ADBE: One Billion Users, but What Reaches Owners?

The 60-second read

Five things to know

1. Adoption is expanding. AI-first ARR grows more than 150%, but the current dollar base and paid-user conversion are undisclosed.
2. Costs matter. Revenue grows faster than both GAAP and adjusted operating income.
3. Compensation is recurring. $544M of the $620M operating adjustment is stock-based and deferred compensation.
4. Cash remains strong. Trailing FCF less SBC reaches $8.521B, with repurchase discipline still important.
5. Value has a range. Three routes yield $262.68–$314.94; the $300 anchor requires a $225 entry at the stated uncertainty.

Three scenarios, one probability-weighted number

Analyst twelve-month scenarios; probabilities are judgments.

Signposts: what would change our mind

Expected earnings-review months are not confirmed release dates. Preserve currency, acquisition and compensation definitions when grading these analyst decision rules.

The tape

An early quote is available; a complete post-earnings session is not.

Adobe is a HOLD with conviction 3/5 and High uncertainty. Our central fair value is $300 per share, against a saved after-market trade of $244.20 on September 10. That appears attractive on a headline earnings multiple, but the discount to value is 18.6%, below the 25% cushion we require for this high-uncertainty case. A $225 entry would meet that price test only if the underlying earnings and cash assumptions remain intact. The valuation is an appraisal of a business in transition, not a prediction that the stock must reach a target within a specified trading window.

The regular session closed at $248.83. The saved after-market trade at 20:33:38 UTC is 1.86% lower, while the nearby bid/ask observation is $243.40/$244.42. A small trade and a spread are less definitive than a completed session, so this packet fixes the timestamp rather than pretending it captures the whole market response. The ranking feed’s 2.36% magnitude describes the regular-session decline and is unsigned in the ranking object. It cannot be substituted for an earnings reaction. No full September 11 closing reaction exists at the research cutoff.

Historical reactions show how demanding the market has been, but they do not prove that every decline was caused by one earnings metric. We calculate the first regular close after an after-market release against the release-day close, using the same convention across eight completed observations. The current quarter is deliberately left pending in that comparison. The prior channel episode framed Q2 as a record quarter with a raised guide and a falling share price. This packet takes a different question: how much value do AI reach, recurring contracts and repurchases create after paying the costs that remain with shareholders?

The print

Revenue and EPS beat; profit growth is less uniform.

The primary Exhibit 99.1 covers the three months ended August 28, 2026. Revenue is $6,760M, up 12.9% from $5,988M. Subscription revenue of $6,582M accounts for almost all the total; product revenue is $67M and services and other revenue is $111M. The release reports 13% growth, or 12% in constant currency, using rounded company percentages. The machine dataset preserves the underlying dollars and computes its own growth consistently. Its prior-period columns are August 29, 2025, and the nine-month columns are separate from the quarterly columns.

Non-GAAP diluted EPS is $6.13, above the saved vendor estimate of $6.08 and above management’s prior $6.05–$6.10 target. GAAP diluted EPS is $4.62. Those are distinct definitions, and the comparison to consensus uses the adjusted number because the vendor’s actual matches that reconciliation. Revenue also exceeds the vendor’s $6,693.883M estimate and management’s earlier $6,670–$6,720M range. The beat is verified, but it does not by itself identify the quality or durability of the incremental profit.

GAAP operating income grows 8.3% to $2,354M, slower than sales. Non-GAAP operating income grows 7.2% to $2,974M. Net income grows just 3.1%, to $1,827M, while GAAP EPS grows 10.5%. A reduced diluted share count contributes to the per-share improvement: 395M this quarter versus 424M a year earlier. That distinction does not invalidate repurchases. It means business growth and denominator reduction should be evaluated separately, particularly when the company is spending more on repurchases than cash left after a compensation charge.

The new fiscal-year sales range is $26.576–$26.626B, versus $26.50–$26.60B previously. The midpoint rises $51M, a modest change against a roughly $26.6B business. Non-GAAP EPS guidance moves to $24.45–$24.50 from $24.35–$24.45, a $0.075 midpoint increase. The quarter was better than expected, yet the annual raise should not be described as a dramatic acceleration. The fourth-quarter targets still require $6.80–$6.85B sales and $6.30–$6.35 adjusted EPS. Those are the immediate measurements against which the next update can be graded.

