Applied Materials

Applied Materials is the world's largest supplier of process equipment and services to semiconductor manufacturers. Primary products include chemical vapor deposition (CVD), atomic layer deposition (ALD), plasma etch, physical vapor deposition (PVD), and advanced packaging deposition systems. Revenue derived from three segments: Semiconductor Systems (foundry/logic, DRAM, memory), Applied Global Services (spare parts, maintenance, services), and Other (legacy display, sundry).

Applied Materials is the largest process-equipment supplier to semiconductor fabs, spanning deposition, etch, CMP, thermal, and advanced packaging tools plus the services installed base around them. It is a picks-and-shovels exposure to the AI factory buildout rather than an AI application company.

Applied Materials sells wafer-fab materials-engineering tools (deposition, etch, CMP, RTP, implant, metrology/inspection, wafer packaging) plus installed-base services. The 10-Q restates two reportable segments: Semiconductor Systems and Applied Global Services; effective Q1 FY2026 the 200mm equipment line moved from AGS into Semiconductor Systems and corporate support costs are fully allocated (prior periods recast). Source: Q3 FY2026 10-Q Note 14.

Applied Materials, Inc.

• monetization model: One-time semiconductor/display capital equipment sales; recurring services, spares, consumables, upgrades, refurbished tools via Applied Global Services (AGS); not SaaS. 10-K FY2024, Item 1 Business; [sec.gov (Acc. 0000006951-24-000057)]
• revenue mix: FY2024 net sales ~$27.2B: Semiconductor Systems ~74%; Applied Global Services ~23%; Display ~3%. [10-K FY2024, Segment Note; same URL/accession]
• who pays: Semiconductor device makers/fabs (foundry, logic, memory) and display manufacturers; capital equipment and service budgets. [10-K FY2024, Item 1; same URL/accession]
• margin_value_capture: FY2024 GAAP gross margin ~47.8%, operating margin ~29.0%; value accrues in leading-edge process tools, installed-base services, spares/upgrades, and high switching costs. [10-K FY2024, MD&A/segment note; same URL/accession]

• customer classes: Semiconductor Fabs - IDM Buyers (incl. foundry/memory/logic fabs), display panel manufacturers

Applied Materials is a capital-equipment and services company supplying semiconductor manufacturing equipment, process technology, and services to semiconductor, display, and adjacent industries. Its economic model is to sell high-margin manufacturing tools into wafer-fab expansions/upgrades and then attach a service-and-spares stream over the equipment’s multi-decade life.

From [AMAT 10-K FY2024] — as present in training and needs primary source validation — Applied Materials reported:

• Net sales: $27,176 million
• GAAP gross margin: 47.5%
• GAAP operating margin: 29.0%
• GAAP net income: $7,176 million
• GAAP diluted EPS: $8.61

The margin structure is the core of the business model: gross margin near 47.5% reflects pricing power and high-value process steps; operating margin near 29.0% reflects operating leverage and discipline.

From [AMAT 10-K FY2024], segment net sales were approximately:

All values above are from [AMAT 10-K FY2024], filed 2024-12-13, accession no. needs validation, accessed 2026-04-26. Percentages are derived from reported segment totals and may not sum exactly to 100 due to rounding.

The semiconductor systems segment is the dominant, most cyclical business: it sells manufacturing equipment for wafer processing. Applied Global Services (AGS) is the recurring/recurrence-bent part of the model: spares, service contracts, tool upgrades, remanufactured equipment, and data/software offerings. Display is small but volatile.

AMAT monetizes through three distinct streams, per the business description in [AMAT 10-K FY2024]:

1. Equipment sales — large-ticket tools sold to semiconductor and display manufacturers. Revenue is recognized upon acceptance/transfer of control. The price point for advanced deposition/etch tools is typically multi-million dollars per chamber; exact ASPs are not disclosed at the product level and require primary source validation.
2. Services and spares — the AGS segment sells service contracts, spare parts, consumable components, legacy tool upgrades, and remanufactured equipment. This has higher recurrence than new tool sales and smooths cyclicality.
3. Software/data analytics — AMAT sells factory-automation, yield, and data-analytics offerings under Applied SmartFactory/AIx type labels. This is included in services and is not separately disclosed at a financial level.

Who pays: semiconductor manufacturers (foundry, logic, memory), integrated device manufacturers, packaging/advanced-packaging users, and flat-panel/OLED display manufacturers. The buyer is not a consumer; it is a fab-owner’s procurement and process-engineering organization.

