Adobe

IThe business

Adobe makes the programs that turn an idea into an image, a video, a document or a web page, and the systems that big companies use to deliver that content to customers. Photoshop, Illustrator and Premiere are in the first group. Acrobat and the PDF format are in the second. Adobe Experience Manager and Adobe Experience Platform, which run the websites and customer data of about 20,000 large companies, are in the third.⁸¹⁰ It employed 31,360 people at the end of fiscal 2025, half of them in the United States.⁸ Its year ends on the Friday nearest 30 November, so "fiscal 2026" ends on 27 November 2026.⁸

Revenue comes almost entirely from subscriptions. In fiscal 2025 they were $22.9B of $23.8B, or 96%, and product licences and services were the remainder.⁸ Since the start of fiscal 2026 the company reports two customer groups, not three divisions. Business Professionals & Consumers sells Acrobat, Adobe Express and the free Reader, and had $6.50B of subscription revenue in fiscal 2025, up 15%. Creative & Marketing Professionals sells Creative Cloud, Firefly and the enterprise marketing products, and had $16.30B, up 11%.¹⁶ The 10-Q now reports a single segment.³ About 59% of revenue comes from the Americas, 27% from Europe, the Middle East and Africa and 14% from Asia-Pacific.⁸

The picture over ten years is one of steady compounding. Revenue grew from $5.9B in fiscal 2016 to $23.8B in fiscal 2025, 17% a year, and operating income grew faster, from $1.5B to $8.7B.⁴ The record has one odd year. Fiscal 2013, the year Adobe moved its main creative products to subscription, showed revenue of $4.06B, down from $4.40B.⁴ Growth then resumed every year. That history is why investors give Adobe credit for surviving a change in how it sells, and why they ask whether the next change is the one it can't handle.

The numbers that matter

Adobe's headline measure is total annual recurring revenue, ARR: the yearly value of its subscription contracts. It was $27.50B at 28 August 2026, up 11.2% on the same date a year earlier.³ Each year the company revalues the figure at fixed exchange rates, so the starting book for fiscal 2026 was $25.66B, not the $25.20B first reported.¹⁶ The company targets 10.2% growth by year end.² That would put ARR near $28.3B C (calculated).

Look at what is inside the 11.2%. Semrush, a search-visibility software company bought for $1.87B in cash on 28 April 2026, added about $480M of ARR, and the company says so.³⁵⁹ That is 1.9 points of growth. Take it out, as a rough step that assumes Semrush has not grown since, and ARR growth was about 9.2% in the third quarter and 10.5% in the second, against 11.5% for fiscal 2025 as a whole E (estimate).¹⁶¹⁷³ The year-end target of 10.2% is about 8.3% without Semrush E (estimate).

Net new ARR, the amount added in a quarter, tells the same story. Between the second and third quarters ARR rose $0.40B, from $27.10B to $27.50B. A year earlier, on the same exchange rates, it rose $0.66B, from $24.08B to $24.74B C (calculated).¹⁷³ That is 39% less on figures rounded to the nearest $10M. An analyst on the call put the fall at 36% to 37%, which is close. To reach 10.2% by year end Adobe must add about $0.78B in the fourth quarter, nearly twice the third C (calculated).¹⁰ The company says the fourth quarter is its strongest, and the fourth quarter of fiscal 2025 was a record for net new ARR.¹⁰²⁰

The second measure is remaining performance obligations, RPO: contracted revenue not yet recognised. It was $22.16B at 28 August, up 8% from $20.44B a year earlier, and 67% of it is due within a year.²³ An analyst on the call noted it was the first single-digit growth since early fiscal 2023, and management said the pattern across a year is a step up in the fourth quarter and little change afterwards.¹⁰ RPO was $22.52B at the end of fiscal 2025, so it has fallen $0.36B in three quarters.¹⁶²

What changed this year

Four things happened, and the first three are about people, not products. On 12 March, with the first-quarter results, Narayen said he would leave the chief executive role once a successor was found. The stock closed 7.6% lower the next day C (calculated).²¹¹ On 8 June the chief financial officer, Daniel Durn, told the company he was resigning, effective 15 June, and on 11 June the company named Steven Day, a senior vice president, interim CFO.⁷ A press report says Durn is joining Marvell Technology as its finance chief.²² Shares fell 6.2% on 11 June and 6.8% on 12 June C (calculated), and reached a closing low of $193.41 on 25 June.¹ On 2 September the board appointed Anil Chakravarthy to succeed Narayen from 1 December, and on the same day David Wadhwani, who led the Creativity and Productivity business, said he was stepping down.⁶ The stock fell 6.7% on 4 September C (calculated).¹

The fourth thing is about the business. At the second-quarter call, management told investors that it would put free products first and defer planned changes to Creative Cloud price lines, which an analyst on the call described as a possible price rise. Narayen said about half of the hit to second-half ARR came from deferring those price changes and half from the push to free users.⁹⁵ The target for ARR growth stayed at 10.2% only because Semrush was added to it.⁵ I treat that as the most important disclosure of the year, and Section VII returns to it.

The stock has recovered 29% from the June low, to $250.23, and is still 40.5% below the 19 May 2025 close of $420.68 C (calculated).¹

The bet the company has made

Adobe says AI widens its market. Its argument is that there are over a billion people a month using its products, 900 million of them on Acrobat and Express, up more than 25% in a year, and that free access to its tools will turn some of them into paying customers.²¹⁰ The creative free tier passed 100 million monthly users, up more than 70%.¹⁰ AI-first ARR, from products built around AI, such as the Firefly app and credit packs, passed $650M and grew more than 150%.²¹⁰ For scale, that is 2.4% of ARR C (calculated).

