Adobe Is Down 57%. The Financials Say Wall Street Is Wrong.
I went looking for evidence of the AI apocalypse. Instead I found a company the market can't properly price.
Chegg’s stock is down 99%. Revenues collapsed. The company turned unprofitable. ChatGPT didn’t merely “pressure” the business or “introduce uncertainty.” It replaced the core use case in plain sight. That’s what AI disruption looks like when it actually arrives: a product that used to be scarce becomes abundant, and the tollbooth operator discovers the road has been rerouted.
Adobe is being priced like that story.
Adobe’s stock is down 57% while the broader market is up 39%, a spread that usually signals either a structural break or a spectacular misunderstanding. But Adobe’s operating reality doesn’t resemble Chegg’s. Adobe’s revenues grew 10% and operating profits grew 29% in fiscal 2025, according to its FY2025 financials. Digital Media ARR hit $19.2 billion, growing 11.5% year-over-year. Management is still talking about double-digit ARR growth heading into fiscal 2026.
So I went looking for the evidence that Adobe belongs in the “AI victim” bucket. I wanted to find the early signs: customer defections, collapsing pricing power, a usage cliff, or at least a credible mechanism explaining why Adobe’s trajectory should soon start rhyming with Chegg’s.
What I found instead was a company that’s already lived through a transition Wall Street swore would break it, and a market that seems increasingly unable to distinguish between companies AI actually destroys and companies AI might strengthen.
The question worth asking isn’t “can AI generate images?” Of course it can. The question is why the market is treating Adobe’s fundamentals like a lagging indicator of doom when, so far, they look like a leading indicator of adaptation.
The Business Most People Don’t Understand
Adobe gets discussed like it’s one product: Creative Cloud. That’s understandable. Photoshop is cultural infrastructure at this point. But the part that matters for the AI debate isn’t just whether a model can spit out a decent image. It’s how Adobe’s products fit together inside organizations that have real budgets, real compliance constraints, and real workflow inertia.
Start with the core engine. In fiscal 2025, Digital Media generated $17.4 billion in revenue, about 75% of total revenue, at roughly 95% gross margins, per Adobe’s FY2025 financials. This is the subscription machine most people associate with the company: Creative Cloud plus Document Cloud, sold as recurring relationships rather than one-time transactions.
In practice, it’s a set of lanes professionals and teams learn to drive on. Creative Cloud is where the creation happens: Photoshop, Illustrator, Lightroom, Premiere Pro, and now Firefly inside the suite. Those tools are not interchangeable for many working professionals because the workflows are not just “make an image.” They’re “make an image that can be revised, versioned, handed off, localized, and shipped.”
Premiere Pro is a useful reality check for the “AI will replace it all” narrative, because it forces you to confront a market structure that has already been competitive for years. Premiere Pro still holds an estimated 35% to 40% market share in professional video editing, according to industry data compiled here. If the product were merely “a timeline editor,” that share would be harder to explain. The stickiness is in the workflows, file handoffs, plugins, and the fact that teams don’t just edit video. They collaborate around it.
Then there’s the second engine that gets underweighted in casual analysis: Digital Experience. In fiscal 2025 it generated $5.4 billion, about 25% of revenue, at roughly 72% gross margins, again per Adobe’s FY2025 financials. This is the enterprise side: the software that helps large organizations manage content, orchestrate campaigns, measure performance, and operationalize personalization.
Digital Experience didn’t appear by accident. It was built deliberately through acquisitions that look, in hindsight, like a long campaign to make Adobe’s creative tools harder to dislodge inside enterprises. Adobe bought Omniture for $1.8 billion, Magento for $1.7 billion, Marketo for $4.75 billion, and Workfront for $1.5 billion, among others, as captured in Adobe’s acquisition history. You can argue about the price tags. What’s harder to argue with is the strategic shape: Adobe wanted to own not only creation, but the management and deployment layer that sits downstream.
That’s the loop most people miss.
A designer creates in Creative Cloud. The organization manages and routes work. The content is deployed through Experience products. Performance is measured. The team iterates. The value isn’t only “this tool makes pixels.” The value is “this tool is embedded in how we run marketing, commerce, and content operations.”
