI, Black Forest Labs, Luma, Runway and others — metered through its generative-credit system [28]. That is a deliberate bet that the durable moat is the workflow and the interface where professional work happens, not the underlying model — a sensible hedge if model quality is commoditizing, and an implicit admission that owning the best model is not the plan.

The read, and what would change it

The evidence points to a wide but maturing moat. The switching costs are structural — file-format standards, a 750-million-user funnel, a backlog larger than a year of segment revenue — and they have kept ARR compounding at double digits straight through the arrival of credible AI-native rivals. Adobe's own AI is, so far, doing defensive work well: it is lifting engagement and retention across a book that is still growing, and the standalone AI products, while tiny, are growing from a base of zero.

The strongest fact against a complacent read is the growth line itself. Five consecutive years of deceleration, the first-ever naming of AI-first competitors, and a decision to embrace rival models rather than beat them are not the profile of a franchise widening its lead. The most likely path — a good business growing a little slower each year — is also what an early, gradual erosion would look like, and the two are hard to tell apart from a single year.

What would decide it is observable and falsifiable. Re-acceleration of net-new Digital Media ARR in dollars, and AI Direct ARR scaling from hundreds of millions toward the billions, would show AI turning from a retention lever into a growth engine — the moat widening. ARR growth breaking below roughly 10%, or net-new ARR rolling over in absolute terms, would be the first hard sign that AI-first tools are pulling work out of the funnel. Those are the two lines to watch; the franchise is durable enough that they, not the headlines, will settle the question.