What we said in September — and what happened
The price fell with software, the big tests wait for 9 December, and six details were wrong
On 16 September we published our second X-ray of Adobe at $250.50, nine cents above the price of our first in July. Our view was that Adobe's numbers had moved in cost rather than in demand: revenue was growing faster, but the gross margin had slipped and the contracted order book had grown only 8%. We said to keep accumulating, to add toward $200, and to hold the position at half the size we would have chosen in July, because the payoff depended on 2027 and the chief executive who chose the strategy was leaving. We said the number to watch was the order book in December.
That December report has not yet come. Adobe's fiscal year ends on 27 November and it reports on Wednesday 9 December, eight days into its new chief executive's tenure. So this is an unusual follow-up: no new quarter, a 13.1% fall, and a filing, the third-quarter 10-Q published six days after our report, against which we can check our own numbers. Here is the account.
| What we said on 16 Sep | What happened by 20 Nov | Grade |
|---|---|---|
| 'Accumulate; add toward ~$200; half size' | Down 13.1% to $217.71, against 12.7% for the software basket (IGV). The low was $216.95; $200 was never reached. | ◆ Open, unrewarded |
| 'The order book failed' (RPO +8%) | Confirmed in the 10-Q: $22.16bn, up 8%, and lower than in May. | ✓ Right |
| Net new ARR ~$0.40bn, down ~36% (an analyst's figure, unchallenged) | Our own constant-currency arithmetic from the 10-Q gives −39%. | ✓ Right |
| 'Watch RPO in December… management promised it steps up' | Not yet reported. The finance chief described a seasonal pattern ('this year is no different to prior years'), not a promise. | ◆ Open; we overstated it |
| Gross margin 88.1%, out of an 88.9–89.3% band | 88.7%, down 0.6 points, out of a band of 89.0–89.8%. Still the lowest since early 2024. | ✗ Wrong number, right direction |
| 'An 11% cash yield; 9.1× free cash flow' | Trailing free cash flow was $10.59bn, not $10.94bn: 10.8% and 9.25× at the time. | ✗ Overstated |
| 'Figma pays the same bill with far less cash' | Figma's third quarter: revenue +43%, far above its own forecast, and an adjusted operating margin of +12%. | ◆ A stronger rival than we said |
| A new chief executive on 1 Dec; the creative head leaving; an interim finance chief | All true, and none of the gaps filled: no permanent finance chief and no head of the creative business named. | ✓ Right, unresolved |
| Topaz Labs: '$700m–$1bn is speculation' | The 10-Q gives the price: about $340m, mostly cash. | ✓ Right to doubt |
Our errors. Re-checking the September report against the 10-Q and Adobe's releases, we found six details that were wrong. First, the gross margin. We wrote that it fell to 88.1% from a band of 88.9–89.3%. The correct figure is 88.7% ($5,997m of gross profit on $6,760m of revenue), down 0.6 points from a year earlier, and the eight quarters before it ran from 89.0% to 89.8%. Our number came from our own data feed, which moved $40m of amortisation into the cost of sales for that one quarter (fault one in Part IX). The direction was right; the size was double. Second, free cash flow over twelve months was $10.59bn, not $10.94bn: we added the wrong quarter. The data services we dismissed in a footnote had it right. At $250.50 that was 9.25 times free cash flow and a 10.8% yield, not 9.1 times and 11%. Third, the market value was $97.9bn, not $99.6bn, and it was not the first time below $100bn since 2018: it was about $77bn at the June low. Fourth, we said total ARR growth for the first quarter was not disclosed. The first-quarter 10-Q discloses it: 10.9%. Fifth, the trailing price-to-earnings ratio was 14.0, not 13.6, as another part of the same report said. Sixth, we presented Shantanu Narayen's remark that the payoff "will play out, I think, over 2027" as part of the September story. It was said on the June call; the September call does not mention 2027. We also quoted the phrase "deferred, but not closed" about price rises. It is not in Adobe's corrected transcript, and we withdraw it.
