Adobe Inc. (Nasdaq: ADBE)

1. Executive Snapshot

2. What the Company Does

Professionals and businesses need to make images, video, layouts and documents to a standard that clients, printers, agencies and colleagues can open and keep working on. Adobe sells those tools: Photoshop, Illustrator, Premiere Pro and InDesign inside Creative Cloud; Acrobat for PDF; Express and Firefly for lighter and AI-generated work. To large enterprises it also sells Experience Cloud — Adobe Experience Platform, analytics, content management, Workfront and GenStudio — to plan, produce, personalise and measure marketing content (FY2025 10-K).

The unit of economics is one paid subscription seat. A designer on the Creative Cloud Pro annual plan pays $69.99 a month in North America (adobe.com, web). Adobe collects through adobe.com, app stores, resellers, or multi-year enterprise term licence agreements, and recognises the revenue ratably over terms of one to 36 months, which left $7.0B of deferred revenue at FY2025 year-end (FY2025 10-K). The seat costs almost nothing to serve: Digital Media cost of revenue was $841M against $17.6B of revenue, a 95% gross margin (FY2025 10-K). The economics are therefore about acquiring and keeping the seat, not about delivering it.

Subscriptions were $22.9B of FY2025 revenue (96%). Perpetual product sales ($325M, −16%) and services ($540M, −10%) are shrinking remnants (FY2025 10-K). Adobe also cuts subscription revenue by customer group. Creative & Marketing Professionals — Creative Cloud flagship apps plus all of Digital Experience — was $16.3B (+11%). Business Professionals & Consumers — Acrobat and Express — was $6.5B (+15%) (FY2025 10-K).

Source: FY2025 10-K, segment note. Gross margins as reported.

The two businesses have different economics. Digital Experience carries hosting, data and implementation services, so it earns 72% gross margin against 95% for Digital Media. Despite roughly $7.9B spent on Magento, Marketo and Workfront (FY2018–FY2021 10-Ks), Digital Experience stayed at 25–30% of revenue over the decade (FY2016–FY2025 10-Ks, inferred). The profit engine has always been the creative and document franchise.

Adobe is losing two small lines. Its standalone Adobe Stock business, about a $450M book, is declining faster than planned as generative images replace licensed stock photos (Q1 FY2026 call). Publishing & Advertising took a $70M goodwill impairment in Q2 FY2026 (Q2 FY2026 10-Q). Both are cases of content that AI now produces rather than tools that AI assists.

From FY2026 Adobe reports a single segment and has retired Digital Media ARR in favour of Total Adobe ARR (FY2025 10-K; Q2 FY2026 10-Q). The 95% versus 72% margin split above is therefore the last one investors will see.

3. Industry, Competitive Position & Moat

The industry turns intent into production-ready visual, video, interface and document assets. The chain has four layers: capture and stock content, generative models, authoring and workflow tools, and distribution (social, web, ad networks and marketing stacks). Disclosed profit sits in the tools layer: Adobe's Digital Media earns 95% gross margin, Figma 88% non-GAAP (FY2025), and Autodesk about 91% (FY2026) (company filings, web). Stock content is squeezed — Shutterstock's content revenue fell 5% in Q3 2025 as subscribers declined (Shutterstock IR, web). Model labs disclose no margins and, for tool vendors, are a cost input.

The relevant market is global and splits by user type, and professional editing is highly concentrated. The UK CMA found Photoshop held over 80% of raster editing with the next rival near 5%, and Illustrator over 70% of vector editing with the next below 10% (CMA provisional findings, Nov 2023, web). In UI and product design the position reversed: Figma held over 80% and Adobe XD 5–10%. In template-based design for non-designers Canva leads, with about $4B ARR and 265M monthly users in 2025 (company figures reported by trade press, web). No independent share data after 2023 was found.