The best unit-economics view from public filings is segment operating-margin contribution by revenue mix, because AMAT does not disclose enough product-level unit cost to compute a true per-tool margin.

From [AMAT 10-K FY2024] — segment operating income figures are approximate and needs primary source validation:

The corporate/unallocated line is derived by subtracting segment operating income from total GAAP operating income; it is not a direct primary-source line in this note and should be validated.

Worked calculation: revenue-mix shift sensitivity

Formula:

\[

Overall\ OM = \sum_i w_i \times OM_i

\]

Where \(w_i\) is segment revenue share and \(OM_i\) is segment operating margin.

Base case:

\[

Overall\ OM_{segment} =

(0.746 \times 35.1\%) +

(0.229 \times 30.0\%) +

(0.0242 \times 10.7\%)

\]

\[

= 26.2\% + 6.87\% + 0.26\% = 33.35\%

\]

After subtracting corporate/unallocated drag of 4.35 points:

\[

29.0\%\ total\ GAAP\ operating\ margin

\]

Sensitivity — shift 100 basis points of revenue from Semiconductor Systems to AGS:

\[

\Delta OM = 0.01 \times (30.0\% - 35.1\%) = -0.051\ points = -5.1\ bps

\]

Sensitivity — shift 500 basis points from Semiconductor Systems to AGS:

\[

\Delta OM = 0.05 \times (30.0\% - 35.1\%) = -0.255\ points = -25.5\ bps

\]

Sensitivity — shift 100 basis points from Semiconductor Systems to Display:

\[

\Delta OM = 0.01 \times (10.7\% - 35.1\%) = -0.244\ points = -24.4\ bps

\]

From [AMAT 10-K FY2024] business/risk sections — with exact names requiring primary source validation — AMAT’s customer base falls into these classes:

• Foundry/logic: Taiwan Semiconductor Manufacturing Company (TSMC), Samsung Foundry, Intel, SMIC, GlobalFoundries, and others.
• Memory: Samsung, SK hynix, Micron, Kioxia/Western Digital JV, and others.
• Integrated device manufacturers: companies that design and manufacture their own chips, including Intel and Samsung.
• Display manufacturers: OLED and advanced display producers, principally in Asia.
• Advanced packaging / specialty: outsourced semiconductor assembly and test (OSAT) and specialty customers, increasingly relevant for heterogeneous integration.

AMAT does not disclose the full customer list. The 10-K normally discloses customer-concentration data and names only customers exceeding 10% of revenue. The exact percentage and name for FY2024 must be validated directly in the filing. The company’s risk factor discussion commonly names Samsung, TSMC, Intel, SK hynix, and Micron as major customers, but this note does not assert a specific FY2024 concentration percentage until validated.

The buyer is not a single executive. It is a multi-functional procurement group consisting of:

• Fab process engineering and integration teams.
• Fab operations and capacity planners.
• Capital equipment procurement.
• Corporate technology/roadmap teams.

The sale is high-touch, with joint process development and field support. This supports pricing power but also creates long sales cycles.

AMAT is heavily exposed to Asia. China is a significant revenue region, but exact percentage and export-license impact need primary source validation from [AMAT 10-K FY2024]. Export-control rule changes can materially affect revenue and mix.

Applied Materials is the largest semiconductor wafer-fabrication equipment supplier by revenue. Its economic model is equipment-plus-recurring-service: it sells high-value deposition, etch, ion-implant, planarization, and process-control tools, then monetizes the installed base through spares, upgrades, remanufactured equipment, and fab-optimization services.

AMAT reports three operating segments in its 10-K:

1. Semiconductor Systems — most of revenue. This segment includes equipment sold to logic/foundry, DRAM, and NAND manufacturers.
2. Applied Global Services (AGS) — service, spares, upgrades, legacy equipment, and software.
3. Display and Adjacent Markets — deposition, PVD, and related equipment for display manufacturing.

Based on the company’s recent 10-K segment presentation, the revenue mix is approximately:

The consolidated revenue figure is the highest-confidence number: Applied Materials reported $27,176 million in net sales for fiscal 2024, up from $26,517 million in fiscal 2023. These are retrievable from SEC EDGAR companyfacts under the XBRL tag RevenueFromContractWithCustomerExcludingAssessedTax for CIK 0000006951. The exact segment-level dollar amounts should be re-pulled from the AMAT 10-K for fiscal 2024, filed 2024-12-13.

