On Thursday evening Adobe reported results for the third quarter of its fiscal 2026, and by every measure a software company can fit into a press release, it was a good quarter. Revenue rose 13 percent to $6.76 billion, above the $6.69 billion consensus. Adjusted earnings per share came in at $6.13 against the $6.07 analysts expected. Full-year guidance was raised, to $26.58–26.63 billion in revenue and $24.45–24.50 in adjusted earnings per share. More than one billion people now use at least one Adobe product every month. And the company announced that Shantanu Narayen, chief executive since December 2007, will hand over to Anil Chakravarthy on December 1 and become executive chairman.
The stock fell 2.3 percent to $248.95 in Thursday’s regular session, dropped as much as five percent after hours, and closed Friday at $252.23, up 1.4 percent on a day when the broad market gained almost one percent. You can read that as a shrug. You should read it as a verdict. Adobe has lost a third of its market value in twelve months, from a high of $370.86 to a low of $190.12, and at a market capitalisation of $100 billion it now trades at 9.4 times the earnings it is itself guiding to for this year. That is the multiple of a tobacco company or a telephone utility, not of a business growing 13 percent with a 44 percent operating margin. The question of this commentary is therefore not whether the quarter was good. It was. The question is what the market knows, or believes it knows, when it prices a growing company as if it were shrinking.
The number that was not in the headline
Every quarter Adobe reports annualised recurring revenue, or ARR. It reached $27.50 billion, up 11.2 percent year on year. That sounds solid until you compute the prior year: $27.50 billion divided by 1.112 is roughly $24.73 billion, so the increase over twelve months was $2.77 billion. What interested analysts on the call, however, was not the stock but the flow, the net new ARR added in the quarter. And after several firms pressed the point, that flow was 36 to 37 percent below the level of the same quarter last year.
Management did not dispute the number. It explained it. For several quarters Adobe has been deliberately routing a growing share of the traffic on its websites not into the checkout funnel but into free products: Adobe Express, the free tiers of Firefly and Acrobat. Monthly active users of the free creative products passed 100 million in the quarter, up more than 70 percent. Acrobat and Express together exceed 900 million monthly users. The price of that strategy is that people who would once have entered a credit card after a seven-day trial now pay nothing at all, at least for now. Chief financial officer Steven Day said in so many words that the backlog trend “does reflect our focus to accelerate new user acquisition through the freemium business model.” Remaining performance obligations, the contracted backlog, grew eight percent to $22.16 billion, the slowest growth since early 2023, while revenue grew 13 percent. Backlog is tomorrow’s revenue. When it grows more slowly than today’s revenue, the balance sheet is telling you the growth rate is going to fall.
Growth per share and growth of the company are two different things
The second look belongs to the income statement, where there is a discrepancy you only see with a calculator. Adjusted earnings per share rose 15 percent, from $5.31 to $6.13. Adjusted net income, from which that figure is derived, rose only from $2.252 billion to $2.424 billion, a gain of 7.6 percent. The difference is the share count. It fell from 424 million to 395 million diluted shares in twelve months, a reduction of 6.8 percent. Roughly half of the per-share earnings growth in Adobe’s headline came not from the business but from buying back its own stock. In the quarter alone the company repurchased 9.5 million shares for $2.232 billion, an average of about $235 apiece, and the board’s authorisation permits a further $24.55 billion. On a $100 billion market capitalisation, that is a quarter of the company.
On a GAAP basis, that is under US accounting rules and including stock-based compensation, the picture is starker still: net income of $1.827 billion against $1.772 billion, a gain of 3.1 percent, while GAAP earnings per share rose eleven percent. None of this needs to be read as a trick. A buyback at 9.6 times earnings is a good use of cash for the remaining shareholders, far better than most of the acquisitions software companies have made in recent years. But one should know what one is buying: a company whose profit grows in the high single digits and whose earnings per share grow in the mid-teens because it is buying itself.
What the price is saying
At $252 and roughly $24.50 of adjusted earnings per share guided for this year, the earnings yield is 9.7 percent; on the current consensus for next year, according to Stockanalysis data, it is 10.6 percent. The ten-year Treasury yielded just under five percent on Friday, its highest in almost three years, and futures price a rate increase at Wednesday’s Federal Reserve meeting with roughly 90 percent probability. Adobe therefore offers, arithmetically, double the risk-free rate, from a business with a 44 percent operating margin that generated $2.52 billion of operating cash flow in the quarter, a record for a third quarter.
A market that values such a company at 9.4 times is not saying the numbers are wrong. It is saying they are perishable. The multiple implies that earnings will not grow 13 percent but will, within a foreseeable horizon, fall, and not a little but structurally. The narrative behind that is familiar: generative models from OpenAI, Google and a string of smaller vendors produce images, video and layouts that used to require Photoshop, Premiere and InDesign, and increasingly they do so inside tools Adobe does not own. If that is right, the freemium pivot is not offence but defence: Adobe is giving away what it may soon be unable to sell, in order at least to keep the users. And the market is pricing exactly that.
The second transition under the same chief executive
It is worth knowing the story Narayen himself told on Thursday. When he took over in 2007, Adobe sold software in boxes, the Creative Suite, at more than a thousand dollars a licence, with a new release every eighteen months. In 2012 he announced the move to subscription, the Creative Cloud, and in 2013 the company stopped selling boxes altogether. Revenue fell in fiscal 2013, and in 2014 it was still below the 2012 level. The stock stagnated through the transition and analysts wrote of a risky experiment. Since then revenue has risen more than six-fold, and the shares rose more than twenty-fold from the low to the 2021 high. It is one of the most successful business-model transitions in the history of software.
Narayen argues he is now doing the same thing a second time: deliberately forgoing near-term revenue to build a model that is larger in five years. Asked by an analyst why Adobe had deferred planned price increases in Creative Cloud, he replied that he was “actually really happy that we did not focus on the pricing actions,” because while they might have provided some short-term relief, that would not be as critical as continuing to drive new user adoption. Chakravarthy, who joined from Informatica in 2020, where he had been chief executive, and has since run the marketing and customer-data business, added that the company is “first focused on making sure that we are acquiring new users,” and only then on the intensity of their AI usage. It is a consistent argument. It has one catch: in 2012 there was no alternative to Photoshop. In 2026 there is, and part of it is free.
Where the AI business stands, in numbers
Since this year Adobe has disclosed a metric it calls “AI-first ARR,” recurring revenue from products built around generative AI from the outset: Firefly, the Acrobat AI Assistant, GenStudio for automated advertising production. That metric passed $650 million in the quarter and is growing more than 150 percent. Firefly ARR alone rose 40 percent quarter on quarter, and monthly active users of the Acrobat AI Assistant doubled from the prior quarter. In the enterprise business, ARR for the marketing platforms, Experience Manager, GenStudio and Experience Platform, grew more than 20 percent.
Those numbers deserve context. $650 million is 2.4 percent of total ARR of $27.5 billion. If the AI business grows another 150 percent over twelve months, it will be $1.6 billion, or six percent. That is impressive, and it is not yet what carries the company. What carries the company is the $4.65 billion of quarterly revenue from creative and marketing professionals, up 13 percent, and the $1.91 billion from business professionals and consumers, essentially Acrobat, up 16 percent. Both segments are growing double digits. Both are being priced by the market as if they were in liquidation.
What it means for investors, and for Figma, Canva and the rest
The cleanest way to see what the market thinks of Adobe is to look at what it pays for the companies eating at its edges. Figma, which Adobe tried to buy for $20 billion before regulators in Brussels and London forced the deal to be abandoned in late 2023 at the cost of a $1 billion break fee, has traded on the New York Stock Exchange since its 2025 listing at a multiple of revenue that is a multiple of Adobe’s multiple of earnings. Canva, still private, is the freemium model Adobe is now copying, a decade in. And the frontier labs, OpenAI and Google above all, ship image and video generation as a feature of a chatbot rather than as a product. Adobe’s answer, that Firefly is trained on licensed content and therefore safe for commercial use, is a real advantage for the enterprise buyer and a matter of indifference to the hundred million people making a birthday card.
For a US investor, the absence of a dividend matters. Adobe returns all its capital through buybacks, which means no qualified-dividend income to report annually and no tax event until sale; the gain is taxed at long-term capital-gains rates after a year, plus the 3.8 percent net investment income tax above the thresholds. Anyone who sold Adobe during the slide to $190 and wants back in should mind the wash-sale rule’s thirty-day window. And anyone who owns it should be clear that a stock at 9.4 times earnings with a quarter of itself authorised for repurchase is either a value investment or a value trap, and that the difference lies not in the multiple but in whether the earnings hold. The broader software sector has spent 2026 answering that question, mostly downward; Adobe is merely the largest and most profitable company to be asked it.
The counter-argument
The bear case is not weak, and it deserves to be taken seriously. First, fourth-quarter guidance of $6.80–6.85 billion implies only 0.6 to 1.3 percent sequential growth over the third quarter’s $6.76 billion. For what should seasonally be the strongest quarter, that is thin, and it was below what parts of the market expected, which is why the stock fell five percent after hours before recovering. Second, a decline of more than a third in net new ARR is not a rounding error. Management says it is intentional. Intentional and good are not the same thing, and whether 100 million free users ever pay is a bet, not a metric. Third, free users cost money. Every image generated in Firefly consumes compute that Adobe pays for and no customer pays for. The 44 percent adjusted margin is remarkably stable, but it is not a law of nature when users grow 70 percent a year and revenue 13.
Fourth, and this is the point the bulls prefer to skip: a chief executive who ran the company for nineteen years and rebuilt it once successfully is handing over in the middle of the second rebuild to a successor from the data business, not from creative software. That may be the right choice; the agentic era Chakravarthy spoke of is a data problem. But markets usually price transitions at a discount, not a premium. And fifth, a ten percent earnings yield means something different at five percent Treasury yields than at one percent. Part of the compression in Adobe’s multiple is not an Adobe story at all but the story of the entire software sector in a year when the Fed is raising rather than cutting.
How we will know who is right
The dispute between a company that says it is investing in users and a market that says it is giving away its product will not be settled by the revenue line. It will be settled by three figures Adobe will deliver over the next two quarters. First, fourth-quarter backlog: Steven Day himself said RPO and current RPO step up seasonally in the fourth quarter. If the step-up fails to appear, or growth falls below eight percent, the freemium argument is refuted, because then the installed base is slowing too. Second, fourth-quarter net new ARR, the first quarter in which the prior-year comparison already includes the beginning of the freemium redirection. If the decline is still a third, it is not a base effect. Third, the conversion rate of the 100 million: Adobe did not disclose it. If the company discloses it in December, at the new chief executive’s first appearance, that is a sign it is presentable. If not, the market has its reason.
Until then the sober reading stands: Adobe is a company growing 13 percent, earning a 44 percent margin, carrying no net debt worth worrying about, and available at 9.4 times earnings. That is either one of the cheapest software stocks in the world or a company whose earnings will be half their current size in five years. The market, with Friday’s 1.4 percent gain, chose neither reading. It merely noted that a record quarter is not enough to answer the question. That, for all its sobriety, is the real finding: at Adobe, quarterly numbers no longer count. What counts is whether 100 million people who pay nothing today pay something in two years. There is no metric for that answer, only a price, and the price is nine times.
Try TradingView Free for 30 Days
Plus get a $15 discount on your first subscription through this link.
Read more in our topic hub: Topic Hub: Quarterly Earnings Tracker 2026