Customers and reporting

The customer groups are not standalone profit segments.

Adobe changed its formal reporting structure in the first quarter of fiscal 2026. Digital Media, Digital Experience and Publishing and Advertising were combined into a single operating and reportable segment as management shifted toward unified selling and integrated product development. The latest available 10-Q explains that the CEO evaluates consolidated net income and resources. Customer-group subscription revenue is useful supplemental information, but it is not a disclosure of two independently measured operating profits. Allocating corporate profit to those groups by revenue share would create numbers the company has not reported.

Creative & Marketing Professionals subscription revenue reaches approximately $4.65B, up a reported 13%. Business Professionals & Consumers reaches $1.91B, up 16%. The former includes Experience Cloud and Creative Cloud flagship applications; the latter includes Acrobat and Express. The June investor datasheet backfills both groups across earlier quarters, enabling a meaningful revenue comparison despite the changed financial presentation. Its values are rounded to hundredths of a billion, so totals can differ from the sum of the displayed components. Those rounding differences are retained rather than forced to reconcile by altering a group.

The segment change improves alignment with management’s customer strategy but reduces the usefulness of old segment-margin narratives. Historical Digital Media gross profit and Digital Experience gross profit remain available in the saved filings; they should be labeled historical and are not carried forward as current group profitability. The current release does not independently reveal Firefly profit, the profitability of free Express users, or the cost of serving a particular AI cohort. Missing disclosure is part of the analytical boundary, not a license to infer a favorable margin from a product label.

Semrush further changes the comparison. Adobe completed the acquisition on April 28 for approximately $1.87B, principally cash. Q2 included about $40M customer-group subscription revenue and $480M ARR from the acquisition, while cash paid net of acquired cash was $1.56B in the first half. The Q3 release does not separately disclose the acquired revenue contribution. Consolidated growth therefore cannot be relabeled organic growth. The strategic fit with search and brand visibility may be useful, but the investment earns a return only when incremental cash compensates for the purchase price and integration costs.

Profit bridge

Growth has a cost, even before the non-GAAP adjustment.

Gross profit rises $651M to $5,997M, but cost of revenue grows 18.8%, faster than sales. The calculated gross margin slips roughly 57 basis points to 88.7%. Research and development increases 18.4% to $1,288M, sales and marketing increases 11.5% to $1,827M, and general and administrative costs increase 19.6% to $488M. Operating-expense amortization adds $40M. These are the disclosed cost lines; the release does not provide a clean incremental AI inference-cost bridge. Assigning the whole margin movement to generative AI would overstate what the available evidence establishes.

GAAP operating margin is approximately 34.8%, down about 147 basis points year over year. The adjusted margin is approximately 44.0%, down about 232 basis points. Both views therefore show a margin constraint rather than a discrepancy caused only by one accounting charge. Management still targets roughly 45% adjusted operating margin for the full year and roughly 44% for Q4. Those ranges support substantial profitability, but they make it difficult to justify a valuation that requires an immediate return to the older, higher adjusted-margin run rate.

Below operations, interest expense is $66M, investment gains are $21M and other income is $48M. Their net contribution is only $3M. Pretax income is $2,357M and tax expense is $530M, an effective rate of approximately 22.5%. The prior-year tax provision was $415M on $2,187M pretax income. A higher tax burden helps explain why net income growth is slower than operating growth. This is an important distinction for an investor looking at EPS: neither the lower share count nor a favorable adjusted tax convention turns tax expense into a nonexistent shareholder cost.

Earnings quality

Compensation is the largest recurring reconciliation item.

The $620M gap between GAAP and adjusted operating income includes $544M stock-based and deferred compensation, $58M amortization of intangibles and $18M acquisition-related expense. Compensation accounts for 87.7% of the operating adjustment. The cash-flow statement’s SBC add-back is $539M, slightly different from the broader non-GAAP compensation category. They are not interchangeable labels. Our owner-cash proxy deducts the actual cash-flow SBC add-back from free cash flow; the non-GAAP bridge preserves management’s separate $544M line.