Which barriers actually bind

• Formats and workflow bind. PSD, AI and PDF are the files clients, printers and agencies exchange, so leaving Adobe means breaking hand-offs with other people, not just learning a new app. Cheaper tools such as the old perpetual-licence Affinity had not dented Photoshop's or Illustrator's share by 2023 (CMA, web).
• Trained labour binds. Studios hire people who already know Photoshop and Premiere, so a tool switch carries retraining cost across the whole team (inferred).
• Enterprise breadth binds. Only Adobe spans raster, vector, video, PDF and marketing under one contract, with governance and IP indemnity for Firefly outputs (FY2025 10-K; Adobe Firefly legal FAQ, web). Over 150 customers exceed $10M ARR (Q4 FY2025 call).
• Code and capital do not bind. Adobe's own 10-K describes "limited barriers to entry" and "customer price sensitivity" (FY2025 10-K). Figma showed a browser-native, multiplayer tool can take a new category from a file-based incumbent.

Competition now comes from below and from the side. Adobe's 10-K stopped naming rivals in FY2021; it had listed Figma, Affinity, InVision and Sketch by FY2019 and added "AI companies" in FY2023 (FY2016–FY2025 10-Ks). Outside evidence shows faster growth elsewhere: Figma's revenue grew 48% in Q2 2026 with 136% net dollar retention (Figma release, web). Canva's revenue grew 25% in Q2 2026 (shareholder update via trade press, web), and it made Affinity free in October 2025 (web). Apple's Creator Studio bundle costs $12.99 a month against $69.99 for Creative Cloud Pro (Apple Newsroom; adobe.com, web).

Generative AI adds entrants that skip the tool entirely. ChatGPT users made over 700M images in the first week of its 2025 image launch (OpenAI via TechCrunch, web), and Google reported over 230M Veo 3 videos (Alphabet Q3 2025, web). Adobe's response is to aggregate rather than out-build them. Firefly now hosts more than 30 models, including Google, OpenAI, Black Forest Labs and Runway (Q1–Q2 FY2026 calls). For partner models Adobe leaves the commercial-safety judgment to the creator (adobe.com partner-models page, web).

Two things could weaken the position. Regulators have effectively certified Adobe as dominant: the CMA's theory of harm was losing Figma as a future entrant into raster and vector editing, and Adobe paid a $1B fee to walk away (2023-12-18 8-K; CMA). Buying the next Figma is no longer available as a defence. Meanwhile the price gap is widening, with Adobe pricing up at the top while Canva, Apple and Blackmagic price down at the bottom.

The winners in this industry own the canvas where work is done and shared, run a free or cheap funnel, control their model costs, and offer enterprise governance (industry research, inferred). Adobe has the canvas and the governance in professional workflows. It has been weak on the free funnel, and the June 2026 freemium pivot is a direct, costly attempt to fix that. On model cost it depends partly on partners. Adobe is that kind of company in its professional core, and only partly at the edges where the users are multiplying.

4. Growth Engine

Revenue grew from $5.85B in FY2016 to $23.77B in FY2025, a 16.8% compound rate (FY2016–FY2025 10-Ks, inferred). The early decade combined the end of the shift from boxed software to subscription (subscription share rose from 78% to 92% by FY2021), strong seat growth, and acquisitions. Digital Media ARR, as reported each year-end, went $4.01B, $5.39B, $6.83B, $8.40B, $10.18B, $12.24B, $13.97B, $15.17B, $17.33B and $19.20B from FY2016 to FY2025 (10-Ks). Adobe revalues ARR at new exchange rates every year, so those year-ends cannot be chained exactly.

In FY2025 growth was almost entirely organic, since acquisitions cost only $17M (FY2025 10-K cash flow). Management said over 75% of Digital Media net new ARR came from more subscriptions, cross-sell and upsell, with the rest from value-based pricing (Q4 FY2025 call). The FY2018–FY2019 period was different. Revenue grew about 24% in each year, but Digital Experience jumped 31% in FY2019 on the Magento and Marketo purchases, and the acquired contribution was never separately disclosed (FY2019 10-K).