If you believe AI will commoditize creation, this loop becomes more important, not less. When creation gets cheaper, organizations usually produce more variants, not fewer. They need more governance, more approvals, more measurement, more brand consistency, and more integration with the systems that decide where content goes and what it’s supposed to achieve.
With the business structure clear, the AI question becomes less mystical. It turns into something you can actually interrogate: where does AI plausibly break this loop, and where does it reinforce it?
The Competition That Predates AI
One reason I’m skeptical of the “Adobe gets Chegg’d” framing is that Adobe has been “supposed to be disrupted” for a long time. The threats just keep changing costumes.
The most credible low-end pressure point isn’t AI. It’s Canva, which has built a massive business by making design easier, cheaper, and more collaborative for non-designers. Canva is not a toy anymore. Sacra estimates Canva at $3.5 billion in revenue and a $42 billion valuation, and notes its acquisition of Affinity as a move upmarket in its Canva analysis. That’s the real strategic arc: start where Adobe is too complex and expensive, then climb into professional territory as users’ needs mature.
Canva’s threat is not that it beats Photoshop at Photoshop. It’s that it changes who gets to “do design” in the first place. If the person generating marketing assets inside a small business is no longer a freelancer with Creative Cloud but an operator inside Canva, Adobe loses seats at the margin.
Then there’s Figma, the rare competitor that spooked Adobe enough to pull out the checkbook. Adobe tried to acquire Figma for $20 billion, then abandoned the deal after regulators in the EU and UK raised antitrust concerns, as Figma described in its own announcement. Figma’s strength has always been native collaboration and tighter integration with product and engineering workflows. It’s not just a design tool. It’s a shared workspace for how digital products get built.
What’s interesting now is that the relationship isn’t purely adversarial. Many teams use both Figma and Adobe tools depending on the stage of work and the output required. That’s not a victory lap for Adobe. It’s a reminder that creative tooling often becomes a stack, not a single winner-take-all product.
Video editing is another category where the “Adobe is doomed” story should have shown up years ago if it were going to. Editors have choices: Final Cut Pro, DaVinci Resolve, Avid, and fast-growing mobile pipelines like CapCut. Yet Premiere Pro still holds that 35% to 40% professional share estimate, per the same market share compilation. That doesn’t mean Adobe is invincible. It means competition doesn’t automatically translate into collapse, especially when the customer is a team and the output has to move through a pipeline.
Underneath all of this sits a less glamorous but brutally real moat: formats and expectations. PSD and PDF aren’t just file types. They’re coordination standards. Clients ask for editable PSDs. Legal teams expect PDFs. Agencies hand off source files in Adobe-native formats because that’s what the next person in the chain can open without drama.
This is the part that reminds me of the Teams versus Slack dynamic. Slack was, in many users’ eyes, the cleaner product. But Microsoft bundled Teams into Office and won by integration and distribution. Adobe’s suite advantage works in the same direction: the product isn’t only the app, it’s the fact that the app is embedded into an organization’s production system.
These competitive threats are real, and Adobe has had to respond to them. But they’re legible threats. AI is supposed to be different. It’s supposed to be the kind of shift that makes the suite irrelevant.
So I tried to get specific about what, exactly, the market is pricing.
The AI Threat Everyone’s Pricing
Chegg is the cleanest counterfactual because the disruption mechanism is obvious. When a general-purpose AI assistant can answer homework-style questions instantly, the value of a specialized homework-help subscription collapses. Forbes described Chegg’s situation bluntly: the stock is down 99%, revenues are declining, and the company has turned unprofitable as AI competition reshapes the category in this analysis. In other words, the product got substituted, not complemented.
If Adobe were in the same situation, I’d expect to see the early financial symptoms. Instead, Adobe is posting growth and margin expansion. That doesn’t prove the threat is imaginary. It does suggest the market is discounting a future shock that hasn’t shown up in the numbers yet.
When I break down the fear narrative around Adobe, it tends to cluster into three claims.
First: AI tools replace designers entirely. If “make an image” is a prompt, why pay for professional software?
This is the most emotionally compelling claim, and the least operational. In real organizations, “design” is rarely just image generation. It’s brand constraints, legal constraints, localization, version control, approvals, platform specs, and endless revisions. AI can speed up parts of that work, but the work doesn’t disappear. It changes shape.