One thing the corrections do to the argument deserves saying plainly. The first and second both made Adobe look slightly worse than it was, the first in the business and the second in the price. The margin dent is real but half the size, and the cash yield was a little lower than we said. Everything else we re-checked was right: revenue, subscription revenue, operating income and margins, earnings per share, the share count, the order book, the recurring-revenue series, AI-first ARR, the guidance, net debt, the buyback authorisation and the facts of the succession.
★ Down 13%, in step with software
Ahead of the basket for two months, then four bad days
| Date | What happened | ADBE that day |
|---|---|---|
| 22 Sep | No company news found; the 10-Q was filed after the close. The software basket fell 0.4%. | −4.5% |
| 12 Oct | Wells Fargo cut Adobe to Underweight with a $225 target (as reported). | −1.7% |
| 15 Oct | Adobe and NVIDIA announced a wider partnership (as reported). | +4.6% |
| 22 Oct | No cause found. The high since our X-Ray, 8.0% above our price. | +6.1%, $270.55 |
| 4 Nov | Citi cut Adobe to Sell with a $207 target, warning of 'a hard reset' (as reported). The shares rose. | +1.0% |
| 10–12 Nov | Adobe MAX, the annual product conference: many AI launches, no financial targets. | −0.3%, +0.4%, +0.4% |
| 17 Nov | Software sold as a group ahead of Anthropic's stock-market listing (MarketWatch, as reported). | −4.3% |
| 18 Nov | Adobe fell while the basket rose 1.3%. No company news found. | −4.0% |
| 20 Nov | The basket fell 4.6%; Adobe 1.8%. Lowest close since July. | −1.8%, $217.71 |
No news, one direction. Since 16 September Adobe has published no results, no forecast and no 8-K. Its most important announcements were products at its MAX conference in Miami Beach, and the shares barely moved on them. The fall came in two stages. Through 16 November Adobe was down 1.8% while the software basket was down 5.7%. In the four trading days from 17 to 20 November it fell 11.5%, against 7.4% for the basket, and the two ended within half a point of each other. The market did not learn anything new about Adobe in those four days. It sold the group, and on 18 November it sold Adobe on its own.
The analysts moved first. Three firms downgraded the shares after our report: Wells Fargo to Underweight on 12 October, HSBC to Hold on 13 October, and Citi to Sell on 4 November. Citi's argument, as reported, is the clearest statement of the bear case now. Adobe chose to put its AI product, Firefly, ahead of price rises; Citi estimates that this costs about 2.5 points of recurring-revenue growth in fiscal 2027, cut its forecast for the gross margin for "unmonetized AI inference costs", and expects recurring revenue to grow only about 5% next year. The average analyst target is about $267 and the consensus rating is Hold (stockanalysis.com).
Where the price sits. At $217.71 the shares are 13% below our price, 13% below the $250.76 Adobe's outgoing chief executive received for $31m of shares in September, 7% below the average price Adobe paid for its own shares in the third quarter, and 12% above the $194.51 a director, David Ricks, paid for 10,000 shares in June. The market value is about $84.7bn, 68% below the peak of November 2021.
★★ The test on 9 December
What the order book and new recurring revenue must show, and what Adobe's own targets imply
In September we said that revenue had been the wrong number to watch, because at a subscription company it describes contracts signed long ago. The two numbers that look forward are the order book, which Adobe calls remaining performance obligations (RPO), and net new annual recurring revenue, the recurring revenue added in a quarter. We said Adobe would pass if the order book returned to double-digit growth and net new ARR stopped shrinking compared with a year earlier. Both answers come on 9 December. We can already say a good deal about them.
The order book. At the end of August it stood at $22.16bn, 8% higher than a year earlier and slightly lower than in May. The order book always jumps in Adobe's fourth quarter, because that is when large customers renew. Last year it rose by $2.08bn, to $22.52bn. To show growth of 10% this time it must reach $24.77bn, a jump of $2.61bn, a quarter more than last year's. If Adobe repeats last year's jump exactly, growth will be 7.6%. When we wrote that management had promised a step-up, we overstated what was said. Steve Day, the interim finance chief, said the book would "step up in our fourth quarter and then remain fairly flat over the next three quarters, and this year is no different to prior years in that shape." That is a description of the calendar, not a forecast of growth.