FY2026 headline growth now includes acquired revenue. Semrush closed on 28 April 2026 for $1.874B and brought about $480M of ARR (Q2 FY2026 10-Q and call). The Q3 10-Q confirms Total Adobe ARR includes Semrush but gives no pro forma figures, calling the impact not material; management also declined to quantify it on the call (Q3 FY2026 10-Q and call). The decomposition below strips it out.

Sources: FY2025 10-K; Q3 FY2026 10-Q ($27.50B vs $24.74B revalued prior-year ARR); Q2 and Q3 FY2026 calls. Acquired share = ~$480M Semrush ARR over the prior-year ARR base. The FY2026 target was 10.2% before Semrush (December 2025) and stayed at 10.2% after including it (June and September 2026), so the organic target fell by roughly Semrush's size.

Net new ARR in Q3 FY2026 was about $0.4B, which an analyst put at down 36–37% year on year; management did not dispute it (Q3 FY2026 call). Two deliberate choices explain part of it. In June 2026 Adobe deferred planned Creative Cloud price increases and pushed new users into free tiers first, and management attributed roughly half the ARR shortfall to each (Q2 FY2026 call). The Stock decline explains a further 0.3 points of Q1 ARR growth (Q1 FY2026 call).

Drivers, ranked by current impact

Two drags run the other way. Generative AI is replacing standalone stock-photo licences, and the freemium pivot delays paid conversion into 2027, according to management (Q2 FY2026 call). Whether the pivot is an investment or a concession depends on free-to-paid conversion, which Adobe does not disclose.

“I'm actually really happy that we didn't focus on the pricing actions, because that, while it may have provided some short-term relief, would not be as critical as continuing to drive new user adoption.”

Shantanu Narayen, Q3 FY2026 earnings call, 10 September 2026

5. Margin, Cash & Capital Allocation

Gross margin runs at 89% because the marginal cost of a software seat is hosting and, increasingly, AI inference (FY2025 10-K). Operating margin rose from 25.5% in FY2016 to 36.8% in FY2021 because sales and marketing fell from 32.6% to 27.3% of revenue and G&A from 9.9% to 6.6% (FY2016–FY2025 10-Ks, inferred). Since then operating leverage has stalled. R&D climbed back to 18.1% of revenue by FY2025, and stock-based compensation rose to 8.2% of revenue, a real cost that free cash flow does not show.

GAAP operating margin by year, FY2016 to FY2025: 25.5%, 29.7%, 31.5%, 29.3%, 32.9%, 36.8%, 34.6%, 34.3%, 31.3% and 36.6% (10-Ks, inferred). FY2019 was pulled down by Marketo and Magento amortisation, and FY2024 by the $1B Figma termination fee (about 36.0% without it, inferred). Non-GAAP operating margin was about 46% in FY2025 (2026 proxy) and is guided to about 45% in FY2026 (Q3 FY2026 call).

AI cost is starting to show in cost of revenue. Digital Media cost of revenue rose 24% in FY2025 while its revenue rose 11%; the 10-K names AI inferencing among subscription costs (FY2025 10-K). In Q3 FY2026 subscription cost of revenue rose 24% against 14% subscription revenue growth, with hosting and data centres contributing 22 points (Q3 FY2026 10-Q). That is the first quarter where it shows in the total: gross margin fell to 88.7% from 89.3% a year earlier, and GAAP operating margin to 34.8% from 36.3% (Q3 FY2026 10-Q, inferred).

Cash conversion is very high. FY2025 operating cash flow was $10.03B, capex only $179M (0.8% of revenue), and free cash flow about $9.85B, or 41% of revenue (FY2025 10-K, inferred). Upfront annual billing funds the business before service is delivered, so working capital is a source of cash.

Sources: 10-Ks for each year; margins and free cash flow (operating cash flow less capex) are my calculation. FY2016 and FY2021 were 53-week years. FY2024 includes the $1B Figma fee in operating income and operating cash flow. Segment boundaries changed in FY2018 and FY2020, and ASC 606 was adopted in FY2019 without restating prior years; the consolidated lines above are unaffected, but segment trends across those years are not comparable.