Second: seat dilution. This is the argument I take most seriously because the math is simple. If AI makes a 10-person design team as productive as 2 people, why would the company keep paying for 10 Creative Cloud seats at roughly $40 a month? Even if the remaining designers are more productive, Adobe could lose revenue if the unit of monetization is the seat.
Third: a unified AI platform emerges that doesn’t need Adobe. The idea here is that the model layer becomes the product, and Adobe becomes a legacy UI on top of a commodity capability.
This is where the story gets more subtle. Adobe’s own positioning emphasizes that there is no single dominant model and that users will want access to multiple models inside their workflows. Adobe frames its approach as a multi-model ecosystem, with Adobe’s models and partner models accessible within its apps, as described in its Firefly and generative AI approach overview. The implication is straightforward: as models commoditize, advantage shifts to whoever owns distribution and workflow.
Seat dilution is the bridge between the bear case and the bull case, because it forces you to ask what Adobe is actually selling.
Adobe is responding by changing the unit of value from “a seat” to “a seat plus usage.” In the Q4 FY2025 materials, Adobe describes a shift toward subscriptions that include usage-based generative credits, typically 2,000 to 4,000 credits per month, with the ability to purchase more, per the Q4 FY2025 transcript. That’s not a cosmetic pricing tweak. It’s an attempt to align monetization with the thing AI increases: output volume.
If AI makes designers more productive, they generate more variations, more assets, more iterations. That consumes credits. The risk shifts from “fewer people need the tool” to “fewer people need the tool, but those who do use it more intensely.” Adobe is betting the intensity effect can offset the seat effect.
The honest answer is that nobody knows the long-run balance yet. But the market’s current pricing often assumes the seat effect is guaranteed and the intensity effect is marketing spin. The early ARR and profit numbers don’t support that assumption.
The next question I had was whether Adobe has ever pulled off a transition like this before, where the market priced catastrophe and the company rewired its model.
It has.
The Transition They’ve Done Before
Before 2013, Adobe’s model looked like a relic from another era: perpetual licenses. Customers paid roughly $1,200 to $2,500 per seat for major software packages, then upgraded on a cycle that was lumpy and often driven by new features rather than continuous value. Piracy was a persistent problem. Revenue was spiky. The business depended on periodic upgrade waves. The arc of that change and the subscription launch mechanics are captured in this overview of Adobe’s cloud transformation.
In 2013, Shantanu Narayen pushed Adobe into Creative Cloud subscriptions, roughly $40 to $60 per month depending on the plan. At the time, the fear wasn’t AI. It was churn. Investors worried Adobe was trading high upfront cash for lower monthly payments and exposing itself to cancellation risk.
But the transition worked. One analysis describes recurring revenue going from near-zero to over 90% of the business, with churn around 5%, as summarized in this business-focused write-up. Whatever you think of those exact figures, the direction is undeniable: Adobe turned a lumpy licensing business into a compounding subscription machine.
That history matters because it shows something about management behavior under pressure. Adobe didn’t defend the old model until it broke. It cannibalized itself before someone else could.
Now the fear has changed. Instead of “subscriptions will kill the business,” it’s “AI will kill the business.” The mechanism is different, but the pattern rhymes: a technological shift changes how value is created and captured, and investors assume the incumbent can’t move fast enough.
Narayen is leaning into the idea that Adobe is not being displaced by AI, but pulled deeper into it. In Adobe’s FY2025 earnings release, he framed the year like this:
“Adobe’s record FY2025 results reflect our growing importance in the global AI ecosystem and the rapid adoption of our AI-driven tools. By advancing our innovative generative and agentic platforms... we are excited to target double-digit ARR growth in FY2026.”
I don’t read that as fluff. I read it as a statement about where Adobe thinks the battleground is: not whether AI exists, but whether Adobe becomes the interface through which AI gets used in commercial creative work.
If you accept that premise, the next place you have to look is Firefly, because that’s where Adobe’s AI strategy becomes concrete, monetizable, and legally exposed.
Firefly and the Commercial Safety Moat
If you want to understand why enterprises might choose Adobe’s AI tools even when cheaper or flashier generators exist, you have to look at a constraint that becomes existential at scale: IP risk.