New recurring revenue. Adobe still aims for total ARR growth of 10.2% this year, the target it set at the start of the year. The year began with $25.66bn of recurring revenue at this year's exchange rates. Reaching 10.2% means ending at about $28.28bn, and the company stood at $27.50bn in August. So the target implies about $0.78bn of net new ARR in the fourth quarter, against about $0.92bn a year earlier: a fall of about 15%. Adobe's own target therefore implies that net new ARR shrinks for another quarter. If Adobe hits its targets exactly, both of our tests fail.
| The test | Last year's fourth quarter | What Adobe's targets imply | To pass |
|---|---|---|---|
| Order book (RPO), growth | +$2.08bn in the quarter | ~+7.6% if repeated | ≥ $24.77bn (+10%) |
| Net new ARR, constant currency | ~$0.92bn | ~$0.78bn (−15%) | > $0.92bn |
| Total ARR growth, fiscal 2026 | — | 10.2% (target kept since December 2025) | Above 10.2% |
| Fourth-quarter revenue / adjusted EPS | — | $6.80–6.85bn / $6.30–6.35 | Usually beaten |
| Fiscal-2027 targets | First given in December | Consensus: revenue +9.1%, EPS +12% (stockanalysis) | The real test |
Why we still think the test is right. A buyer of Adobe at nine times earnings is not paying for the last quarter's revenue. The buyer is paying for whether people still sign multi-year contracts for Creative Cloud and Acrobat in a world of free AI tools. The order book and new ARR are the closest public measures of that. If they fail again, the bear case gains a second data point. That would not prove the moat is gone. In September revenue still grew 12.9% and subscription revenue 13.7%. It would show that the company is renewing its customers on less generous terms than before.
Why December may say less than it seems. Three things will blur the reading. First, the new chief executive, Anil Chakravarthy, will set the fiscal-2027 targets eight days after taking office, and new chief executives usually set targets they can beat. Second, Adobe chose in June to put off price changes in Creative Cloud; Narayen said that about half of the resulting effect on this year's ARR came from that decision and half from the free offer. A price rise in 2027 would flatter the numbers; its absence would depress them. Third, the order book now includes Semrush, which Adobe bought in April. We will judge December on two questions: how much the order book and new ARR fell short, and whether the fiscal-2027 targets show the gap closing or widening. A miss is not by itself a reason to sell. A cut to the long-run outlook would be.
The margin dent, measured again
Half the size we printed, and still real
Our September letter began with a confession that "something in the accounts moved", and the number we led with was a gross margin of 88.1%. The right number is 88.7%. We owe readers a second look at what the correct figure means.
The quarterly gross margin was 88.7%, against 89.3% a year earlier and 89.2% in the second quarter. It is the lowest since the first quarter of fiscal 2024. Over twelve months the margin is still 89.3%. That is a dent, not a break. But the reason for it is the one we described, and the detail is in the 10-Q. In the third quarter the cost of delivering subscriptions rose 24%, while subscription revenue rose 13.7%. The gross margin on subscriptions alone fell from about 91.2% to 90.4% (our arithmetic). Adobe now runs and pays for AI models that generate images, video and sound for its users, including models from Google, OpenAI, ByteDance, Kling, ElevenLabs and Topaz inside Firefly, according to its MAX announcements. Each generation costs money. Many of them are given away to bring in users.
| Quarter | Gross margin | Note |
|---|---|---|
| Q3 fiscal 2025 | 89.3% | A year earlier |
| Q4 fiscal 2025 | 89.5% | |
| Q1 fiscal 2026 | 89.6% | |
| Q2 fiscal 2026 | 89.2% | Semrush joins in April |
| Q3 fiscal 2026 | 88.7% | Lowest since early 2024 (we printed 88.1%) |
| Subscription cost vs subscription revenue, Q3 | +24% vs +13.7% | The AI bill |
How big can the bill get? In September we wrote that if the gross margin settled at 85% instead of 89%, Adobe would earn about a billion dollars a year less. That arithmetic still holds and it is the right scale for the risk: about 4% of revenue, or roughly a tenth of free cash flow. Citi's forecast, as reported, trims next year's margin by 0.7 points, which is about $200m. The rivals show the same pattern. Figma's gross margin fell from 84% to 80% in a single quarter as its customers used more AI, according to its own release.
