THE PARADOX. Five straight EPS beats, and the stock is down 37% this year. The market isn't pricing Adobe's results — it's pricing the end of the reason the results existed.
THE NUMBERS. Adobe is down roughly 37% in 2026 and about 60% since January 2024. It trades near 13 times earnings — a multiple you'd expect from a mature hardware company, not the software franchise that once traded above 40. And here's the strange part: Adobe keeps beating earnings. It has beaten EPS five straight quarters.
That combination — good results, collapsing stock — is the market telling you something specific. It isn't saying the business is broken today. It's saying it no longer believes the reason the business worked will keep working.
WHAT THE MOAT ACTUALLY WAS. For twenty years Adobe's power came from two things that had nothing to do with software quality.
First, the skills barrier. Photoshop, Illustrator and Premiere were hard to learn. That difficulty was the product: if you wanted a professional result, you hired someone who had spent years mastering the tools, and that person paid Adobe $60 a month for the privilege of being employable.
Second, the file-format chokepoint. PSD, AI, PDF. The industry exchanged files in Adobe's formats, and the only reliable way to open them properly was to pay Adobe.
Generative AI broke the first one in about eighteen months. The image generation tools don't require you to understand layers, masks or curves — they require a sentence. The skills barrier that justified the subscription evaporated, and with it the argument that professionals must pay forever to stay competent.
THE CAPTAIN LEFT. In early 2026, CEO Shantanu Narayen — the architect of the subscription pivot that made Adobe a Wall Street darling — exited unexpectedly. A search is underway. The company announced a $25 billion buyback.
Both facts matter. A buyback is what a company does when it believes its own stock is the best available investment — which is a real signal, and also a confession that it doesn't have a better place to put the money. And a leadership vacuum during a technological transition is the worst possible timing. Every strategic bet now gets made by a committee that knows it might be replaced.
THE DEFENSE. Adobe's case isn't nothing. Firefly, its generative model, is trained on licensed content and carries commercial indemnification — meaning a brand can use the output without worrying about a copyright suit. For enterprises, that's not a feature, it's the whole decision.
The numbers show it's working, just not fast enough: AI-first annual recurring revenue has passed $650 million, up more than 150% year over year. That's real growth. It's also a rounding error against the core business, which means the new thing is growing fast enough to celebrate and too small to replace what's being lost.
THE SEPTEMBER MISS. In September 2026, Adobe guided to sales that narrowly missed estimates, and the stock slipped again. In October, an analyst double-downgrade cited shifting AI budget priorities, near-term guidance risk, and doubt about converting free users into paying ones.
Read that last part carefully: freemium conversion. Adobe's growth strategy now depends on giving away more and hoping people upgrade. That's the strategy of a company whose product has become a commodity at the entry level — which is exactly what AI did to image editing.
THE STRUCTURAL PROBLEM NOBODY WANTS TO SAY. Adobe sells seats. Its revenue is a function of how many people need professional creative tools. AI's entire promise is doing more with fewer people.
Every marketing department that uses generative tools to produce ten variations of an ad without a designer is a department with fewer seats. Adobe is selling shovels into a gold rush that is actively reducing the number of miners. The company can grow ARR per user while the number of users shrinks — and that math eventually stops working.
WHAT WOULD PROVE THIS WRONG. Two things, and both are plausible. First, if enterprises decide that legal indemnification is worth more than model quality — that a brand will pay a premium for output it can defend in court — Adobe's position becomes the safest one in the market, not the weakest one. Its Firefly pitch is precisely that bet.
Second, if AI turns every company into a content company, the total market for creative output expands faster than the price per seat falls. More ads, more variations, more video, more localisation — Adobe could sell fewer seats to each customer and still be bigger overall, because there are vastly more customers.
Those are real outcomes. They're just not the ones the multiple is pricing anymore, and a company can't argue with a market that has decided to wait and see.
SPLASH'S READ: Adobe isn't dying. It's being de-rated, which is a slower and more interesting kind of fall. The business still makes money, still beats estimates, still has the enterprise trust that startups can't buy. What it lost was the story that justified a premium multiple: that professional creative work requires professional tools, and that those tools cost money forever.
The lesson generalizes past Adobe. A moat built on difficulty evaporates when the difficulty does. BlackBerry's moat was a keyboard. Kodak's was film chemistry. Adobe's was the learning curve — and a machine just flattened it.
The company that survives this will be the one that stops selling access to software and starts selling accountability: we generated it, we licensed it, we'll defend you in court. That's a business AI can't commoditize, because it's a promise, not a feature.