Adobe’s Firefly approach is explicitly built around commercial safety. Adobe says Firefly is trained on Adobe Stock, public domain content, and licensed content, as laid out in its generative AI approach page. The point is to mitigate the legal uncertainty that hangs over models trained on scraped data. In a consumer context, people tolerate ambiguity. In an enterprise context, ambiguity becomes a procurement blocker.
Firefly’s second advantage is that it lives inside the tools where professionals already work. Generating an image is only the first step. Enterprises need post-generation editing, compliance review, resizing, localization, and workflow integration. AI output often requires correction. The correction happens in professional tools.
The third advantage is volume. AI doesn’t just create one asset. It creates a thousand variations. That’s valuable only if you can test, measure, and deploy at scale. This is where Adobe’s Digital Experience segment loops back into the story. Creative output becomes more abundant, and the management layer becomes more valuable.
What I found most important in Adobe’s own commentary is that the company is already tying AI to revenue, not just demos.
On the Q4 FY2025 earnings call, CFO Daniel Durn put a hard marker down:
“Total new AI-influenced ARR now exceeds one-third of our overall book.”
Narayen reinforced the direction of travel on the same call:
“In fiscal 2025, Adobe delivered significant AI-influenced... ARR, which accelerated through the year.”
Adobe also pointed to usage momentum. Firefly credit use tripled quarter-over-quarter, and Adobe referenced enterprise traction for Firefly Services with deals involving Coca-Cola and IKEA, per the Q4 FY2025 transcript.
None of this proves Adobe is immune to disruption. But it does establish something that the market’s “Adobe is Chegg” pricing tends to ignore: Adobe is not standing still while AI eats the category. It’s actively trying to become the monetization layer for AI-driven creative work, with a positioning that matters specifically to enterprises.
At this point, the bullish case almost writes itself. That’s exactly when I force myself to slow down and list the risks that are real, not rhetorical.
The Risks
The easiest way to get blindsided in an AI narrative is to assume the only threat is a better model. In practice, distribution shifts and platform power matter more than model quality.
One real pressure point is the Meta and Google ecosystem for small businesses. These platforms increasingly automate ad creation, targeting, and testing inside their own walled gardens. For a small business focused on direct response and return on ad spend, the workflow becomes: feed the platform a few inputs, let it generate variants, and let it optimize delivery. That can erode demand for freelance design work, which in turn can erode entry-level Creative Cloud usage.
This doesn’t kill Adobe’s enterprise base. But it can hollow out the bottom of the funnel where future professionals and small agencies are born. Over time, that matters.
Another risk sits in the data layer that powers Digital Experience. Modern architectures like zero-copy data warehouses promise app-agnostic access to data without forcing everything into one vendor’s stack. That could weaken the “behavioral data moat” for experience platforms if customers can more easily swap orchestration layers.
Adobe’s response so far looks like adaptation rather than denial. Snowflake and Adobe announced a partnership aimed at customer experience workflows, described in Snowflake’s partnership post. That suggests Adobe is choosing interoperability rather than fighting the trend. But the strategic risk remains: if the data layer becomes more portable, experience platforms can become more substitutable.
Then there’s seat dilution again, because it’s the risk that can be true even if Adobe’s products remain best-in-class. AI can make teams smaller. That is not a moral argument. It’s a budget reality. Adobe’s credit model is designed to capture value from higher output, but the balance between fewer seats and more usage-based revenue is genuinely unclear. It will only become clear over time as customers renew and as Adobe discloses more about how consumption evolves.
Finally, I pay attention when smart investors walk away. Dev Kantesaria has publicly said he sold Adobe because he foresees obsolescence except at the high end, in this public commentary. I don’t treat that as proof. But it’s a credible articulation of the bear thesis: AI collapses mid-market need for professional tools, leaving Adobe as a premium niche rather than a broad platform.
These risks don’t negate the financial performance. They explain why the market might be willing to discount the future so aggressively. Which brings me to the part of the story where the narrative meets the numbers.
What It Means
The market is currently pricing Adobe like an AI casualty. The business is behaving, at least so far, like an AI beneficiary.