Where the cash went, FY2016–FY2025

• Buybacks: $45.7B, about 80% of cumulative free cash flow of $56.8B. In FY2024 and FY2025 buybacks exceeded free cash flow; FY2025 alone was $11.28B (FY2016–FY2025 10-Ks).
• Acquisitions: $9.7B, 65% of it in FY2018 (Magento, Marketo), plus the $1B Figma fee in FY2024.
• Capex: $3.0B, about 1% of revenue over the decade.
• Debt: net $4.2B added ($9.4B issued, $5.2B repaid). Net cash fell from about $2.9B to about $0.4B (inferred).
• Dividends: none (FY2016 and FY2025 10-Ks).

The ranking shows a management team that treats share count as its main lever for per-share growth. Diluted EPS compounded at about 24.5% against 16.8% for revenue (inferred). The pattern continued in FY2026. In the first nine months Adobe bought back $6.82B (26.1M shares) and spent $1.56B net on acquisitions, against free cash flow of about $7.47B (Q3 FY2026 10-Q, inferred). At Q3 it held $5.64B of cash and short-term investments against $6.15B of notes ($1.35B due within a year) and $250M of commercial paper, a net debt position of about $0.8B (Q3 FY2026 10-Q, inferred). A new $25B authorisation runs to April 2030, with $24.55B left at Q3; the previous $25B was used in under 11 quarters (Q3 FY2026 10-Q; Q2 FY2026 call).

M&A has shifted to tuck-ins since the Figma deal was blocked. Semrush ($1.874B, of which $1.251B goodwill) closed in April 2026 (Q2 FY2026 10-Q). Adobe agreed on 24 June 2026 to buy Topaz Labs, a private AI video- and image-enhancement company, for about $340M, mostly in cash; it is expected to close in Q4 FY2026 (Q3 FY2026 10-Q). It is not in any reported figures above.

6. Cyclicality, Constraints & What to Monitor

End-market exposure is by user type rather than industry. Creative & Marketing Professionals were about 72% of FY2025 subscription revenue and Business Professionals & Consumers about 28% (FY2025 10-K, inferred). By region, revenue was 59% Americas, 27% EMEA and 14% APAC, almost unchanged from FY2016 (10-Ks). No customer or distributor exceeds 10% of revenue (FY2025 10-K).

Macro sensitivity is low. Revenue grew 15% in pandemic-hit FY2020 (FY2020 10-K), because a seat is a small, necessary cost for the people who use it and subscriptions are recognised over time. The sensitivities that do exist are currency — FX added about $116M to Q2 FY2026 revenue (Q2 FY2026 10-Q) — and enterprise spend scrutiny, which the 10-K links to tariffs and longer AI contracting cycles (FY2025 10-K).

Where Adobe sits now: record margins at the slowest growth of the decade. Gross margin of 89.3% in FY2025 was the highest in the series, and GAAP operating margin of 36.6% sat within 0.2 points of the FY2021 peak (10-Ks, inferred). Revenue growth ran 25%, 24% and 24% in FY2017–FY2019, 15% in FY2020 and 23% in FY2021 (a 53-week year), then 10–11% in every year from FY2022 to FY2025 (10-Ks, inferred). Organic ARR growth is now about 8–9%, and cRPO growth of 9% in Q3 FY2026 was the first single-digit reading since early FY2023 (Q3 FY2026 call); Q3 gross and operating margins were also both below the prior year (Q3 FY2026 10-Q). This points to maturing demand rather than a cyclical dip, and the margin peak may already be passing (inferred).

The downside mechanism is an unbundling squeeze rather than a recession. Free or cheap professional tools (Affinity, DaVinci Resolve, Apple's bundle) and AI-native generation would hollow out individual and small-business seats first. Because revenue is recognised over terms of up to 36 months, the damage would appear in net new ARR and cRPO a year or more before it reaches reported revenue (FY2025 10-K risk factors, inferred). Rising inference cost would hit gross margin at the same time; Canva cut its 2026 growth forecast from 30% to 20% while working down AI serving costs (trade press, web).