It also changes how it sells. Instead of only selling seats, it sells credits for AI generation and, in the enterprise, usage and outcome-based contracts.⁹ Acrobat now runs inside ChatGPT, Chrome, Microsoft Edge and WhatsApp, which means Adobe is trying to be the engine behind other companies' front doors as well as its own.¹⁰ I read that as sensible and as an admission. The front door of creative work may no longer be an Adobe window.

On the cost side, the company is spending more. In the third quarter research and development rose 18% and general and administration 20%, against revenue growth of 13%.² GAAP operating margin was 34.8%, against 36.3% a year earlier, and non-GAAP margin was 44.0%, against 46.3% C (calculated).² The FY2026 target is about 45%, against 46.2% in fiscal 2025.²¹⁶

Two smaller items

In March 2026 Adobe settled the Department of Justice lawsuit, filed in 2024, that alleged it did not clearly disclose subscription terms and made cancelling hard. It recorded a loss contingency tied to a legal settlement in the first quarter, and I did not find the settlement terms in the 10-Q.³¹⁹ A shareholder suit over the Figma competition is on appeal after being dismissed.³ I don't treat either as large in dollars. The cancellation case matters more for how Adobe is allowed to sell subscriptions.

IIThe moat

A durable business keeps its customers and its prices without spending more each year to do it. Adobe's record on that test is among the best of any software company. The test it now faces is different: the main product may soon be one a free AI tool does reasonably well.

Evidence it is wide

The first evidence is margin. Gross margin was 89.3% in fiscal 2025, and GAAP operating margin has risen from 25.5% in fiscal 2016 to 36.6% C (calculated).⁴ Non-GAAP operating margin, which leaves out stock pay and deal amortization, was 46.6% in fiscal 2024 and 46.2% in fiscal 2025.¹⁶ A business without pricing power doesn't keep margins like that for ten years while growing 17% a year.

The second is habit. Photoshop files, Illustrator files, the PDF itself and Premiere projects are the working formats of whole industries. A design team's templates, plug-ins and brand libraries live in them. Switching is a retraining cost, a file-conversion cost and, for a studio, a risk to a client deadline. For enterprises the cost is higher: Adobe Experience Platform holds customer data, and the company says retention across the enterprise base remains strong.¹⁰ It doesn't publish a customer retention rate or a seat count, which I regret, because the answer to the main question sits there.

The third is the way the company uses its power. It had planned pricing changes on Creative Cloud for fiscal 2026 and put them off. Narayen said Adobe is deferring them, not dropping them, and that he is confident the creative business is stable.⁹ A company without pricing power doesn't have a rise to defer. What it tells me is that Adobe thinks the price lever works and chose to give up growth now for users later.

Evidence it is narrowing

The first sign is growth. Organic ARR growth has slipped from 11.5% to about 9% in three quarters, net new ARR fell 39% from a year earlier in the third quarter, and the company's own target for the fiscal year implies the slowing continues.¹⁶³¹⁷ The company gave reasons: the price deferral and the push to free users. I accept the reasons. They are also the signs of a business that can no longer raise price and grow at the same time without something else giving.

The second is cost. Cost of subscription revenue rose 24% in the third quarter, against subscription revenue up 14%.² That moved subscription gross margin from 91.2% to 90.4% C (calculated). It is a small move. It is the kind that AI products cause, because every image or video a user generates costs real compute, and a free user costs the same as a paid one. Management says it is not cutting spending on models, cloud or marketing to protect margin.⁹

The third is the product itself. Adobe's 10-K says its markets have limited barriers to entry, short product cycles and customer price sensitivity, and lists AI-first creative tools among its competitors.⁸ Firefly, Adobe's own image and video model, now sits beside models from other companies that Adobe also offers inside its apps. A company that includes its rivals' models in its own product knows that models are not where it wins. It wins, if it wins, on the workflow around them.

The fourth is a smaller product line. The 10-Q says ARR growth in the first quarter was "partially offset by a decrease from Adobe Stock", the library of licensed images, which an image generator can substitute for, in my view.¹⁹ In the second quarter Adobe wrote off $70M of goodwill in its publishing and advertising unit.³ Both are small. I mention them because they are where AI has already shown up in the numbers.

The question no filing can settle

A professional using Photoshop pays a monthly subscription. The question is what happens when a model can make a good-enough version of the same output from a text box, free or for a few dollars. Three outcomes are possible. The professional keeps paying for control and precision, and Adobe keeps its price. The professional pays less, because part of the work moves to cheaper tools. Or the work shrinks: fewer designers, fewer editors, fewer seats. I can't know which, and neither can Adobe, which is why the company is spending its next year's growth to buy a larger audience.

My judgement: the switching costs are real, the margin record is real, and the evidence that they protect the price for the next five years is thin. If the first outcome holds, my base case is conservative. If the third does, my bear case is closer than it looks. The company's own shift from raising price to growing users is what I would watch, because it is the signal that comes from the people who see the renewal data.

IIIThe competitive advantage

The 10-K lists no single rival. It says Adobe competes with software companies, AI companies, hardware makers, operating-system developers and social media companies, and it names the groups: productivity tools and design platforms for business users, professional creative suites and AI-first tools for creators, and enterprise software and clouds for marketers.⁸ Below I compare Adobe with the listed software companies that sit closest, using the same simple owner measure for each.

The table says Adobe is the cheapest of the seven on owner cash, at 8.6% of its market value, and has the highest owner margin, 32.8%, while growing 12.9% in its latest quarter, behind Microsoft, ServiceNow and Autodesk. Intuit, which came from the same screen, sits at 7.0% and Salesforce at 6.1%. ServiceNow and Workday, which the market has not marked down as far, sit at 2.2% and 2.5%. Investors are paying less for each dollar of cash from the companies they think AI can reach. That is a belief, not a fact. It is still a large gap.

