Is AI a Substitute for Adobe, or an Amplifier? I Followed Both Arguments All the Way Down

The stock fell 66% while the business set records

If I had to name the single most hated stock in the US market right now, I would name Adobe (ADBE). It is down roughly 66% from its all-time high, and at the recent lows the drawdown approached 75%. It is down 33% year to date alone. A stock that traded near $700 five years ago sits at $227 today.

Now overlay the operating results from that same stretch and you get a strange picture. Through the entire collapse, Adobe posted record revenue and record profit quarter after quarter. Revenue in the most recent quarter grew 13% year over year, annual recurring revenue passed $27 billion, and management is still guiding for more growth.

The price broke while the business improved. That gap is, to me, the entire argument.

Two things stacked on top of each other to cause the fall. First, the whole software sector corrected. After a long bull run, subscription software names — SaaS — fell out of favor all at once. The second reason is the real one, and it aimed straight at Adobe: artificial intelligence. For decades, making a beautiful image or editing a video meant learning powerful, complicated Adobe software and building real skill. Now a typed request produces a finished image in seconds for almost nothing. Wall Street watched that and drew its conclusion: AI is going to make Adobe useless.

When the market decides a company's future is in danger, it sells first and asks questions later. That is exactly what happened here.

The whole fight compresses into one sentence

Is AI a substitute for Adobe, or is it an enhancement that makes Adobe better?

The bears say substitute. I lean enhancement. But a conclusion reached without understanding the other side properly is worth nothing, so let me lay out the bear case first — all three arguments, taken seriously, because none of them are dumb.

Bear case 1 — AI started selling the finished product, not the tool

The most fundamental threat is that what people need could change entirely.

The old way is opening Photoshop, building layers, masking, and spending real time refining an image. The new way is typing: make me five finished ad designs that match my brand. In that world, the user does not care about the tool. They want the result.

Where the value migrates is the whole question. If only the output matters, value slides away from Adobe's apps and toward whoever owns the AI engine doing the work. Adobe could add every AI feature in the world and still lose, simply because customers need fewer programs, fewer seats, and fewer hours inside the software. You can win the feature race and still lose the structure.

Bear case 2 — Adobe could lose an entire generation

This is the scariest of the three to me.

The professional market is locked in tight. The real danger is not today's pros — it is the people just starting out. Teenagers making content on phone apps, small businesses for whom a simple cheap tool is plenty, creators who reach for AI-native software from day one. They may never find a reason to learn Adobe at all.

Adobe is already surrounded: web and mobile-first design apps, dead-simple content tools, AI-first creativity software, slick presentation platforms. If a whole generation grows up creating without Adobe, then Adobe keeps its existing customers and slowly loses the future. At some point it quietly becomes the software your parents used. What makes this scenario frightening is that it never shows up in a quarterly report. By the time it does, it is late.

Bear case 3 — the AI math might not add up

Adobe likes to point out that its AI products already bring in more than $500 million a year. That sounds big until you set it against more than $27 billion in total recurring revenue — it is less than 2%.

The problem is that building and running AI is wildly expensive. Terry Smith, who runs one of the largest funds in Europe, put it bluntly: he cannot see how Adobe makes money from AI. Adobe does not own the powerful engines underneath, so it may end up paying other companies for the AI while its own customers simply expect those features to be included in the subscription they already pay for. Costs up, no extra revenue in.

Smith raised a second concern along the same line. Adobe once offered a jaw-dropping $20 billion to buy Figma, a hot young design company. His reading: if Adobe was willing to spend that to swallow an upstart, maybe Adobe sees a threat it cannot simply build its way past.

Stack all three and you get the nightmare. Customers expect AI free inside what they already pay for, Adobe's costs shoot up, AI lets people get by with fewer subscriptions, and the new AI revenue stays small. AI squeezes profits instead of growing them.

I do not disagree with any of that. These are real concerns. And this is exactly where the other side begins.

Bull case 1 — a cash machine that buys time

The first thing I look for is the ability to buy time.

Last quarter Adobe pulled in about $6.5 billion in revenue, the vast majority of it steady subscription money that arrives like clockwork. It converted that into over $2 billion of profit and over $2.5 billion in cash. In one quarter. Free cash flow last year was $10.3 billion.

This matters because Adobe does not have to guess the AI future perfectly on day one. That river of cash lets it fund AI research, give away free apps to pull people in, and buy back its own stock — while surviving several mistakes along the way. It recently used that cash to acquire Semrush to strengthen its marketing tools, it keeps pushing free and cheap products like Express to hook the next generation, and it is repurchasing shares aggressively while the price is depressed.

Weak companies die during big transitions. Rich ones adapt. Blockbuster is the instructive counterexample, and most people miss the actual cause of death. Blockbuster was doing well on its streaming pivot. What buried it was not the streaming fight — it was a debt-heavy balance sheet meeting the financial crisis. With a stronger balance sheet, you might be watching something different on your TV tonight.