Why it matters less than the order book. A margin that falls from 89% to 87% still leaves Adobe among the most profitable software companies in the world, and its operating margin before stock pay is about 44%. What would matter is a margin that keeps falling while growth does not rise, because that would mean Adobe was paying for AI and getting nothing for it. So we will read the gross margin and the order book together in December. A lower margin with a stronger order book means the spending is working. A lower margin with a weaker order book means it is not.
★ Three empty chairs and a $31m sale
A new chief executive, an interim finance chief, no head of the creative business
On 1 December Anil Chakravarthy becomes Adobe's chief executive and joins its board, and Shantanu Narayen, chief executive since 2007, becomes executive chair. We described the succession in September. Since then very little has been settled. David Wadhwani, who ran the creative and productivity business, left that role on 27 September and stays on as a senior adviser. No successor has been named. Steve Day has been interim finance chief since Dan Durn left for Marvell in June, and no permanent appointment has been announced. Adobe said in September that it would file the new chief executive's pay terms in an amendment; it had not done so by 22 November. On 9 December Adobe will report with a chief executive eight days in office, a finance chief without the permanent title, and nobody named to run the business that earns most of its money.
Retention. The 10-Q filed on 22 September includes a "Form of Retention Agreement" for senior executives. It promises a year's salary and a full target bonus, a year of health cover and accelerated vesting of shares to anyone dismissed without cause, and half of a separate grant of "Retention RSUs" to those let go before 15 July 2027. It lapses a year after the permanent chief executive starts. Forms 4 show large share grants on 15 July to the chief legal officer, the chief people officer and the chief marketing officer, 31,662 shares each, worth about $7m at the time, and 22,616 shares to Day. Retention agreements are common in a succession. This one shows that the board was worried enough about departures to pay to prevent them.
The sale. On 16 and 17 September, the day we published and the day after, the Narayen family trust sold 125,000 shares at an average of about $250.76, about $31.3m. The Form 4 says the sale was "primarily made for the reporting person's tax and estate planning purposes", and it was not made under a pre-arranged trading plan. The trust still holds about 243,000 shares. A departing chief executive selling for estate planning is not evidence of bad news, and we do not read it as such. But Adobe was buying its own shares in the same weeks, at about the same price, with shareholders' money, and the person with the best view of the business was selling. Owners should notice that.
| Role | Who, at 22 November | Since |
|---|---|---|
| Chief executive | Shantanu Narayen until 30 Nov; Anil Chakravarthy from 1 Dec | Board decision 2 Sep |
| Executive chair | Shantanu Narayen from 1 Dec | |
| Chief financial officer | Steve Day, interim | 15 Jun 2026 |
| Head of the creative and productivity business | Vacant (David Wadhwani left the role) | 27 Sep 2026 |
| Lead independent director | Frank Calderoni (led the search) |
What it means for the verdict. We lower our score for management and capital to 4 from 5. Not because anyone has done anything improper; the governance is ordinary. But in September we said the number of things that had to go right before the end of 2027 had gone up, and the people responsible for them had changed. Ten weeks later, two of those people have still not been named. Chakravarthy built Adobe's business with large companies and knows the enterprise customers well. The creative professionals, whose subscriptions are the heart of the company, will be led by someone not yet chosen.
The moat, rechecked
The file formats still hold the professionals; the entry point is still being given away
| The claim (September) | Evidence now | Verdict |
|---|---|---|
| The professional standard: Photoshop, Illustrator, Premiere | At MAX Adobe let outside AI agents open and edit native Photoshop, Illustrator and Premiere files through a standard connection, and put Google's and OpenAI's models inside its own apps. The files stay Adobe's. | ✓ Intact |
| Acrobat and the PDF | A Google partnership puts Acrobat inside Google Workspace (MAX release). | ✓ Intact |
| The free entry point: 100m free creative users | No new user numbers. Firefly is now an 'AI creative studio' with dozens of outside models, most of it free to start. | ◆ Unproven |
| Pricing power | Price rises put off since June; none announced at MAX. Citi puts the cost at ~2.5 points of next year's ARR growth (as reported). | ✗ Unused |
We keep the moat at 6. Its centre is the same: professionals who have spent years learning Adobe's tools and whose clients, studios and printers expect Adobe's files. The most interesting decision at MAX was to open that centre to the AI laboratories rather than compete with them. Adobe now lets an outside agent, built by Anthropic, OpenAI or Google, work directly in a Photoshop file. That is a bet that the file format, not the menu, is the moat, and it is a sensible one. A business that controls the standard can let anyone build on it, as long as the work still ends in the standard.