Where Adobe wins

Breadth first. I know of no other company with leading products in each of imaging, video, design, documents and enterprise content. A marketing team can plan a campaign, make the assets, check them against the brand, publish them and measure the result inside one vendor. Management calls this the content supply chain, and says its ARR for the products that handle it, including Experience Manager and GenStudio, grows above 20% a year.¹⁰

Distribution second. Over a billion monthly users, 900 million of them on Acrobat and Express, is the largest top of funnel in the category.¹⁰ A start-up with a good model has to buy that audience. Adobe sends search traffic into a free tool and tries to convert some of those users into subscribers. It is paying with near-term ARR for this, which I count as a cost.

Third, trust. Adobe trains Firefly on licensed and public-domain content and sells that as "commercially safe" for companies that fear copyright claims.⁸ Management named Disney, Publicis and the Premier League among its enterprise wins in the last quarter.¹⁰ Whether this survives a world where AI-generated images are routinely used is something I can't judge, but in the large-company market it is a real selling point today.

Where it loses

Figma is the clearest proof that Adobe can lose. Figma is the design tool Adobe tried to buy; Adobe walked away and paid a $1B termination fee in fiscal 2024.⁸ Figma is now listed. Its revenue was $1.28B in the twelve months to June 2026, and $370M in the June quarter against $250M a year earlier, growth of 48% C (calculated).²⁶ That is small next to Adobe's $26B. It shows that a focused rival can win a creative job from Adobe at scale.

Canva, Midjourney, and the image and video models from OpenAI, Google and others attack from below. Canva is private and Midjourney does not report, so I can't give numbers. I list them because Adobe's own 10-K does: new entrants, AI-first tools and creative features built into broad platforms.⁸ Microsoft and Google bundle design and AI into suites that businesses already pay for. In the enterprise marketing market Adobe competes with Salesforce, Oracle, HubSpot and the large clouds, and its own 10-K says internally built apps are a risk.⁸

The deepest challenge is the one in Section II. If an AI assistant, not an Adobe screen, becomes the place where a user starts a task, Adobe's tools are one of several that the assistant can call. That would still be a business. It would not be the one that earned 89% gross margin from selling the screen.

IVManagement

This is the section where I most want to be careful, because the company has changed its chief executive, its finance chief and the head of its largest product group in six months. I start with what changed, then with how the people are paid, because pay says what they are being asked to do.

Narayen has been chief executive for eighteen years. He told staff the company grew from under $1B to over $25B of revenue and from about 3,000 to over 30,000 people on his watch.²¹ That record is among the best of any software chief executive, and the board has kept him close: he stays as executive chair. Chakravarthy joined in January 2020 to lead the enterprise marketing business, took on the global sales force that year, and was president of that business from December 2021. Before Adobe he spent four years as chief executive of Informatica, a data software company.⁶ Two things stand out. Both finalists for the job were inside the company, Narayen said on the call.¹⁰ And Chakravarthy comes from the part of Adobe that sells to enterprises, while the creative business, where the AI question is sharpest, lost its president.⁶ An analyst asked on the call whether that choice signals where the board thinks growth lies. Narayen said the new chief executive will own the whole strategy, and Chakravarthy said he sells all of Adobe to its largest customers.¹⁰ I note the question and can't judge the answer.

The market's first verdict was negative. The stock fell 6.7% on the day after the announcement, and went on falling, from $285.75 on 3 September to $248.83 on 10 September C (calculated).¹ I don't read much into that. Prices move for many reasons in a week with an earnings report.

How they are paid

Narayen's pay for fiscal 2025 was $51.2M, 217 times the median employee's $235,989. Of it, $45.4M was stock and $3.1M a cash bonus.²⁸ The bonus pays only if Adobe reaches 95% of its revenue target and 95% of its earnings per share target, and pays up to 155% of target.²⁹ The stock is in performance shares, and they are well designed in one way. Half pay out on the company's total return against the Nasdaq-100 over three years, with nothing below the 25th percentile and a cap at target if the absolute return is negative. The other half pay on net new sales, which the committee defines as growth in total ARR against a target set each year. That measure leaves out the effect of material acquisitions and exchange rates.²⁹ So material deals such as Semrush are meant to be excluded from it. I give the committee credit for that.

Now look at what the plan pays them to do. It pays them to grow revenue and ARR, to beat 100 large Nasdaq companies on share price over three years, and to raise earnings per share. The earnings per share goal is the one to watch, because buybacks raise it and stock pay does not reduce it. No measure charges management for the capital it uses, whether on buybacks or deals. I found no return on capital, no cap on stock pay as a share of revenue and no cash flow per share in the plan.²⁸²⁹ The relative return test also helps in a bad market for software: if every company falls and Adobe falls less, it can pay.

Shareholders noticed. At the April 2026 annual meeting 148.8M shares voted for the pay of the named executives, 145.0M voted against and 0.7M abstained, so 49% of the votes for or against opposed it.¹⁴ A year earlier about 80% approved.²⁸ It passed by 3.8M shares. Four shareholder proposals all failed, and the director who attracted the most opposition, Daniel Rosensweig, received 90.7M votes against, 31% of those cast for or against him.¹⁴ I found no announced change to the plan after the vote. The January 2026 program already ran on the same two measures.²⁹

How much of their own money is in

The proxy lists 438,975 shares for Narayen at 17 February 2026, of which 430,352 were in a family trust.²⁸ Since then he has sold 200,000 shares: 75,000 on 28 April at about $243.50 and 125,000 on 16 and 17 September at about $250, for $49.6M in all C (calculated). His filings give "tax and estate planning" as the reason.³⁰ The trust held 238,724 shares after the September sales, and about 244,000 after a vest on 15 October, worth about $61M at $250.23 C (calculated).³⁰¹ That is about 40 times his $1.5M salary, against a guideline of 20 times.²⁸ He made no open-market sales in 2025 that I found. I have no evidence the sales mean anything about his view. The September sales came less than two weeks after the new chief executive was named, six days after the third-quarter report, and after the stock had fallen 12% from its 3 September close C (calculated).¹