Durable vs borrowed

Leading indicators

7. Risks, Unknowns & Questions for Deeper Work

• AI substitution at the individual seat. If a prompt produces a good-enough image, the casual Creative Cloud buyer lapses. The Stock book is the leading case: a ~$450M line shrinking faster than planned (Q1 FY2026 call). The BLS projects US graphic-designer jobs falling 2% over 2025–2035 (BLS, web). This compounds with the widening price gap in Section 3.
• Model dependence and cost. Firefly now relies partly on Google, OpenAI and other labs, and the 10-K warns other companies may "prevent, limit or interfere" with access to third-party models (FY2025 10-K). The same partners compete for end users, so Adobe could end up paying its rivals for the capability its customers value most.
• Leadership during the product transition. Anil Chakravarthy, who ran the enterprise marketing business, becomes CEO on 1 December 2026. The creative business head, David Wadhwani, leaves on 27 September, and the CFO seat has been interim since June (2026-09-08 and 2026-06-11 8-Ks). The creative franchise — 74% of revenue at 95% gross margin — loses its product owner in the middle of its hardest transition. Say-on-pay passed with only about 50.5% support in April 2026 (2026-04-21 8-K, inferred).
• Less disclosure. A single segment from FY2026 and the retirement of Digital Media ARR remove the tools to see whether creative margins or growth are eroding. Deterioration in the core could hide inside a consolidated figure that includes acquired marketing revenue.
• Balance-sheet drift. Buybacks above free cash flow moved Adobe from about $2.9B of net cash in FY2016 to about $0.8B of net debt at Q3 FY2026, with $1.35B of notes to refinance in 2027 (Q3 FY2026 10-Q, inferred). This is survivable alone, but combined with slowing growth it removes the cushion that has funded both buybacks and M&A.
• Constrained defence. After the Figma decision, buying a fast-rising rival is effectively off the table (CMA). The March 2026 DOJ settlement ($150M, with rules on cancellation-fee disclosure and easy cancellation; DOJ, web) removes part of the friction that used to slow churn.

What the sources could not answer

• Paid seat counts, ARPU, gross and net retention, and churn — never disclosed (NOT FOUND).
• The split of ARR between individuals, teams and enterprise agreements (NOT FOUND).
• Dollar AI inference cost, and gross margin by product after the FY2026 segment merger (NOT FOUND).
• Free-to-paid conversion from the freemium funnel, and the timing of any restart of Creative Cloud price increases (NOT FOUND).
• Semrush's revenue and ARR contribution after the acquisition — the 10-Q calls it not material and management declined to quantify it (NOT FOUND).
• Independent market-share data after 2023, PDF-editor share, and Canva margins (NOT FOUND).
• Whether FY2026 performance-share ARR targets were reset after the freemium pivot, and the new CEO's capital-allocation stance (unknown).

Before forming a thesis, an investor would need to settle whether seat counts in the creative professional base are still growing, and whether AI revenue per user is rising faster than AI cost per user.

8. Investor Takeaways

• What it is: the toll road for professional visual content and PDF, sold as subscriptions, with a slower enterprise marketing business attached.
• Economic engine: a seat that costs almost nothing to serve (95% gross margin) produced ~$9.85B of FY2025 free cash flow, nearly all of which went to buybacks.
• Main growth lever: broadening the paid base through Acrobat, Express and free creative tiers, then monetising AI through credits and higher plans — which has to offset a professional core now growing in single digits organically.
• What could break it: AI tools and cheap bundles thinning the individual seat base while inference costs rise, during a leadership change and with less disclosure to see it.
• What to monitor: organic net new ARR and cRPO, freemium conversion, gross margin under AI load, and Figma's and Canva's growth.