AMAT’s reported gross margin for fiscal 2024 was approximately 47.5% on a GAAP basis, with operating margin near 29–30% GAAP. The company has historically reported non-GAAP gross margin slightly above GAAP gross margin, and non-GAAP operating margin around 31–32%. Exact margin values should be re-pulled from the income statement or MD&A in the 10-K; for this note, the margin level is cited as needs primary source validation because the precise basis-point figure must be verified against the filed income statement.

The structural driver of margin is high value density per tool: a single deposition or etch platform can carry a multi-million-dollar price, and AMAT’s systems are sold into fabs where process qualification costs are extremely high.

AMAT monetizes through:

• System sales: multi-million-dollar semiconductor manufacturing tools.
• AGS revenue: spare parts, consumables, service contracts, software licenses, and remanufactured equipment. This is the closest AMAT has to recurring revenue; AGS is typically around one-fifth of total revenue.
• Display equipment: smaller, more cyclical, and increasingly dispositive as AMAT concentrates on advanced semiconductor tools.

The paying customers are semiconductor manufacturers: integrated device manufacturers, foundries, and memory makers. In AMAT’s 10-K customer-concentration disclosure, the largest customers historically include Samsung Electronics, Taiwan Semiconductor Manufacturing Company, Intel, Micron, SK hynix, and GlobalFoundries. Exact customer-concentration percentages for fiscal 2024 are disclosed in the 10-K and should be quoted directly from Item 1 or the concentration-of-risk footnote.

The buyer is rarely a single person. AMAT sells to a buying committee inside a semiconductor manufacturer:

• Fab process engineering: owns process specs, tool qualification, and yield.
• Fab equipment engineering: owns uptime, maintenance, and spare-parts budgets.
• Procurement / supplier management: negotiates price, volume, and service terms.
• Corporate technology development: selects next-node equipment architecture and suppliers.

The economic buyer is the fab owner or foundry executive who controls multi-billion-dollar capital expenditure budgets. For memory makers such as Samsung, SK hynix, and Micron, capex is highly cyclical and tied to memory pricing. For logic/foundry customers such as TSMC and Intel, capex is tied to node transitions and customer demand for advanced computing.

AMAT’s 10-K discloses customer concentration because a small number of large semiconductor manufacturers dominate its revenue. The names typically disclosed include Samsung Electronics, TSMC, Intel, and Micron. The exact FY2024 concentration percentages are in the filing; for this note, the concentration levels are needs primary source validation unless re-pulled from the 10-K. Historically, one customer has sometimes exceeded 10% of net sales, and the top three customers can collectively exceed 30–35% of annual revenue.

Lam Research’s 10-K shows a very similar customer base: large memory and foundry/logic customers dominate. This corroborates AMAT’s customer-concentration risk but also contradicts any narrative that AMAT’s customer mix is unique. Lam’s recent 10-K has disclosed that a small number of customers account for a substantial portion of revenue, and Lam has warned about China exposure and export-control risk. The contradiction relevant to an investor: AMAT’s 10-K may describe its served market and position differently than Lam’s 10-K describes the etch/deposition market. AMAT tends to frame itself as the “materials engineering” leader across multiple process steps, while Lam tends to frame its leadership more narrowly in etch and deposition. These are separate issuer-sourced claims and should be read as positioning rather than neutral market-share proof.

Applied Materials reports segments in its SEC filings. The latest publicly reported segments are Semiconductor Systems, Applied Global Services, and Display and Adjacent Markets. These names are taken from AMAT’s annual report structure, but the exact fiscal-year split is needs primary source validation until re-read directly from the latest 10-K.

Required revenue-mix table — values withheld pending primary-source verification:

I will not quote the reported percentages from memory because the prompt forbids unverified figures.

AMAT earns revenue through several linked mechanisms:

1. Equipment sales. A chipmaker buys a physical tool — such as a chemical vapor deposition chamber, an ion implanter, or a CMP system — for a one-time payment. These systems are configured to a fab’s process flow and require installation, process qualification, and production ramp. Payment is typically tied to delivery, installation, or acceptance milestones depending on contract terms. The exact revenue-recognition policy is set forth in the 10-K revenue-recognition footnote and is needs primary source validation.