Chakravarthy held 61,788 shares in the proxy and about 64,000 after his October vest, worth about $16M C (calculated). His guideline is 10 times a salary of $0.8M.²⁸ He has made no open-market sale in the period I checked.³⁰ Steven Day, the interim CFO, holds about 8,000 shares.³⁰

The directors have put in some of their own money at low prices. David Ricks, a director and the chief executive of Eli Lilly, bought 10,000 shares at $194.51 on 25 June 2026, about $1.9M, on the day of the 52-week low close; he had bought 2,250 shares at about $444 in January 2025.³¹²⁸¹ Durn bought 1,300 shares at about $390 in March 2025 and sold 1,646 shares at $294.85 in January 2026.³² These are small amounts in a company worth $98B. They show that at least one director thought $194 was cheap. I can't tell more than that.

The people running Adobe own very little of it as a share of the company: the proxy lists all directors and current executive officers together at 803,767 shares, 0.2% C (calculated).²⁸ BlackRock and Vanguard hold 9.3% and 10.2%.²⁸ I checked the June 2026 13F filings of fifteen well-known investors, and the March filing of Pershing Square, whose June filing I could not find. Only Polen Capital held any Adobe: 7,581 shares worth $1.55M, down from 11,868 in March.³³

What they did with capital

Here I look at the full record, not one good year. From fiscal 2016 to 2025 Adobe produced $56.77B of free cash flow and paid $11.11B of it in stock pay, which leaves $45.66B for owners C (calculated).⁴ Over the same ten years it spent $45.71B on buybacks, the same amount, and $4.10B on tax for employee share vesting. It spent $9.75B on acquisitions C (calculated).⁴

The recent record matters more. From the start of fiscal 2023 to the end of the third quarter of 2026, Adobe generated $32.1B of free cash flow and $25.1B of owner cash. It spent $32.0B on buybacks, $2.0B on tax at share vesting and $1.6B on acquisitions C (calculated).⁴³ So buybacks alone used all of free cash flow, and 128% of owner cash. The gap was met by debt, which rose from $3.6B to $6.4B, and by $1.3B received from employee share purchases C (calculated).⁴³ Net debt is about $0.8B.

What did the buybacks buy? 85.9M shares at an average price of $373: $383 in fiscal 2023, $543 in 2024, $366 in 2025 and $261 in the first nine months of 2026 C (calculated).⁸³ The shares are worth $21.5B at $250.23, $10.5B less than was paid. Put simply, the company paid 49% more per share, on average, than the price today. Against my base value of $372 they were close to fair. Against my bear case of $218 they were expensive. The pattern is the usual one: a company with a rising stock and a growing cash flow buys the most when the price is highest. Adobe's $9.5B in fiscal 2024 bought at the highest average price in the record. I don't blame anyone for not knowing the future. I do note that the cash went out at prices well above today's.

The share count fell from 471M in fiscal 2022 to 395M in the third quarter, 16% in under four years C (calculated).⁴² Shares are the one place where the buybacks did show up in the per-share numbers. They are also the main reason earnings per share on a non-GAAP basis grew 15% in the third quarter when net income grew 8% C (calculated).²¹⁰

The deals are a smaller part of the story. Adobe spent $6.3B on acquisitions in fiscal 2018 and $2.7B in 2021, and its attempt to buy Figma ended in a $1B termination fee in 2024.⁴⁸ In fiscal 2026 it spent $1.87B on Semrush and agreed to pay about $340M for Topaz Labs, an AI image and video enhancement company.³ Semrush added about $480M of ARR, so the price was about 3.9 times ARR C (calculated).³⁵ That looks a modest multiple of ARR to me. Whether it is a good price depends on how much of that ARR keeps growing, and Adobe has not said. It has reported no separate Semrush profit.³

The new debt matters too. At 28 August Adobe had $6.15B of senior notes and $0.25B of commercial paper, against $5.64B of cash and short-term investments.³ Of the notes, $1.35B falls due within a year.³ A company with a 36% GAAP operating margin is in no danger from this. I mention it because the new chief executive and an interim CFO inherit a balance sheet that has changed direction. At the end of fiscal 2024 it held $7.9B of cash and investments against $5.65B of notes. It now holds $5.6B against $6.4B.⁴³

VFinancials

Here is the ten-year record. Reported earnings are a poor guide here for four reasons, so I start with those.

• Tax in fiscal 2020. A net tax benefit of $1.08B, helped by a $1.5B deferred tax gain, turned a pretax profit of $4.18B into net income of $5.26B.⁴
• The Figma fee in fiscal 2024. A $1B termination fee cut operating income and operating cash flow by that amount in fiscal 2024.⁸
• Non-GAAP profit. The company's preferred measure leaves out stock pay, amortization of purchased intangibles, impairment and some legal costs. It was about $11.8B of operating profit in the last twelve months, against $9.3B on a GAAP basis C (calculated).²¹⁶ Stock pay is 82% of the gap.
• A 53-week year. Fiscal 2027 has 53 weeks. The extra week adds about 1.9% to that year's revenue and flatters its growth, and weakens fiscal 2028's.⁸ I leave it out of my model; it is worth less than $1 a share E (estimate).

The table uses numbers as filed. Fiscal years end near 30 November. Diluted shares are weighted averages.