Current net income of $1,827M reconciles to $2,424M adjusted net income after the operating additions, removal of $21M investment gains and a $2M income-tax adjustment. Adjusted EPS therefore answers a different question from the amount earned under GAAP. Adobe can legitimately use that measure to manage operations, but owners still surrender value through equity compensation or the cash needed to offset its dilution. Deducting SBC from free cash is an analytical way of acknowledging that cost; it does not claim employees were paid the same amount in cash this quarter.

The nine-quarter history also contains genuine unusual items. Fiscal 2024 includes the $1B Figma termination payment, while Q2 fiscal 2026 includes a $70M Publishing & Advertising goodwill impairment. Earlier fiscal 2026 loss-contingency charges also enter the reconciliation. These items should be explained, not treated as interchangeable with recurring stock awards. The abandoned Figma transaction is resolved for eligibility because a primary December 2023 filing documents termination, yet its economic cost remains relevant when judging acquisition discipline. A failed deal can be over without its financial lesson becoming irrelevant.

The balance of quality evidence is positive but qualified. Operating cash exceeds net income this quarter and over the trailing year. Revenue remains overwhelmingly subscription-based. Against that, current gross and operating margins are lower, tax rates have increased, and the company continues to make substantial compensation adjustments. The quality score is therefore not a binary rejection of non-GAAP reporting. It is a requirement that the valuation charge recurring costs, retain taxes and distinguish the repeatable cash engine from transactions that should not recur every year.

Cash and capital

The franchise generates cash; distribution discipline still matters.

Quarterly operating cash flow is $2,523M, versus $2,198M a year earlier. Capital spending of $85M leaves $2,438M free cash flow, up 14.7%. Deducting $539M SBC produces a $1,899M owner-cash proxy. The business converts earnings into cash well, and the small capital-expenditure line reflects a software model rather than a retailer’s physical expansion. However, the operating cash statement also includes deferred-revenue movements and other working-capital changes. A single quarter should not be annualized without considering collections and contract timing.

The three audited fiscal years provide a stronger base. Free cash flow increased from $6,942M in fiscal 2023 to $7,873M in fiscal 2024 and $9,852M in fiscal 2025. After the SBC charge, those amounts are $5,224M, $6,040M and $7,910M. The latest trailing four quarters produce $10,592M FCF and $8,521M after SBC. This is a substantial cash franchise. The valuation question is whether that owner cash can grow as AI products expand usage while prices, distribution and serving costs change.

Repurchases are not automatically value creation. Adobe paid $2,232M for stock this quarter, compared with $1,899M of free cash after our compensation charge. The cash purchase exceeds that proxy by $333M. Management describes approximately 9.5M shares repurchased, implying a rough $234.95 average cash price. That is below our $300 appraisal, which supports the economic case for the purchases, but the share number is rounded and settlement timing can differ from cash-flow presentation. The estimate is a diagnostic, not an execution-price audit.

Quarter-end cash of $4,359M plus short-term investments of $1,280M totals $5,639M. Current and long-term debt carrying amounts total $6,363M, leaving $724M net debt on that accounting basis. The current earnings release does not include the complete principal, swap and discount reconciliation. The Q2 10-Q reported $6.15B senior notes and $500M commercial paper; current-quarter financing cash includes $250M debt repayment. We use current carrying debt explicitly and flag the missing par bridge. We do not silently reconstruct it from incomplete movements.

Ordinary lease expense remains inside operating earnings and cash flows, so the enterprise models do not subtract the full $399M lease liability a second time. Semrush’s completed purchase is already reflected in cash, and no future repurchase is assumed in the model denominator. The release’s 395M diluted weighted-average shares are an analyst valuation convention, not an exact current common-share count. The latest available filing-cover count is 397.5M at June 11; Q4 guidance assumes approximately 389M diluted shares. Showing all three prevents false precision and the temptation to count a future share reduction without paying for it.