That tension is the whole case study. Adobe’s stock has fallen 57% while the broader market rose 39%, per this market analysis. Yet Adobe’s fiscal 2025 performance shows 10% revenue growth and 29% operating profit growth, per Adobe’s FY2025 financials. Adobe also reported $7.1 billion in net income, $7.9 billion in free cash flow, and $9.5 billion in buybacks over the trailing twelve months in the same FY2025 materials. That is not what a collapsing franchise looks like.
The multiple captures the market’s skepticism. Adobe has traded around 14x trailing earnings/FCF, per that same Forbes market analysis. For a software company guiding to double-digit ARR growth, that’s a valuation that implies either a coming growth cliff, a durable margin hit, or a belief that the business model is about to be structurally impaired.
And yet, inside Adobe’s own numbers, AI is already showing up as a revenue contributor. Durn’s line that “Total new AI-influenced ARR now exceeds one-third of our overall book” forces a simple update: if AI were purely corrosive to Adobe’s monetization, this number should be small or shrinking, not large and prominent.
Even outside Adobe, some market observers are noticing the mismatch between narrative and metrics. Investment analyst Shay Boloor put it this way:
“No evidence GenAI is eroding Adobe’s core... Digital Media ARR ~12%”
That quote matters less as “authority” and more as a concise testable claim. If AI is eroding the core, where is it in the ARR line?
This is where I landed on a framework that feels more useful than arguing about whether AI is “good” or “bad” for Adobe.
AI disruption is priced correctly when the AI product directly substitutes the core use case and customers can switch without friction. That’s Chegg. The end user got a cheaper, faster substitute that met the need well enough. Switching costs were minimal. The financials broke quickly.
AI disruption is often mispriced when AI commoditizes a capability but increases the value of distribution, workflow integration, and commercial safety. That’s closer to Adobe’s situation. If models become interchangeable, the advantage shifts to the platform that sits inside the workflow, has enterprise relationships, and can offer legal comfort.
Who benefits in that world? Companies with distribution and governance, not necessarily the companies with the best raw model. Adobe’s combination of a commercially positioned model (Firefly), integration into professional workflows, and an enterprise deployment layer through Digital Experience is a plausible “AI as a feature” winner.
Who gets pressured? Pure-play generation companies without enterprise distribution or clear commercial safety positioning. Freelance designers serving small businesses that increasingly rely on Meta and Google’s automated ad tools. And companies like Chegg where AI is not an enhancement but a replacement.
What to Watch
I don’t think the right way to hold this story is as a permanent verdict. It’s a live question that will resolve in the numbers.
ARR growth trajectory. Adobe guided to double-digit ARR growth, and the current growth rate was 11.5% for Digital Media ARR in FY2025. If Adobe’s ARR growth decelerates below high single digits for two consecutive quarters, that’s a meaningful tripwire.
Firefly credits versus seats. The key uncertainty is the balance between seat dilution and usage-based revenue. The tell will be whether credit adoption grows fast enough to offset any seat pressure. Adobe doesn’t fully disclose the ratio today, but any future disclosure or proxy metrics will matter.
Canva’s enterprise penetration. Canva’s move upmarket is the most credible product threat to Creative Cloud’s breadth. If Canva starts winning enterprise workflows at scale, that’s when Adobe’s stickiness story gets tested.
Meta and Google automation adoption among small businesses. This is the quiet channel conflict that can reshape the low end of the creative economy. If small businesses stop hiring designers, the cultural pipeline into Adobe weakens over time.
There’s also a structural uncertainty I don’t ignore: whether AI partners could ever cut off Adobe’s access to models. Adobe’s positioning assumes a multi-model world where users want choice. If the model layer becomes more closed, distribution advantages can get constrained. What would resolve that uncertainty is clear behavior: do leading model providers keep expanding partnerships and integrations, or do they retreat into vertically integrated stacks?
Chegg’s business got replaced. Adobe’s business, so far, is getting augmented. The market priced both the same way.
The interesting question is whether the market can distinguish between companies AI actually destroys and companies that might be strengthened by it. Right now, it can’t. That’s either a buying opportunity created by narrative panic, or a warning that the financials haven’t caught up to the disruption yet.
The financials will tell us which.