Three things stand out. Revenue multiplied by 4 in nine years and operating income by nearly 6, almost all of it organic until the Semrush deal. Operating leverage is real: sales and marketing fell from 33% of revenue in fiscal 2016 to 27% in 2025, and general and administration from 10% to 7%, while research and development rose from 17% to 18%.⁴ And the best part of the record is the latest part: operating income rose 29% in fiscal 2025 on revenue up 11%, the first year without the Figma fee.⁴ In the last twelve months, though, the leverage stopped. Research and development in the third quarter rose 18% and sales and marketing 11%, against revenue growth of 13%.²

Cash, and what an owner gets

Free cash flow reached $10.6B in the last twelve months, 41% of revenue C (calculated). Owner cash, after stock pay, was $8.5B, or 32.8% of revenue. The company does not define free cash flow; this is operating cash flow less capital spending. Capital spending is tiny, 0.8% of revenue, because Adobe buys its hosting from others, and it fell from 3.3% in fiscal 2020 C (calculated).⁴ Hosting and model costs sit instead inside cost of revenue, which is why gross margin is the line to watch.

Taxes explain why reported cash flow is better than it may look. Cash paid for income taxes was $2.2B in fiscal 2025 against a tax charge of $1.6B, and $1.5B in the first nine months of 2026 against $1.7B a year earlier.⁴³ So cash flow is already after a heavy tax bill.

The deferred revenue line, which was the support for cash flow at many software companies, has weakened. It added $0.77B to operating cash flow in fiscal 2025 and $0.07B in the first nine months of fiscal 2026 against $0.28B a year earlier C (calculated).⁴³ I model a small working-capital gain, 20% of each year's added revenue.

Stock pay

Stock pay was 6.0% of revenue in fiscal 2016 and 8.0% in the last twelve months, after a high of 8.9% in fiscal 2023 C (calculated).⁴ The trend is flat to slightly down, and it is the right direction. Two things keep me from relaxing. First, $4.03B of expense on unvested awards is still to be recognised, over a weighted 2.5 years.³ Second, a lower share price means more shares for the same dollars: employees bought shares in the stock plan at an average of $200.78 this year against $321.93 last year.³ At $250.23 it takes about 29% more shares to deliver the dollar value that $321.93 delivered a year ago C (calculated). The company offsets this with buybacks, and the buybacks are paid out of the same cash.

Margins and costs

The company is guiding to a lower margin. Non-GAAP operating margin was 46.2% in fiscal 2025, is guided to about 45.0% for fiscal 2026 and about 44.0% for the fourth quarter, against 45.6% in the fourth quarter of fiscal 2025 C (calculated).¹⁶² Management said on the second-quarter call that it is spending on models, cloud and marketing for the push into free users, and that it will not be short-term about it.⁹ I take that at face value. I also take it as a sign that the margin I use for the next ten years, 43.5%, is more likely too high than too low.

Gross margin, at 89%, is not the problem. The drop in operating margin comes from spending lines. Spending on research and development rose from 16.7% of revenue in fiscal 2016 to 18.1% in fiscal 2025, and was 19.1% in the third quarter C (calculated).⁴² I expect the number to keep rising.

VIValuation

Before any numbers I start from the far end. Say it is 2031 and this has gone badly. What happened? I can see three stories, and I know of no way to rule any of them out.

The three ways this goes badly wrong

The price of the thing Adobe sells falls. AI tools make good-enough images, video and layouts for free or close to it. Professionals keep their Photoshop seat but stop paying more, buyers of mid-priced plans drop to cheaper tools, and the deferred price rises never return. Growth runs three points below my base every year and margin three points lower, with more of the spend going to compute for free users. My base value falls to $255, about the price today E (estimate). If it goes further, with revenue falling from $27.4B in fiscal 2027 to $24.7B in 2030 and non-GAAP margin falling to 33%, the value is $141, 44% below the price E (estimate).

Free users don't turn into paying ones. The 100 million creative free users and the 900 million Acrobat and Express users are real. They are also an audience that can cost more than it earns. Management chose this path knowing that it lowers ARR in the second half of the year and that half of the shortfall is a price rise it put off.⁹ If conversion disappoints, Adobe has given up price now for users later and received only users. That is one cause of my bear case: growth near 1% for most of a decade and margin falling to 37%.

The capital goes the wrong way. The new chief executive comes from the enterprise side and an interim CFO runs the numbers. A company facing slower growth has two easy answers: buy growth and buy back stock. Adobe did the second at an average of $373 a share since fiscal 2023 and the first with Semrush at about 3.9 times ARR. If it spends 5% of revenue on deals every year and stock pay stays three points above my path, my base falls to $303 E (estimate). I find this the most likely of the three, because the pay plan does not discourage it and the buyback authorisation is $24.55B.³

These can arrive together, since slower growth makes deals more tempting and a lower share price raises the shares needed for stock pay. So I don't add them and I don't put them in my cases. They are why my margin-of-safety price sits well under my base value.

What an owner can really take out

Now the cash. Reported figures give several answers, and they are far apart.

So after my deals allowance an owner's cash today is about $8.0B, 17% below non-GAAP net income of $9.6B and 10% above GAAP net income of $7.3B. The gap to non-GAAP is mostly stock pay, which I count as a real cost: Adobe pays 8 cents of every revenue dollar in shares. On this number the stock trades at 11.6 times owner cash before my allowance, 12.4 times after it. On the company's targets for fiscal 2026 the shares trade at 10.2 times non-GAAP earnings of $24.45 to $24.50 and 13.8 times GAAP earnings of $18.12 to $18.17 C (calculated).²

On what it costs to stand still: capital spending is small. The cost of standing still is stock pay and the compute behind free users. If you take stock pay out of my model, the base value goes from $372 to $448. That $77 a share is what you are asked to ignore when someone quotes free cash flow E (estimate). If you take out my deals allowance, it goes to $397.