Valuation

Three routes; one explicit set of costs and uncertainty.

The fair-value anchor is $300. It comes from a $299.28 blend of three methods rather than from an old share-price high or a consensus target. The routes deliberately disagree: the direct earnings and operating-profit approaches are above $300, while the discounted-cash-flow result is approximately $263. That lower cash result is useful discipline. It shows that even a business with substantial reported cash does not deserve an unlimited multiple when future growth, competition and recurring compensation remain uncertain.

Route one applies 17 times management’s fiscal-year GAAP EPS midpoint of $18.145, producing $308.47. This route retains stock compensation and the other charges embedded in the annual GAAP outlook rather than silently normalizing them away. The multiple assumes the installed base and recurring contracts retain economic value while growth moderates. Seventeen times is an analyst judgment, not a company target or an assertion that Adobe must regain its historical premium. Because this is an equity earnings multiple, no additional cash is added to the resulting per-share value.

Route two assumes $27.5B revenue over the coming twelve months and a 35% normalized GAAP operating margin, giving $9,625M operating income. This is a forward model, distinct from current reported revenue and from the $27.5B ARR measure that happens to have the same numerical magnitude. At thirteen times operating income, enterprise value is $125,125M. Subtract $724M net debt and divide by the fixed 395M diluted-share convention to obtain $314.94 per share. The 35% margin includes recurring compensation and normal rent. We do not apply an adjusted-profit multiple while forgetting the largest recurring adjustment.

Route three begins with $8.2B unlevered owner cash after recurring compensation and ordinary reinvestment. This is an explicit forward assumption, not a relabeling of reported operating cash. It is below the observed $8.521B trailing levered FCF-less-SBC proxy, even before considering an interest add-back, providing a conservative starting point without inventing a precise conversion. The model grows that cash 5% annually for five years, discounts at 11% and uses 2.5% perpetual growth. It assumes continuing relevance and modest expansion, not domination of every AI application market.

The five cash flows are $8.200B, $8.610B, $9.0405B, $9.4925B and $9.9672B. Their present values sum to approximately $33.154B. The discounted terminal value is approximately $71.328B, giving $104.482B enterprise value. Net debt reduces equity value to $103.758B; divided by 395M model shares, the result is $262.68. Terminal value remains a large share of the answer, which is why the sensitivity grid and uncertainty classification matter. A well-formatted model is not precise merely because it produces two decimals.

Weights of 40% for forward GAAP earnings, 35% for operating profit and 25% for DCF produce $299.28, rounded to $300. The price-implied DCF requires approximately $7.627B of initial cash under identical growth and discount assumptions. Our $8.2B starting point provides a cushion, but not one that makes the thesis insensitive to adverse outcomes. At $244.20, the model upside is 22.9% while the discount to fair value is 18.6%; those percentages have different denominators. The High-uncertainty buy threshold remains $225, not a rounded version of today’s price.

Our bear/base/bull cases are $190/$300/$410 with 25%/50%/25% probabilities, giving a weighted value of $300. The bear combines roughly $15.83 EPS and twelve times earnings; the base uses $18.75 and sixteen times; the bull uses roughly $22.78 and eighteen times. These are scenario conventions, distinct from each detailed valuation route. The downside allows for weaker pricing and margin pressure; the upside requires sustained monetization and operating efficiency. The probabilities are editorial judgments and must not be interpreted as observed market frequencies.

Wall Street

Available targets precede the new earnings release.

The saved target feed has no post-print analyst response at the research cutoff. Stifel’s $225 target is dated September 9; Mizuho’s $260 target is dated September 8; Barclays’ $295 target is dated September 3; RBC Capital’s $315 target is dated September 2. They provide a range of pre-earnings expectations, not a current consensus judgment on the newly disclosed quarter. Calling them fresh reactions would give the audience a false sense that analysts had already evaluated the release and conference call.