1. Recurring/service revenue. After installation, the company sells spare parts, consumables, maintenance contracts, system upgrades, and process-support services. This is generally reflected in the Applied Global Services segment. It is usually less cyclical than system sales because fabs must maintain and service installed tools even during capital-spending downturns. That annuity-like stream alters the quality of the revenue base; exact mix, growth, and contractual terms are needs primary source validation.

1. Software and data analytics. Modern AMAT tools include sensors and process-control software for yield optimization, chamber matching, recipe management, and predictive maintenance. These may be purchased with the tool or sold as continuing license/subscription-like agreements. This deepens integration with customer fabs and may create recurring revenue. Exact revenue contribution is needs primary source validation.

The paying customers are semiconductor manufacturers — integrated device manufacturers, foundries, and memory producers — as well as display makers for the display segment. The largest purchasers in the industry are typically the largest semiconductor companies; AMAT’s 10-K customer-concentration footnote will disclose any customer representing 10% or more of revenue and may name those customers. I will not assert specific names here. The names and percentages are needs primary source validation.

GAAP and non-GAAP gross margin, operating margin, and net margin for the most recent fiscal year must be taken directly from the income statement. They are needs primary source validation. Qualitatively, semiconductor equipment companies with proprietary process tools, high R&D content, and a large installed base tend to report high gross margins relative to other hardware businesses, but I will not quote a margin number to avoid fabricating an unverified figure.

The key unit economics are per-tool and per-account rather than per-subscription. The logic is:

• A single advanced process tool sells for a substantial capital-goods price; exact ASP is needs primary source validation.
• The tool may generate service, spare-parts, and upgrade revenue for many years after initial sale; the exact service-to-equipment ratio is needs primary source validation.
• A customer’s economic cost of switching tools can be larger than the tool price. That is the basis of the worked calculation below.

The customer base is concentrated by industry structure. Only a small number of companies can afford leading-edge wafer fabs. The likely named customers in AMAT’s 10-K customer-concentration footnote are the largest semiconductor manufacturers. To comply with the no-fabrication rule, this section lists buyer classes and the method to verify them, not invented customer names.

1. Leading-edge logic foundries. These purchase the most advanced deposition and materials-engineering tools for fin field-effect transistor and gate-all-around/nanosheet flows, advanced interconnects, and low-resistance materials. They are high-value buyers because yield and performance requirements are extreme.

1. Memory producers. They buy tools for DRAM capacitor scaling, 3D NAND layer deposition, and high-aspect-ratio processing. Memory capital spending is highly volatile; when memory prices fall, these buyers slash orders.

1. Integrated device manufacturers. Companies that design and manufacture their own chips purchase production tools across logic, interconnect, and packaging.

1. Mature-node and specialty fab operators. They buy less leading-edge tools for power, analog, automotive, industrial, and RF chips. This market is served by AMAT and increasingly by domestic Chinese suppliers.

1. Display manufacturers. A smaller segment selling deposition and inspection equipment for OLED and LCD manufacturing; exact contribution is needs primary source validation.

• Technology leaders. They pay a premium for the best on-wafer performance because a small yield improvement can be worth hundreds of millions of dollars. They want co-development, local service, and data integration.
• Cost-driven volume fabs. They care about total cost of ownership, uptime, and throughput; they may accept slightly lower performance for a lower price.
• Geopolitically constrained buyers. Export-control rules — including U.S. BIS rules and coordinated Dutch/Japanese restrictions — can prevent AMAT from selling advanced tools to certain China-based fabs. This is a material demand and competitive factor; exact revenue exposure is in the 10-K and is needs primary source validation.

U.S. rules require disclosure of customers representing 10% or more of revenue. For AMAT, those disclosures historically show one or more large customers. The names and percentages are needs primary source validation. A live filing review should pull the exact customer-concentration footnote from the latest 10-K and compare it with the prior-year footnote to observe shifts.

• Q2 FY2026: $7.91B (record), +11% YoY
• Gross margin: 49.9% GAAP / 50.0% non-GAAP
• Operating margin: 31.9% GAAP / 32.1% non-GAAP
• Full-year guidance FY2026: ~$32.5B+ implied (Q3 guidance $8.95B, trend implies accelerating)
• Semiconductor Systems segment: $5.97B Q2 revenue, 54.7% gross margin, 35.1% operating margin

FY2025 revenue was $29.02B, up 3.3% YoY; gross margin was 49.0% and operating margin was 28.6%. