Creative Destruction

The economist Joseph Schumpeter gave the world a phrase for the way capitalism renews itself: creative destruction, the gale of innovation that tears down established businesses to make room for new ones, the relentless churn by which the new technology buries the incumbent that the old technology built. It is the most important idea in understanding why no corporate moat is permanent. And there is a dark joke in the fact that the company now serving as the market's leading test case for creative destruction in the age of artificial intelligence is Adobe — a company whose entire business is, literally, creative software. Adobe sells the tools of creation. Artificial intelligence has now learned to create. The question the market is asking, with growing anxiety and a falling share price, is whether Adobe rides the gale or is flattened by it.

Start with what Adobe is, because the strength of the franchise is what makes the question so stark. Adobe owns the professional content-creation stack: Photoshop, Illustrator, Premiere, After Effects, InDesign, and the Acrobat/PDF franchise it invented and still dominates. These are not just market leaders; they are standards — the default tools taught in every design school, demanded in every creative job listing, embedded in the workflows of entire industries. Adobe converted that dominance into one of the great recurring-revenue machines in software: an estimated 40-plus million Creative Cloud subscribers (Adobe reports recurring revenue rather than a precise headcount), around $23.8 billion of revenue in its 2025 fiscal year (up about 11%), Digital Media annual recurring revenue of roughly $19 billion, and the kind of high margins that a near-monopoly on an essential professional tool produces. For most of the past decade, Adobe was a quintessential quality-compounder — a stock you bought and held because its dominance seemed unassailable and its subscription cash flows seemed eternal.

The thirteen-times tell

Now look at what the market pays for all of that today. Adobe trades at roughly thirteen times earnings — and that number is the entire story compressed into a single statistic. Thirteen times earnings is not a growth multiple; it is not even an average multiple. The broad market trades around twenty times; the software sector, on the strength of exactly the AI optimism documented elsewhere in this series, trades far higher; and many software companies with slower growth and weaker franchises than Adobe command multiples two, three, even five times Adobe's. A thirteen-times multiple on a dominant, profitable, growing software franchise is the market's way of saying something specific and ominous: that it does not believe the earnings are safe. The stock, which once traded near $700 at its 2021 peak, now sits in a 52-week range whose low is around $224, having fallen roughly 30% in the first months of 2026 alone, carrying its market value down to around $96 billion. The market has re-rated Adobe from a quality-compounder to a value stock — and value multiples on former growth darlings are usually a sign that investors fear the business is not cheap but melting.

What changed was not Adobe's current numbers, which remain solid. What changed was the market's belief about the durability of those numbers, and the trigger was the arrival of generative AI as a credible threat to the very thing Adobe sells. The fear crystallized around specific events — when Anthropic launched a design-focused AI tool in the spring of 2026, Adobe's stock took another leg down — because each new demonstration that an AI can produce professional-quality images, layouts, and video on command is a demonstration that the skills Adobe's software monetizes might be becoming a commodity. The thirteen-times multiple is not a verdict that Adobe is doomed. It is a measure of how much uncertainty the market now attaches to a franchise it used to consider certain.

Squeezed from both ends by the same technology

The reason Adobe's position is so genuinely difficult — and the reason the bear case is not crazy — is that generative AI threatens it from two directions at once, with the same blade. From below, AI-native tools are commoditizing the bottom of the market. Canva, which now claims something like 260 million monthly active users, lets ordinary people with no training produce good-enough designs, social graphics, and marketing content using simple AI-powered tools, eroding Adobe's relevance for the vast population of non-professional creators who never needed Photoshop's full power and now need it even less. The "good enough," produced instantly and cheaply by an AI, is the enemy of the expensive professional tool, because most content in the world does not require professional polish — it requires speed and adequacy, which is exactly what AI delivers.

From the side and above, the frontier AI labs themselves are encroaching on the professional core. The image generators built by OpenAI and others, the video tools from companies like Runway, the design-focused models from Anthropic and the rest — these can increasingly perform tasks that once required Adobe's professional software and a trained operator: generating, editing, and transforming images and video directly from a text instruction. Figma, the collaborative-design upstart, has embedded its own AI agent into its canvas and is growing revenue at better than 40%, taking the enterprise-design workflows Adobe covets. The threat is not one competitor; it is the entire direction of the technology, which is to collapse the distance between wanting a piece of visual content and having it — removing, in the process, the steps that Adobe's tools exist to perform. Adobe is being squeezed from below by commoditization and from above by the frontier, and both pincers are powered by the same generative AI.

Moat, or margin leak?

Adobe's answer to all of this is Firefly, its own family of generative-AI models, baked directly into Creative Cloud, along with a strategy of integrating partner models and offering enterprise customers something the upstarts cannot easily match: AI trained on licensed content, with the intellectual-property indemnification that risk-averse corporations need. It is a credible response, and Firefly has real early traction — its annual recurring revenue crossed roughly $250 million as AI-first revenue more than doubled. The bull case is that AI becomes a new product Adobe can sell, a reason to charge existing subscribers more and to win new ones, turning the threat into an upsell and expanding the moat. In this telling, Adobe's distribution (its tens of millions of subscribers), its data, its enterprise relationships, and its trusted, indemnified models let it ride the gale rather than be flattened by it.

But there is a darker reading, and the early financial evidence leans toward it. The question, as one analyst sharply framed it, is whether generative AI becomes "a pricing engine for Creative Cloud or a cost center that gives users more output while weakening Adobe's moat." Building and running these AI models is enormously expensive, and Adobe is locked in an arms race that forces it to spend heavily on AI research and, crucially, to bundle AI features cheaply or for free to keep subscribers from defecting to AI-native rivals. That is not a pricing engine; it is a margin leak — Adobe paying more to give customers more output at the same price, because the competitive pressure means it cannot charge for the AI the way it hoped. The tell is in the margins: Adobe's Digital Media operating margin has already compressed from roughly 37% in 2021 to around 31% more recently, reflecting exactly this dynamic — heavy AI investment and the pressure to subsidize AI tools, eating into the fat margins that justified the old premium valuation. Firefly's $250 million is real, but against $24 billion of revenue it is a rounding error, and the costs of the AI arms race are not. The fear the thirteen-times multiple encodes is precisely this: that AI does not add a lucrative new revenue stream so much as it raises Adobe's costs while eroding its pricing power — the worst of both worlds.

The company that has survived its own gale before

The single strongest argument in Adobe's favor is historical: it has faced existential technological transitions before and not merely survived them but emerged stronger, and any bear who forgets this is ignoring real evidence. Adobe helped kill the print-and-paste-up world of physical design and replaced it with desktop publishing. It navigated the shift from film to digital photography. And most impressively, in 2013, it executed one of the most celebrated business-model pivots in software history: it abandoned its lucrative, comfortable model of selling boxed software for a one-time fee and forced its entire customer base onto cloud-based monthly subscriptions — a wrenching transition that analysts feared would alienate users and crater revenue, and that instead transformed Adobe into a far larger, more valuable, more durable recurring-revenue machine. A company with that track record of reinventing itself through disruption has earned the benefit of some doubt, and the bull case rests heavily on the belief that AI is simply the next transition Adobe will master, as it mastered the last several.

But the bear's rejoinder is sharp and specific, and it identifies why this transition may be genuinely different from the ones Adobe survived. Every previous disruption Adobe navigated was a change in the distribution or production model — how software was sold, how images were captured, how files were delivered — while leaving intact the core fact that creating professional content required skilled humans using powerful tools. Adobe's moat was never really the boxed disc or the perpetual license; it was the irreplaceable combination of professional skill and professional software. Generative AI attacks that core directly, for the first time. It does not change how the tools are sold; it threatens to remove the need for the tools, and the skill, altogether, by letting the machine do the creating. The subscription pivot of 2013 changed Adobe's business model while its product remained essential. The AI disruption threatens the essentiality of the product itself. That is a different kind of gale, and Adobe's history of surviving the others is not proof it will survive this one — it is, at most, evidence that it is unusually good at adapting, facing a challenge unlike any it has adapted to before.

The upstart Adobe tried to buy

There is a revealing piece of recent history that crystallizes Adobe's predicament: the Figma affair. Figma is the collaborative, browser-based design tool that came up fast as a challenger in interface and product design, and in 2022 Adobe agreed to acquire it for around $20 billion — an enormous sum, and a clear signal that Adobe's own management saw the threat coming and judged that buying the upstart was wiser than competing with it. Regulators disagreed. Antitrust authorities in Europe and the United Kingdom scrutinized the deal as an attempt by the dominant incumbent to swallow a rising competitor, and in 2023 the acquisition collapsed under that pressure, with Adobe paying a $1 billion breakup fee for the privilege of walking away empty-handed. Figma went on to thrive independently, eventually going public, embedding its own AI, and growing revenue at better than 40% — exactly the AI-native competitor Adobe had tried to neutralize, now free and well-capitalized and aimed squarely at Adobe's customers.

The Figma episode is a perfect miniature of the whole Adobe problem. It shows that Adobe saw the disruption clearly enough to try to buy its way out of it — that the threat is not a surprise to anyone, least of all to Adobe itself. It shows that the strategy of absorbing challengers, which served Adobe well in the past, is now foreclosed by regulators wary of exactly that kind of incumbent consolidation. And it shows that the upstarts Adobe most fears are not only surviving but flourishing as independent public companies, carrying the AI-native approach that Adobe's installed base of legacy tools cannot easily replicate. Adobe wanted to own the future of design and was told, by the antitrust authorities, that it would have to compete with it instead. That forced competition, against nimbler AI-native rivals on a battlefield where Adobe's incumbency may be more burden than moat, is the situation the thirteen-times multiple is straining to price.

Bargain, or melting ice cube?

So Adobe presents the same stark binary that ran through the Novo Nordisk story: at thirteen times earnings, it is either a screaming bargain or a value trap, and the two cases sound completely different. The bargain case is genuinely strong. Adobe is still growing, still hugely profitable, still the standard in professional creative work; it has a $25-billion buyback in place to shrink the share count; Firefly is gaining traction; the indemnification and enterprise stickiness are real advantages the upstarts lack; and a dominant franchise trading at a market-discount multiple, oversold on a fear that has not yet shown up in the numbers, is the kind of setup that has rewarded contrarian buyers many times before. If AI turns out to be a tool Adobe wields rather than a force that destroys it, the stock is absurdly cheap.

The value-trap case is equally coherent, and it turns on a single uncomfortable insight: an installed base is only a moat if the next generation of users joins it. Adobe's tens of millions of subscribers are the professionals and students who learned their craft on Adobe's tools over the past two decades. But the creators coming up now are learning on AI-native tools — starting with Canva, with generative image and video models, with whatever is free and instant on their phones — and a generation that never learns Photoshop is a generation that will never pay for it. The moat does not break all at once; it erodes from the bottom, as each cohort of new creators bypasses the standard that the previous cohort considered indispensable. In this reading, today's thirteen-times multiple is not cheap at all; it is sitting on top of earnings that are in the early stages of being competed away, a melting ice cube that looks like a bargain only if you assume the melting stops. No one yet knows which case is right, which is precisely why the stock trades where it does.

The first domino

The deepest reason Adobe matters — the reason it deserves a place in this collection even though it is, by the standards of the other chapters, cheap rather than expensive — is what it represents for the entire edifice of richly-valued software documented throughout this series. The market has paid extraordinary multiples for software companies on a single shared assumption: that artificial intelligence makes established software more valuable, by giving these companies a powerful new capability to sell. Adobe is the live experiment in the opposite hypothesis — that AI might make established software less valuable, by commoditizing what it does and collapsing the moats that incumbency built. And Adobe is not a weak test case; it is one of the strongest software franchises in the world, with a near-monopoly, a vast installed base, real AI products of its own, and every advantage an incumbent could want.

If Adobe can be re-rated from a quality-compounder to a thirteen-times value stock on AI-disruption fear, then the same fear can, in principle, come for any software incumbent whose core product an AI might learn to replicate — which is to say, for a great many of the names the market currently prices for permanence. Adobe is the canary, the first domino, the place where the market is testing, in real money and real time, whether the AI boom is a rising tide that lifts the software incumbents or a gale that uproots them. The earnings the company reports, the trajectory of Firefly, the path of the margins — these will offer the first real evidence either way. But the meta-lesson is already visible in the multiple: the market has decided that, for at least one great software franchise, incumbency in the age of AI is not obviously an asset, and might be a liability. Schumpeter's gale has reached the company that sells the tools of creation, and the most honest thing anyone can say — including Adobe's own management, including the analysts, including the market that has marked it down to thirteen times earnings — is that no one knows yet whether it is a wind to ride or a storm to be destroyed by. That uncertainty, attached to a franchise that two years ago felt as permanent as any in technology, is the whole story. The gale is real. Where it blows the incumbent is the one thing not yet written.