The associated rating feed shows Stifel Hold, Barclays Equal Weight and RBC Outperform on those respective dates. The saved current subset does not establish a matching Mizuho rating, so that label remains unavailable. Our HOLD and $300 appraisal sit above Stifel’s target and near Barclays and RBC, but numerical proximity does not prove agreement about AI economics. Our explicit recurring-compensation charge, fixed share denominator and conservative cash-flow model are the reasons for the appraisal. Those assumptions remain visible even if a broker changes its target after this source cutoff.

Consensus also has a definition boundary. The current quarter’s vendor EPS actual of $6.13 matches non-GAAP EPS; GAAP is $4.62. The estimate comparison is therefore made on that adjusted basis. Annual estimate records are forecasts and can have different update times and adjustments. This packet does not transform old forecast rows into reported history or claim an estimate-revision trend from a single snapshot. Company guidance is the primary forward operating reference, with broker targets serving as a separately dated comparison.

AI and the outlook

Reach, contracted revenue and cash monetization are different measures.

Adobe highlights more than 150% year-over-year growth in AI-first ARR and a milestone of one billion monthly active users across creativity and productivity solutions. Both can be evidence of adoption. Neither tells us how many users pay, what each pays, or the cash margin on the incremental workload. The current release does not state an absolute AI-first ARR level, while Q2 had disclosed more than $500M. It would be wrong to manufacture the current dollar amount by multiplying the earlier lower bound by the new growth rate, which compares a different prior-year quarter.

Total Adobe ARR is $27.50B, but even this familiar metric requires a definition check. Adobe measures annual contract value using currency rates set in December and revalues the previous fiscal year for the next measurement year. The June datasheet shows prior Q3 at $24.74B using December 2025 rates, versus an earlier reported $24.28B using December 2024 rates. Comparing $27.50B with the same-FX $24.74B gives 11.16% growth. The annual reset also changes FY2025 ending ARR from $25.20B to $25.66B. That $460M is currency revaluation, not newly signed contracts.

RPO reaches $22.16B, up 8.4% year over year but slightly below $22.27B in Q2. The company says 67% is current RPO. The datasheet defines that percentage with an exclusion for certain non-cancellable and nonrefundable enterprise committed funds, so it should not be casually multiplied into a precise next-year revenue guarantee. Contract timing can explain sequential variation, but persistent RPO growth materially below subscription sales would warrant closer attention. We want adoption, bookings and cash to tell a coherent story over several quarters.

Leadership will also change. Anil Chakravarthy becomes president and CEO on December 1, while Shantanu Narayen becomes executive chair. The current September release still correctly identifies Narayen as CEO and Steve Day as interim CFO. David Wadhwani is due to step down from his creativity and productivity leadership role on September 27 and serve as a senior advisor during transition. These are meaningful operating changes, not automatic evidence of failure or success. The next team will need to demonstrate how broader usage becomes durable paid demand without sacrificing the margin and cash characteristics that make Adobe valuable.

Competing cases

A valuable installed base can coexist with disruption risk.

The strongest bullish case begins with the existing cash engine. Subscription revenue is growing at a double-digit rate, contracts support visibility, and the company continues to convert income into cash while reducing the share count. AI may expand the number of people who can create useful content and increase the value of integrated workflows. Acrobat, Express, flagship creative applications and enterprise marketing tools can reinforce distribution. If that reach improves paid conversion and product utility, the current earnings multiple could understate the durability of the franchise.

The bearish case does not require Adobe to disappear. It requires only that competitive tools weaken pricing, change workflow ownership or increase the cost of keeping customers engaged. A large free audience may be expensive to serve without paying proportionately. Enterprise customers can demand concessions, and a slower contract-growth rate can eventually reach recognized revenue. Margin contraction already exists in both reported and adjusted results, although its precise causes are not separately quantified. A valuation based on yesterday’s profitability would be vulnerable if these changes prove structural rather than transitional.

Capital allocation can amplify either outcome. Repurchases below value improve per-share ownership; repurchases that offset employee issuance or absorb cash before weaker results do less for outside owners. Acquisitions can strengthen capabilities, but the Figma termination payment and current goodwill/contingency charges show why capital decisions deserve independent scrutiny. Semrush is a completed transaction whose future return remains to be demonstrated. We credit integration opportunities without assuming that every acquired dollar of ARR produces the same return as organic subscription expansion.