The weak point is the one we named in September: the entry point. Adobe is giving more away to win new users, and it has chosen not to raise prices for the users it already has. Both decisions make sense for a company defending its position. But they are also what a company does when it is no longer sure of its pricing power. We will know which it is when Adobe raises prices again and we see whether customers stay.
- That owner earnings, about $21.89 a share today, do not shrink over the next decade. The price is below the value of ten years with no growth at all.
- That AI costs leave the gross margin above about 85%, and that the free users Adobe is buying turn into paying ones over time.
- That the buyback keeps retiring 5–7% of the shares a year, so that flat profits still mean rising profits per share.
The competition
Figma beat its forecast; the AI laboratories became partners inside Adobe's apps
| Rival | Where it meets Adobe | The latest | Read |
|---|---|---|---|
| Figma | Interface and product design; now a broader creative suite. | Q3 revenue $391m, +43%, $16m above the top of its forecast; adjusted operating margin +12%; full-year forecast raised again (Figma 8-K, 4 Nov). | ✗ Stronger |
| Canva | Marketing and everyday design; Affinity for professionals. | No new figures since September. Made its Affinity professional design suite free (as reported). | ◆ Steady pressure |
| AI laboratories (Google, OpenAI and others) | Images and video from a sentence; agents that do creative work. | Now partners: their models run inside Firefly and Photoshop, and their agents can edit Adobe files (MAX release). | ◆ Partner and rival |
| Free and cheap tools | Simple edits, social posts, small businesses. | The reason Adobe is giving away Firefly and Express. | ◆ Taking the bottom |
Figma deserves a correction of tone. In September we used Figma as evidence that Adobe was the only company in its field able to afford the AI transition. Figma had grown 48% with a large operating loss. Its third quarter weakens that argument. Revenue grew 43%, its customers' spending grew 31% on the same accounts, and before stock pay it earned an operating margin of 12%. Its gross margin fell, as Adobe's did, from 84% to 80%. Figma is paying the same bill and growing faster, from a revenue base about a seventeenth of Adobe's. That does not make it a threat to Adobe's profits next year. It does mean that the best young designers, who once grew up on Photoshop, increasingly grow up on Figma.
The laboratories chose to be partners. The other change is that the AI companies Adobe feared are now inside its products. At MAX, Adobe announced Google's models in Firefly, Photoshop, Express and Premiere, Adobe tools inside Google's search, OpenAI's and others' models in Firefly, and connections that let outside agents work in Adobe files. This turns Adobe from a rival of the laboratories into a place where their models are used, and Adobe pays them for it. It is a good defence of the professional business. It also means that more of the value of each image goes to whoever supplies the model.
Management, insiders & capital allocation
A handover in December, one large sale, and buybacks at $235
Capital allocation. In the third quarter Adobe bought back 9.5m shares for $2.23bn, at an average of about $235, and over the first nine months of the year 26.1m shares for $6.82bn, at an average of about $261. That is 91% of the free cash flow of those nine months. The shares outstanding fell from 418.6m in September 2025 to 389.2m in September 2026, a reduction of 7.0% in a year. The first-quarter purchases, at about $306, are now 29% under water; the third quarter's are 7% under water. Adobe has $24.55bn of buyback authority left until 2030, equal to 29% of its market value at today's price.
A buyback is only as good as the price paid. At $217.71 Adobe buys about $12.50 of free cash flow for every $100 it spends. Few uses of its cash would earn more, and we think the buyback is the right use of it. The company has started to fund part of it with debt. Net cash of $385m at the end of fiscal 2025 became net debt of $724m in August, after the $1.9bn Semrush purchase, and $1.35bn of bonds fall due in February and April 2027, which Adobe says it will refinance. It is a small amount of debt for a company with $10.6bn of free cash flow, but it shows the buyback is now running at the limit of the cash.