How I built it

The model runs on fiscal years, 2027 to 2041, with the rest of the fourth quarter of 2026 as a stub. Each year it takes revenue, applies non-GAAP operating margin, takes off stock pay as if it were paid in cash, and charges cash tax at 21% of what is left, which is what the last twelve months show: $2.04B of cash taxes on $9.7B of profit after stock pay C (calculated).³⁴ It then adds depreciation at 2.1% of revenue, takes off capital spending at 0.8% and a deals allowance, and adds 20% of each year's new revenue for customers who pay in advance. The terminal year takes the working-capital gain down to the rate of long-run growth. The stub is my full-year estimate for fiscal 2026 less the nine months reported, which is $2.8B; my estimate for the full year is $8.7B, against $8.5B in the last twelve months E (estimate).

From the value I take off $6.40B of debt at face value, add $5.64B of cash and short-term investments and take off $0.34B for Topaz, which Adobe expects to pay in cash in the fourth quarter.³ I use 391M shares, the number outstanding on 28 August, the date of the balance sheet. The cover of the 10-Q shows 389.2M on 18 September, and I don't mix the two. Unvested awards are not added, because their cost is already in my stock pay line.³

The discount rate

I use 10.0%, the rate I use on every page of this site. As a check, the capital asset pricing model starts from the 4.94% ten-year Treasury yield and adds beta times Damodaran's implied equity risk premium of 4.20% for 1 October 2026.¹⁵³⁴ Adobe's beta to the S&P 500 is 1.37 over 60 months and 1.20 over 36, which gives 10.7% and 10.0% C (calculated).¹³⁵ So my rate is at the low end of what the formula gives. At 11% the base value is $326. The sensitivity table below shows the effect.

Bear, base and bull

Base, $372. Revenue grows 8.5% in fiscal 2027. That sits on the 10.2% ARR target without Semrush, 8.3%, and on the 9% I measured in the third quarter, and it includes about 0.8 points from a full year of Semrush, which contributed about $280M of revenue in fiscal 2026.²⁹ I then fade growth by half a point to a point a year, to 4% in 2036. Revenue reaches $34.9B in 2030 and $45.6B in 2036, 5.5% a year over ten years. Non-GAAP operating margin steps down from 45% to 44.5% in fiscal 2027 and 43.5% after that, a point and a half below fiscal 2026's target, to pay for AI compute and free users. Stock pay slides from 8% to 7.5% of revenue by 2030. My owner margin of 29% in 2036 compares with 33% today, so the base builds in a drop of four points.

Bull, $521. This is the company's plan working. Revenue grows 10% for two years, close to its ARR target with Semrush, then fades. Operating margin returns to 46.5%, close to fiscal 2024 and 2025, and stock pay falls to 7% of revenue. The price rises Adobe put off come back, free users convert, and AI products add to the base. Revenue reaches $58B in 2036 E (estimate).

Bear, $218. Growth falls to 1% by fiscal 2030 and stays there for most of a decade. Margin falls from 45% to 37% by 2031 and stock pay rises to 9% of revenue. Revenue still reaches $32B in 2036. It is what a good business looks like when the market stops growing under it and the owner has to keep paying to defend its position. It is 13% below the price. On this case the stock is not cheap.

I give the cases weights of 30%, 50% and 20%, which comes to $355 E (estimate). I don't pretend those weights are measured. They are a way to show I think the bear is more likely than the bull. To justify the price with only my bear and base cases you have to believe the bear has about a 79% chance C (calculated).

What I can't judge

The most important input is what AI does to the price and number of creative seats, and I can't know it. Adobe publishes no seat count, no retention rate and no revenue split between new customers and existing ones. It reports ARR in total, and now revalues it each year. The evidence that would settle the question, price paid per user at renewal, is not published. What I have is what management chose to do with the information it has, which is to put off price rises and give away more. One reading is that renewal data did not support a rise. Management says the reason is a larger opportunity. I can't tell which.⁹

So I show the range, not a false point. Each row below moves growth in every year. Each column moves margin in every year.

Values run from $218 to $632. The price sits in the top-left corner: four points less growth a year with my margin, or three points less growth and three points less margin. That width is the honest answer. Growth matters more than margin: four points of growth a year is worth between minus $120 and plus $185 a share, and five points of margin about $50 either way.

Where I differ from the screen

This page started from a screen of stocks that fell 35% or more and still looked cheap. I rebuilt the numbers from the filings and found four differences.²³ The screen had net cash of about $0.6B. The 28 August balance sheet has $6.40B of debt against $5.64B of cash and investments, so net debt is about $0.8B, or $1.1B after Topaz.³ It used 10.2% for ARR growth, which is the year-end target. The third-quarter figure was 11.2%, and about 9% without Semrush.²³ Its bear case was $301, above the price, and its base $434, from one constant growth rate for ten years. Mine are $218 and $372. I fade growth, take margin down to 37% in the bear, charge a deals allowance and start from net debt. It also dated the chief executive change to 2 September, the day the board named the successor. Narayen announced on 12 March that he would step down.²¹⁶

What the price implies

Read it plainly. At $250.23 the market is paying for revenue that grows about 1.4% a year for ten years, with growth fading to zero by 2036, and margins like today's. Or for a margin of 31.5% with my growth. Or for a return of 13.5% a year on my base cash flows, not 10%. Any of those would be a hard decade for a company that grew 17% a year for the last nine and 13% in the latest quarter. They are not impossible. They are what the price has decided.

The discount rate and long-run growth

The margin of safety

I set it at $242, 35% below my base value. That is a larger cushion than the 25% I use on steadier businesses, and 5 points more than the 30% I use for the same AI question on my ServiceNow page. The extra five points are for the two changes at the top: a chief executive who starts on 1 December and a finance chief who is still interim. At $242 my bear case would return 8.7% a year and my base 15% on the cash flows, and the owner cash I count is 8.5% of the market value. The price is 3% above this level, so by my own rule it does not meet the test today. The stock closed at $236.28 on 13 October 2026 and at $193.41 on 25 June, so that price has been available this year.¹

Here is what that means and what it does not. It does not say the stock will fall to $242. It says that at today's price I am not paid enough, by my own standard, for the things I cannot see. The price is far below my base value and slightly above the price at which I would be comfortable with being wrong.