Our HOLD reflects the balance of these arguments. The business is profitable enough and the cash record strong enough that a categorical bearish conclusion would ignore substantial evidence. The uncertainty around AI monetization, margin durability, leadership and share count is also too meaningful to waive the margin-of-safety threshold merely because the stock is well below a former high. The model can be wrong in both directions. The response is an explicit price discipline and dated operating tests, rather than a confident story that explains every available number after the fact.

What changes the call

Numeric tests tied to the next operating disclosures.

The immediate checkpoint is the fiscal fourth quarter. Revenue at or above the $6.825B guidance midpoint and adjusted operating margin of at least 44% would support execution. A result below the $6.80B sales floor or below 43% adjusted margin would weaken it. Because the margin measure excludes compensation, we also want the GAAP margin to remain near the guided 34% rather than letting adjusted performance conceal a widening recurring-cost burden. These are analyst decision rules layered onto disclosed company guidance, not additional management promises.

For the following reports, trailing FCF less SBC should remain at least $8.5B to support the owner-cash thesis. A decline below $7.5B would call for a material model review. Same-currency total ARR growth of at least 10% and RPO growth of at least 8% would support the assumed continuing expansion; sustained growth below 8% ARR or 5% RPO would be a warning. Each comparison must retain acquisition and currency qualifications. The tests are designed to assess economic direction, not to reward a favorable change in reporting scope.

The price-led entry is $225 or below with the thesis intact. A cheaper quote after a permanent reduction in earnings power would require a lower fair-value estimate too. Conversely, clear evidence that AI reach converts into paid recurring cash at stable margins could justify a higher cash forecast and a different call before that price is reached. December 2026 and March 2027 are expected review windows, not confirmed earnings-release dates. The leadership transition has a separately announced effective date of December 1. We will grade facts against the same definitions used here rather than changing the test after the result arrives.

Sources and calculations

Every displayed figure has a source or an explicit analyst assumption.

The current earnings source is Adobe Q3 FY2026 Exhibit 99.1, accession 0000796343-26-000147. The latest available quarterly report is the Q2 FY2026 Form 10-Q; the new Q3 10-Q is not yet in the saved submissions. The June 11 investor datasheet supplies historical customer-group and ARR comparisons with its definitions. The current earnings call and Q3 supplemental documents were pending at the source cutoff. This packet can proceed from the public release under the production rule, and it does not pretend to summarize a call it has not read.

The source dataset stores full quarterly income statements, balance sheets, cash-flow tables, GAAP-to-non-GAAP reconciliations and source column headers for nine consecutive quarters. Historical formal segment tables are retained where disclosed; annual Q4 source tables are identified as annual rather than mislabeled quarterly. Three audited cash-flow years are preserved. The current release provides a complete corporate earnings bridge but no exact current cover-page common-share count, current par-debt reconciliation, current standalone group profit or absolute AI-first ARR. Those gaps remain explicit in the dataset.

The packet’s arithmetic is generated by episode-local research code, with the canonical packet renderer and checker used in place. Free cash flow is operating cash less property-and-equipment purchases. Owner cash subtracts the cash-flow SBC add-back. The DCF uses an explicitly assumed unlevered owner-cash stream, after recurring compensation and reinvestment, and subtracts reported net debt once. Forward GAAP P/E adds no cash again. The enterprise operating-profit route uses a GAAP-like margin with rent and compensation included. Price, target and reaction sources are market feeds; they are not mislabeled as company financial statements.

Data cutoffs matter. The after-market trade and bid/ask are timestamped market observations; they are not an average session price. Earlier Street targets remain dated pre-print opinions. Prior unavailable YouTube oEmbed records were independently proven to predate this release by a September 7 manifest snapshot; the latest accessible prior long episode is Q2 FY2026. This research does not reuse its record-quarter-versus-selloff angle. The new analytical emphasis is the relationship between reach, contracted growth, compensation and cash left for owners.