Insiders. Since 16 September the only purchase or sale of shares on the market by an insider was the Narayen trust's sale. The other Forms 4 show only shares vesting and shares withheld for tax. The last purchase was the director David Ricks's 10,000 shares at $194.51 on 25 June, which our September report did not mention.
The numbers
Twelve months to August, a cash yield of 12.5%, and six faults in our feed
Twelve months to August. Revenue was $25.97bn, up 12%. Gross profit was $23.18bn, a margin of 89.3%, and GAAP operating income $9.27bn, a margin of 35.7%. GAAP net income was $7.28bn, or $17.92 a share; adjusted earnings per share, which exclude stock pay and the amortisation of acquisitions, were $23.65. Operating cash flow was $10.81bn and capital spending only $0.21bn, so free cash flow was $10.59bn, 41% of revenue. Stock-based pay was $2.07bn, 8% of revenue, which is low for a software company. Free cash flow less stock pay, our measure of owner earnings, was $8.52bn, or $21.89 a share on the 389.2m shares outstanding.
The quarter, once more. The third quarter, to 28 August, had revenue of $6.76bn (+12.9%), subscription revenue of $6.58bn (+13.7%), a GAAP operating margin of 34.8% and an adjusted one of 44.0% (46.3% a year earlier). GAAP net income rose 3.1% to $1.83bn and GAAP earnings per share about 10% to $4.62, the difference being the buyback. Research and development rose 18.4%. These figures were right in September; we repeat them so that this report can be read on its own.
The balance sheet. Cash and short-term investments were $5.64bn against borrowings of $6.36bn: net debt of $0.72bn, less than a month of free cash flow. A $1.5bn credit line is undrawn. The 10-Q lists no new legal matters. The shareholder suit over the abandoned Figma deal was dismissed and the plaintiff is appealing, and five derivative suits remain stayed.
| What the screen shows | What is wrong, and the number to use |
|---|---|
| Q3 cost of revenue $803m; gross margin 88.1% | $40m of amortisation, an operating expense, was moved into the cost of revenue for this quarter only. The filed figures are $763m and <b>88.7%</b>. This fault produced our own first error. |
| Debt $6.76bn; net debt $2.40bn | Counts $0.40bn of leases as debt and ignores short-term investments. Net debt is <b>$0.72bn</b>. Flagged in September; still unfixed. |
| Business description: three segments | Adobe has reported a single segment since the first quarter of fiscal 2026, with two customer groups. The description is out of date. |
| Revenue by segment: fiscal 2024 total $36.7bn | Rows are duplicated and the series is out of order, so the 'latest' year shown is 2021. Revenue was $21.5bn in fiscal 2024 and $23.8bn in fiscal 2025. Do not use this series. |
| Analysts: 'Buy', 64 analysts; lowest target $190 | Inconsistent with the feed's own monthly history (40 analysts). Current sources give <b>Hold</b>, about 39 analysts, a lowest target of $140, and three downgrades since October. |
| Estimates: fiscal-2029 EPS below fiscal 2028 | An artefact of different analysts in each year ($29.45 from 11 analysts, $30.55 from 15). Treat estimates beyond fiscal 2027 as thin. |
No dividend — where the cash went
Nine-tenths of the free cash flow into buybacks, and 7% of the shares retired in a year
Adobe has never paid a dividend and has given no sign that it will. For a company growing at 12% and earning these margins, that can be the right choice, provided the retained cash earns more inside the company than owners could earn with it outside. Adobe's answer has been to hand nearly all of it back through buybacks: $52.5bn since fiscal 2016, including $6.8bn so far this year. The test of that policy is the price. From 2021 to 2024 Adobe bought at 30 to 60 times earnings, and much of that money has been lost on paper. Today it buys at about 12 times GAAP earnings and 8 times free cash flow, the cheapest in at least five years.