Bull case 2 — Adobe doesn't sell apps, it owns the workflow

This is the piece I think the market misses most.

Photoshop, Illustrator, Premiere, and Acrobat are not a loose bundle of separate programs. They are one connected system that professionals and entire companies have built their process on top of. Files, fonts, brand assets, team approval steps — all of it lives inside Adobe.

So yes, some AI tool doing one specific thing better than Photoshop is entirely plausible. Replacing a company's whole design, video, and document system is a completely different problem, and it is the kind of pain almost nobody volunteers for. Adobe isn't an app on your computer. It is closer to the operating system for professional creative work.

Then there is the second moat people forget: Acrobat and the PDF. The way the entire business world signs, shares, and manages documents runs through Adobe. Oakmark, a legendary value fund, flatly calls Creative Cloud the de facto standard for both professionals and students. De facto standard means default, and defaults do not change easily.

Bull case 3 — AI may actually make Adobe stronger

Adobe is stuffing its own AI, Firefly, directly inside every app. What matters here is that a professional needs far more than one pretty AI image.

A professional needs layers, precise control, brand consistency, correct file formats, and output that plugs cleanly into the rest of the work. Adobe does all of it.

And it holds a decisive weapon for enterprise buyers: Adobe trained its AI on properly licensed material, so companies can use it without fear of getting sued over who really owns an image. A large company will pick the safe, legal AI every single time. No exceptions.

The early proof is showing up. Adobe's AI products crossed $500 million a year and more than tripled in 12 months. There is also a telling signal at the edges — ask a general-purpose chatbot to do serious image work and you can get told that Adobe is the better place for it. Adobe keeps widening its integrations with the major model providers. AI isn't routing around Adobe. In places it is routing traffic toward it.

The two scenarios side by side

Issue	Substitute scenario (bear)	Amplifier scenario (bull)
Shape of demand	Finished-output requests rise, cutting seats and usage hours	Demand for layers and precise control remains; AI folds in as a top-layer feature
Next generation	Starts on cheap mobile tools and never learns Adobe	Free on-ramps like Express funnel users up the stack
AI economics	Costs rise while AI revenue stalls below 2% of the total	AI revenue more than tripled in 12 months; licensing safety commands an enterprise premium
Moat	Peeled off one app at a time	Leaving means replacing the entire workflow and the PDF standard
Financial cushion	Margin pressure erodes cash flow	$10.3 billion in annual free cash flow absorbs multiple mistakes

Where I land

I stand on the amplifier side, and I think the market is overreacting.

The reason is simple. While everyone argues about the future, Adobe's actual business is growing. Revenue is up 13% year over year, AI products more than tripled in a year, and recurring revenue — money effectively locked in for a year — sits above $27 billion. This is not a company falling apart. It is a company being priced like it is falling apart. That gap is where opportunity lives.

Michael Burry, the ultimate skeptic, is long Adobe for a reason that touches the same nerve. His argument is beautiful in its simplicity: large language models, no matter how advanced, will never beat the creative instinct of a talented human. AI can crank out content all day, but it cannot replace real taste and creative judgment, and Adobe is the workshop where that judgment gets executed. Mohnish Pabrai goes further — he argues established software companies like Constellation Software and Adobe have a genuine head start in this AI shift and are far more likely to benefit from it than get killed by it. The market, he says, is likely dead wrong about them.

The bears could still be right. That is why I settle this argument with a price rather than a feeling. I walked through the math in Adobe at 9x Free Cash Flow — Building the Fair Value From Assumptions, One at a Time, and I covered how the same AI fear compressed the broader software sector in Two Software Giants Under AI Fear: Adobe vs. Salesforce Valuation.

FAQ

Q: If Adobe's revenue is still growing, why does the stock keep falling? A: Because the market prices future structure, not current results. If the scenario where AI shrinks demand for creative software is right, today's 13% growth breaks soon. The decline isn't a reaction to weak numbers — it's a bet on that scenario.

Q: Doesn't a slower growth rate mean the business is worse? A: Yes, it does. Revenue grew 11% annually over the last three years and 11.9% over five, against 17% over ten. The deceleration is real. But the investing question isn't whether the business got worse — it's whether the price fell more than the deterioration justifies. I think it did.

Q: Is $500 million in AI revenue actually meaningful? A: In absolute terms it's still small, under 2% of total recurring revenue. What matters alongside the size is the rate — more than tripling in 12 months — and the licensing structure that lets enterprises use it without copyright exposure. Right now the direction and the barrier matter more than the number.

Q: Was the Figma bid evidence of weakness? A: It cuts both ways. Being willing to spend $20 billion can read as recognizing a threat, or as having enough cash to simply buy the threat outright. I lean toward the latter, but I'll grant the former is a reasonable reading.