Adobe's AI Paradox: The Business Is Accelerating While the Stock Prices a Retreat

Adobe has become one of the market's strangest AI stories.

The company is growing revenue at a double-digit rate, generating prodigious cash, shrinking its share count, and putting AI assistants inside the creative tools professionals already use. Yet investors continue to value $ADBE less like a durable software franchise and more like a business approaching structural decline.

That disconnect is the entire investment case. Adobe does not need to win every corner of generative AI for the stock to work from here. But it does need to prove that AI makes Creative Cloud, Acrobat, and its marketing stack more valuable—not merely easier to replace.

As of July 24, the evidence is encouraging. It is not conclusive.

The latest quarter was stronger than the stock's reputation#

Adobe's fiscal second-quarter revenue reached a record $6.62 billion, up 13% year over year, or 11% in constant currency. Subscription revenue rose 14% to $6.39 billion. Both major customer groups grew at double-digit rates: Business Professionals and Consumers increased 16%, while Creative and Marketing Professionals increased 13%.

Those are not the numbers of a franchise already being dismantled.

Total annualized recurring revenue, or ARR, ended the quarter at $27.10 billion, up 12.5%. The recently completed Semrush acquisition contributed about $480 million, so investors should not treat all of that increase as organic. Excluding the acquired ARR, Adobe still grew the base by roughly 10.5% on a comparable basis.

The company also raised its fiscal 2026 outlook to $26.50 billion to $26.60 billion of revenue and $24.35 to $24.45 of non-GAAP earnings per share. At a share price around $224 at the time of writing, the stock trades near 9.2 times the midpoint of that earnings guidance. Even on the midpoint of GAAP guidance, the multiple is about 12.5 times.

That is an unusually low valuation for a subscription software company still posting this combination of growth, margins, and cash generation. It is also the market's way of saying that today's earnings may not be durable.

AI revenue is finally visible—but still small#

Adobe says AI-first ARR more than tripled from a year ago and passed $500 million in the second quarter. The growth rate is impressive. The scale requires context.

AI-first ARR represents less than 2% of Adobe's total ARR today. It is too early to argue that direct AI monetization has transformed the income statement. What it does show is that Adobe is no longer discussing AI only through product demos and engagement metrics. There is now a measurable revenue stream, and it is compounding quickly.

More important, the reported figure likely captures only part of AI's economic impact. Generative features can support pricing, retention, new-user conversion, and higher usage inside broader Creative Cloud or Acrobat subscriptions without being classified as standalone AI-first ARR.

This distinction matters. If investors wait for every dollar influenced by AI to appear in a neatly labeled revenue bucket, they may miss the larger effect. Adobe's real opportunity is not simply selling image-generation credits. It is using AI to expand the number of people who can produce professional work while keeping demanding users inside an editable, governed workflow.

Adobe is turning the interface into an agent#

In June, Adobe expanded its creative agent across Firefly and public betas in Photoshop, Premiere, Illustrator, InDesign, and Frame.io. The practical examples are more interesting than the phrase "agentic AI."

• Premiere can organize footage, rename clips, identify interview questions, add markers, and assemble a starting cut.
• Photoshop can execute multi-step edits across a composite while preserving an editable result.
• Illustrator can generate large batches of versioned files and run production checks.
• InDesign can apply a new brand system across layouts and check print readiness.
• Firefly can build brand kits, short product videos, storyboards, and reusable project elements through a conversational interface.

Adobe is also distributing its tools through ChatGPT, Claude, Microsoft Copilot, Gemini, and Slack. That is strategically important. The company appears willing to let the conversational front door belong to somebody else as long as Adobe remains the trusted execution layer underneath.

This may be the best response to the threat from general-purpose models. A chatbot can generate a striking first draft. Professional work still requires consistency, revisions, color management, rights controls, collaboration, asset history, and delivery across dozens of formats. Adobe's bet is that AI compresses the tedious steps without eliminating the system in which the finished work gets made.

The Topaz deal reveals where Adobe sees the bottleneck#

Adobe's June 25 agreement to acquire Topaz Labs is small relative to the company, but strategically revealing. Topaz specializes in enhancing existing video and images—upscaling, sharpening, stabilization, noise removal, frame interpolation, and restoration.

Those capabilities sit at the boundary between generated media and production-quality output. They also complement Adobe's own models rather than forcing the company to compete in a pure race for the largest foundation model.

Topaz brings another useful asset: technology for running complex models locally on consumer devices. On-device inference can improve speed, lower cloud-compute costs, and keep sensitive creative assets closer to the user. If the deal closes as expected in the second half of 2026, Adobe gains both quality-enhancement models and a more efficient way to deploy them.

The acquisition fits a broader pattern. Adobe is assembling the workflow around AI—generation, editing, enhancement, brand control, distribution, and measurement—not betting the franchise on any single model.

The cash engine gives Adobe time#

For the first six months of fiscal 2026, Adobe generated $5.12 billion of operating cash flow and spent just $95 million on property and equipment. That works out to approximately $5.03 billion of free cash flow before acquisitions.

Adobe used $4.59 billion to repurchase shares during the same period. The diluted share count in the second quarter fell to 402 million from 429 million a year earlier, a decline of roughly 6%. In April, the board added a new $25 billion repurchase authorization; approximately $26.8 billion remained across the company's authorizations at the end of May.

Buybacks do not fix a deteriorating product. They are highly valuable, however, when a healthy business can retire stock at a depressed multiple. Adobe's cash flow gives management room to fund AI development, make targeted acquisitions, and reduce the denominator in per-share results at the same time.

There is one important caveat: stock-based compensation was $1.04 billion in the first half. Investors should focus on the actual decline in diluted shares—not merely the headline dollars spent—to judge whether repurchases are creating value after employee dilution. So far, the net share reduction is substantial.

Why the bear case still deserves respect#

The low multiple is not a free gift. Adobe faces several risks that could make it entirely rational.

• Creation is being commoditized. General-purpose models can produce good-enough images, copy, layouts, and video for users who once needed specialized software or a professional creator.
• AI-first challengers have a simpler starting point. New products do not need to preserve decades of workflows, interfaces, and file compatibility.
• Organic growth is slower than the headline. Semrush added to ARR and will add to reported revenue, so the underlying trend must be separated from acquisition effects.
• AI economics remain uncertain. Model inference can carry real costs, and Adobe must balance generous usage with the high margins investors expect.
• Leadership is in transition. Shantanu Narayen plans to step down as CEO once a successor is named, while CFO Dan Durn departed in June. A simultaneous CEO search and interim CFO add execution risk at a decisive moment.
• Direct AI monetization is not yet large. More than $500 million of AI-first ARR is promising, but it remains a small piece of a $27 billion recurring-revenue base.

The deepest risk is not that Photoshop disappears next year. It is that a growing share of casual creation happens elsewhere, weakening Adobe's funnel while price competition gradually reaches the professional tier.

What would change the argument#

Investors do not need to guess whether every new AI feature will succeed. A handful of operating signals can test the thesis over the next several quarters.

• Organic ARR growth: Can Adobe sustain roughly 10% growth after removing acquisition contributions and currency effects?
• AI-first ARR: Does the figure continue to multiply fast enough to become material to total growth?
• Creative retention and pricing: Do agentic features support renewals and premium plans without driving customers toward cheaper alternatives?
• Margins: Can Adobe absorb inference and infrastructure costs while holding non-GAAP operating margin near its roughly 45% target?
• Net share count: Do buybacks keep producing real per-share shrinkage after stock compensation?
• Leadership: Does the next CEO sharpen the product strategy without disrupting a remarkably profitable operating model?

A decisive slowdown in organic ARR paired with margin compression would support the bear case. Continued double-digit organic growth, rising AI revenue, and a falling share count would make today's valuation increasingly difficult to justify.

The bottom line#

Adobe is not a simple "cheap AI winner." It is a mature, enormously profitable software company attempting to rebuild its interface and distribution model while new competitors attack the easiest parts of creative work.

The market is focused on what AI can replace. Adobe is trying to prove that the larger opportunity is everything AI enables people and companies to create, revise, govern, and publish.

At roughly nine times updated non-GAAP earnings guidance, investors are not paying for a flawless transition. They are paying for a business the market suspects may decline. The latest results show the opposite: double-digit growth, rising guidance, expanding AI revenue, and more than $5 billion of first-half free cash flow.

That makes $ADBE unusually compelling—but not because the AI risk has vanished. It is compelling because the price assumes far more damage than the financial statements currently show.

Market data is a snapshot from July 24, 2026. This article is for informational purposes only and is not investment advice.