The result for an owner is easy to see. In twelve months the number of shares fell by 7%, so each remaining share owns 7% more of the business. Stock pay of $2.07bn was offset four and a half times over. If profits stayed flat for a decade and the buyback continued at this price, earnings per share would still rise by something like 7% a year, the pace of the last twelve months. That is the strongest argument for owning Adobe now, and the reason a flat business can still be a good investment at this price.
| Use of cash | Amount | Read |
|---|---|---|
| Free cash flow, 12 months | $10.59bn (41% of revenue) | ✓ Strong |
| Buybacks, 12 months | ~$9.3bn, ~33m shares at ~$279 | ◆ 88% of free cash flow |
| Buybacks by quarter, fiscal 2026 | Q1 ~$306 · Q2 ~$249 · Q3 ~$235 | Cheaper each quarter |
| Stock-based pay, 12 months | $2.07bn (8% of revenue) | ✓ Offset 4.5× |
| Shares outstanding | 418.6m → 389.2m (Sep 2025 → Sep 2026) | ✓ −7.0% |
| Net cash → net debt | +$0.39bn (Nov 2025) → −$0.72bn (Aug 2026) | ◆ Semrush and buybacks |
| Dividends | None, ever | — |
★ Valuation
Priced below the value of a decade without growth
Our method. As before, we value Adobe on owner earnings, free cash flow less stock-based pay, over ten years at 10% a year, with a multiple on the tenth year. We start from $21.89 a share. We have moved our cases since September for three reasons: the corrected free cash flow, which is slightly lower; the evidence that the AI bill is real but smaller than we printed; and Figma's strength, which makes the most hopeful case a little less likely. We ignore the small net debt.
| Path | Owner earnings start | Growth, years 1–5 / 6–10 | Exit multiple | Value today |
|---|---|---|---|---|
| AI takes the entry point | $19.70 (margin to ~85%) | −2% / 0% | 9× | ~$174 |
| No growth for a decade | $21.89 | 0% / 2% | 10× | ~$231 |
| Central case ◆ | $21.89 | 6% / 4% | 13× | ~$354 |
| Freemium converts | $21.89 | 10% / 6% | 16× | ~$499 |
| What $217.71 implies | $21.89 | ~0% a year, constant | 10× | $217.71 |
What is priced in? At $217.71, and with a multiple of 10 on the tenth year, the price assumes that owner earnings never grow again. With a more normal multiple of 13 it assumes they shrink by about 1.6% a year. Adobe's own guidance is for adjusted earnings per share to rise about 17% this year, and analysts expect about 12% next year. Much of that growth will come from the buyback rather than the business. Even so, the gap between a company growing earnings per share at double digits and a price that assumes decline is very large.
Why not buy it all now? Because the next seventeen days may tell us something important, and because the central case depends on growth we cannot yet measure. If growth in the first five years is 3% rather than 6%, the central value falls to about $312; if the exit multiple is 10 rather than 13, to about $313; if owner earnings fall 10% at the start because of AI costs, to about $318. All of those are still well above today's price. But December's targets will be set by a new chief executive, and a cautious first year is likely. A buyer who wants to own Adobe can buy half now at a price below the no-growth value. The other half is worth waiting for, either at a lower price or until there is better evidence.
Our level. For a business with a real moat and an uncertain growth rate we want about 40–45% off our central value of $354. That gives about $200, the level we set in July and kept in September: 8.2 times this year's adjusted earnings guidance, 9.1 times owner earnings, and close to the price at which even the case where AI takes the entry point earns about 8% a year. Buy the half now; keep the rest for below $200.
Risks, lawsuits & controversies
Verified 22 November 2026 against the Q3 fiscal 2026 10-Q, 8-Ks, Forms 4 and attributed reports
Legal. The 10-Q lists no new matters. The securities suit over the abandoned Figma acquisition was dismissed in March 2025, permission to amend was refused in November 2025, and the plaintiff has appealed. Five shareholder derivative suits are stayed. The FTC and Justice Department case over Adobe's cancellation terms was settled in March 2026.
Governance. The board adopted retention agreements for senior executives during the succession, the outgoing chief executive's trust sold $31.3m of shares outside a trading plan, and the new chief executive's pay terms had not been filed by 22 November.
| Test | Today | What would change our mind |
|---|---|---|
| Order book (RPO), 9 December | $22.16bn, +8% | Above $24.77bn (+10%) passes; below $24.24bn (last year's jump repeated) would be a clear second miss. |
| Net new ARR in the fourth quarter | Target implies ~$0.78bn | Above last year's ~$0.92bn passes; well below $0.78bn would mean Adobe missed its own target. |
| Fiscal-2027 targets | Consensus revenue +9%, EPS +12% | ARR growth targeted below ~7% would confirm a slower business; we would then buy only at the level. |
| Gross margin | 88.7% (Q3) | Below 87% without a stronger order book would mean AI is costing more than it earns. |
| Leadership | Interim CFO; creative head vacant | Strong permanent appointments by the spring would lift our management score. |
| The price | $217.71 | Below $200 with the tests no worse than Adobe's own targets: complete the position. |
In September I told you to keep accumulating Adobe at two hundred and fifty dollars, at half the size I would have chosen in July, and to add toward two hundred. Since then the company has said almost nothing, and the shares have fallen thirteen per cent, almost exactly as much as every other software company. A reader who followed me is down thirteen per cent on that half and has not yet bought the second.
Let me start with what I got wrong, because I got more wrong than I would like. My letter opened with a gross margin of eighty-eight point one per cent. The real figure was eighty-eight point seven. The direction was right and the size was double, and I built a paragraph on it. I also overstated the free cash flow, misdated a remark by the chief executive, and quoted a phrase I cannot now find in the company's transcript. Six details in all, corrected in this report. None of them changes the verdict, but you are entitled to numbers you can trust, and this time I gave you some that were too dramatic.
Now what has not changed. On the ninth of December Adobe will show its order book and the recurring revenue it added in the autumn. I said those were the numbers to watch, and I still think so. I have done the arithmetic on Adobe's own targets, and if the company hits them exactly, both numbers will be weaker than a year ago. So I expect the test to fail a second time. A new chief executive, eight days in the job, will then set targets for next year, and new chief executives seldom promise much in their first week.
So why buy at all? Because of the price. At two hundred and eighteen dollars you pay about eight times the free cash flow, and the company spends almost all of that cash retiring its own shares at the same low price. Seven per cent of the company disappeared in a year. If Adobe's profits never grow again, the shares are still worth more than you are paying. For the price to make sense as a poor investment, the business has to shrink. The professionals are not leaving. The order book is growing more slowly, not falling. And the company has just opened its file formats to the AI companies that were supposed to replace it.
There are things that make me uneasy. The chief executive who chose the strategy sold thirty-one million dollars of shares the day I published, while the company bought at the same price. Two of the people who will run the business for the next three years have not been named. And a rival I described as struggling grew forty-three per cent last quarter. None of these alone is a reason to sell. Together they are a reason not to be in a hurry.
My verdict is the September verdict, at a lower price and with fewer illusions. Buy the half now. Keep the other half for below two hundred dollars, or for a December report that is better than I expect. If the fiscal-2027 targets show recurring revenue growing more slowly than about seven per cent, I will still buy at two hundred, because the price will pay for the risk, but I will not pay more.
— The Buffett Lens. When a company's tests are about to fail and its price already assumes it will never grow, the question is not whether the news will be bad but whether the price already allows for it.
the Rest Below $200A cash machine whose forward tests will probably fail again on 9 December, priced as if it will never grow. ★ At $217.71 — 8.0× free cash flow and below the ~$231 value of a decade without growth — the price sits well under our central value of ~$354, within a range of ~$174 (AI takes the entry point) to ~$499 (freemium converts). Buy half a position now; keep the rest for below $200, about 44% below our central value. Watch the order book, net new ARR and the fiscal-2027 targets on 9 December.
Articles that take this apart
MarketDoes AI kill per-seat software?If software is priced per person and the work stops needing people, the arithmetic looks grim. The place to check is not the theory — it is the renewal rate.
MarketWhat a 5% risk-free rate does to everything elseA hurdle rate is not an opinion. When the safest asset available pays 5.12%, every other asset has to explain itself — and most of ours currently cannot.
ValuationWhen buybacks destroy valueA buyback is not a reward to shareholders. It is the company buying its own stock — and like any purchase, everything depends on the price paid.