Quality and price are two questions

How good is the business? Very good. Margins, cash conversion and a ten-year record say so. It has the weaknesses I have listed: the price lever is paused, growth is slowing, the top of the company is changing. How good is the price? It is 33% below my base value and 15% above my bear case. The news is bad: organic ARR is slowing and net new ARR is falling. So the price is low against my base and not low against my bear, and which one is right depends on what I can't see. A very good business can still be a poor purchase at the wrong price. Here the business is very good, and the price is low against my base and above my bear, so the answer turns on growth that I cannot see.

The next best use of the dollar

A ten-year Treasury pays 4.94% with no risk to the coupon.¹⁵ Adobe's owner cash yield is 8.7% today (8.6% on diluted shares, as in the table), and on my base case the return is 15.0% a year. In my bear case it is 8.3%, still 3.3 points above the Treasury. In the stress case, where revenue shrinks, it is 3.0%, below the Treasury. Those returns use my 10% terms for the sale in fiscal 2041.

The closest comparison on this site is ServiceNow, which sits 11% below my base value there. It grows 20% against Adobe's 13%, and costs about four times as much per dollar of owner cash: its owner yield is 2.2% against Adobe's 8.6% C (calculated).¹³⁴ Microsoft sits 44% above its base value on its page and yields 1.2% on owner cash with its datacenter spending counted as capital spending. Those two cost more per dollar of cash and show less gap to value. If the AI fear about software is overdone, the stock with the widest gap has the most room. If the fear is right, it has the most to lose, but it starts with the most cash per dollar of price. I do not value Intuit, GoDaddy or Booking on this page, though they share the same idea. At 7.0% owner yield, Intuit is the nearest to Adobe.

Adobe, Intuit, GoDaddy and Booking came from the same screen and rest on one idea: that the market has priced in too much harm from AI. They are one bet, not four.

The price and why the market may be right

The price is below my base value. Here is why the market may be right and I may be wrong.

• The growth I use is the company's, and the company is slowing. Organic ARR growth was about 9% in the third quarter, net new ARR was 39% lower than a year earlier, and to reach its own year-end target Adobe must nearly double the quarter's net new ARR in the fourth.³¹⁷¹⁰
• The price lever is switched off. Management chose to defer Creative Cloud price changes for a time, and half the hit to ARR comes from that.⁹ If AI tools limit what Adobe can charge, the deferral becomes permanent.
• Margin is guided down and my base margin may still be too high. Operating costs grew faster than revenue in the third quarter and management says it will keep spending.²⁹
• Per-share growth leans on buybacks. Non-GAAP earnings per share grew 15% in the third quarter on net income growth of 8%.²¹⁰ The company paid an average of $373 for its shares since fiscal 2023, and cannot do that for ever at a high price.
• The team is unproven in this role. Two top roles are changing at once, one is interim, and the president of the largest product group is leaving.⁶⁷
• I may be too kind on terminal value. 38% of my base value comes after 2041. A buyer today pays for cash a long way off from a product category that may look different by then.

Where this could be in 5 and 10 years

The value at a later date is what the owner cash after that date is worth then, at my 10% rate, per today’s 391M shares and in that year’s dollars. It leaves out the $1.1B of net debt I take off today and the cash paid out before the date. Owner cash is what the model counts: after stock pay and the deals allowance.

In the base, value rises from $372 to $440 in five years and $514 in ten, and the company pays out $281 a share of owner cash on the way. Four numbers carry that: revenue growth fading from 8.5% to 4% by fiscal 2036, a non-GAAP margin of 43.5%, stock pay of 7.5% of revenue, and the 10% rate. At five years it needs organic ARR to hold near 8% through fiscal 2028. At ten it needs Adobe to still sell creative software at that margin in 2036. I can see the next five years better than the five after.

Bear and bull differ most in growth and margin. In the bear, growth is near 1% and margin falls to 37%, so value barely moves ($229, then $260) and nearly all the return is cash. In the bull, growth stays at 9% to 10% for a few years and margin returns to 46.5%.

The most likely way the base is wrong is growth. The tests in Section VII put today’s value nearer $320 to $340, and the year-5 and year-10 values would fall with it E (estimate). These are estimates of value, not forecasts of the share price. The market can sit above or below value for years.

VIIThe case for the price

What most people miss

Part of the slowdown is a choice. Narayen said on the June call that about half the pressure on second-half ARR comes from deferring Creative Cloud price changes and half from the push to free users, and that the price changes are deferred, not dropped.⁹ One half is spending management can slow. The other is a price rise it can still make. The cash is large: owner cash is 33% of revenue, the highest of the seven software companies I compared, and Adobe spent $9.3B on buybacks in the last twelve months, 9.5% of its market value C (calculated).⁴³ The audience is large too. More than a billion people use an Adobe product each month, up more than 20%. The free creative audience is over 100 million, up more than 70%, and ARR for the Firefly app and credit packs grew 40% on the previous quarter.²¹⁰ Those are company figures, unaudited. They show that people use the new tools. They do not show that people pay for them.

The bull case

The case is that AI makes the market for creating things much larger, and the company with the tools, the distribution and the trust of big brands gets a share. Narayen said on 10 September that AI lets Adobe reach everyone from casual creators to professionals, and called the billion monthly users a landmark.¹⁰² Chakravarthy said every software company will move to agentic software, as software once moved from installed to subscription, and that Adobe has done that before.¹⁰ Day said free users are how Adobe widens its base.² The company raised its fiscal 2026 revenue target in June and again in September, to $26.58B to $26.63B from $25.9B to $26.1B in December, partly through Semrush.¹⁶⁵² Outside the company the support is thin. Morgan Stanley analysts, per Reuters on 13 March, called his exit "the loss of an iconic leader at a time of peak uncertainty" for software.³⁶ Argus lowered its rating, citing the leadership transition.³⁷ I found no public statement on Adobe from the investors who write most about software, so I cite none. The one action I can verify is David Ricks's $1.9M purchase at $194.51 in June.³¹

Against reality

My base needs organic ARR growth that settles near 8%, a non-GAAP margin near 43.5% and stock pay easing toward 7.5% of revenue. That is revenue of $34.9B in fiscal 2030, 7.0% a year from the fiscal 2026 target. The price needs only $29.8B C (calculated). The base growth of 8.5% in fiscal 2027 is about 7.7% organic plus 0.8 points from a full year of Semrush, a little under the 8.3% organic exit rate in the company’s year-end target E (estimate).

Today is some way from that. Organic ARR growth is 9.2% and slowing, and the fourth quarter must nearly double the third's net new ARR E (estimate). Revenue growth in the latest quarter was 13%, of which about 2 points are Semrush and 1 point exchange rates, so about 10% underlying E (estimate).²⁵ The margin my base needs, 43.5%, is below the third quarter's 44.0%, so it asks for no recovery.

I tried to break the base. Each test changes one thing in my model and leaves the rest E (estimate).

• Growth fades faster. Organic ARR growth was 11.5% a year ago, is 9.2% now and is 8.3% implied for year end. If that slide continues, revenue grows 8.5% in fiscal 2027, then 6%, 5% and 4.5%, and settles at 3%. Revenue is $33.6B in fiscal 2030 and the value is $342.
• Depreciation meets spending. The model adds back depreciation at 2.1% of revenue and spends 0.8% on capital, which flatters cash by more than a point of revenue for fifteen years. If the two meet by fiscal 2036 the value is $359.
• The new team costs margin. A margin one point lower for good gives $361. Stock pay stuck at 8% gives $367. A stumble, with margin of 43.0% and 42.0% in fiscal 2027 and 2028 and then back to 43.5%, gives $370. A lasting change matters; one bad year does not.
• All of it together. Faster fade, depreciation fix, a margin point lower and stock pay at 8%: $317. At an 11% discount rate instead of 10%: $279.

So the base is more likely too high than too low. A fairer central value is probably $320 to $340, and even the combined case sits 27% above the price. Only my bear case, or growth four points lower every year, gets down to $250. I keep $372 as the base because the changes above are judgements, but I read it as the top of a central range.

Multiply it out. My bull needs four things at once: organic ARR back to 10% or more by fiscal 2028, which I give 30%; the deferred price rises returning without losing customers, 50%; margin back to 46.5% while the company spends on free users, 40%; and stock pay toward 7%, 60%. Multiplied, that is about 4% E (estimate). Because good growth makes the other three easier, I allow 15% to 20%, which is why the bull weight is 20%.

What a 10x would take. Ten times $250.23 is $2,502, about $0.98T of market value. At 25 times owner cash that needs about $39B a year, and at a 33% owner margin about $119B of revenue, 4.5 times fiscal 2026, or growth of 16% a year for ten years C (calculated). My bull case has 8.2%. Adobe's own revenue grew 17% a year over the last nine years, starting from a quarter of today's size. The share price has risen 2.3 times in ten years and 6.4 times at the most, from $107.51 in October 2016 to $688.37 in November 2021.¹⁴ I checked the three best-known cases, which is not a sample. NVIDIA, Broadcom and Tesla each went tenfold from a market value near $100B in under eight years, by my approximate count with today's share counts, and each was growing far faster than Adobe C (calculated).³⁸ For Adobe I think a tenfold rise is not plausible.

Where I come down

The organic ARR arithmetic changed my starting point. The headline 11.2% includes about 2 points that Adobe bought, and the 10.2% year-end target is about 8.3% without Semrush. The deferral points the other way, because management can reverse it. After the tests above I think the value is nearer $320 to $340 than $372, which is still well above the price. The price needs revenue growth near 1.4% a year with margins like today’s. The next report settles part of it: net new ARR near $0.78B in the fourth quarter, a permanent finance chief, and a date for the price changes.

VIIIRisks, and what would change my view

The three ways this goes badly wrong are at the start of Valuation. These are the rest.

• Measures that change. Adobe reported Digital Media ARR until fiscal 2025 and now reports Total Adobe ARR, revalued each year at new exchange rates and now including Semrush.¹⁶³ A company that changes its main measure when growth slows deserves close reading. I have tried to reconcile the new one to the old.
• Leadership. The new chief executive has not run a company this size, the CFO is interim, and no permanent CFO or new pay terms have been announced.⁶⁷ Another top-level departure before the next proxy would worry me more than any one number.
• Legal. The 2024 federal case over cancellation settled in March 2026. A securities class action over the Figma episode was dismissed and is on appeal, and shareholder suits against directors are pending. The company cannot estimate a loss.³
• Exchange rates. Over 40% of revenue is outside the Americas.⁸ ARR is held at fixed rates inside a year, so reported ARR growth can differ from revenue growth without anything changing in the business.
• Debt and buybacks. $1.35B of notes fall due within a year and the company has $24.55B of buyback room.³ A balance sheet that is drawn down by choice can be rebuilt by choice. I count that as a lever, not a risk, but a new management team could use it differently.
• One idea, four stocks. Adobe, Intuit, GoDaddy and Booking came from the same screen because the market fears AI will take over the customer relationship in software, travel booking and web services. If that fear is right, they fall together. If it is wrong, they rise together. They are not independent bets.
• Key people below the top. Of the five named executive officers in the 2026 proxy, four have left, are leaving or are changing role this year.²⁸⁶⁷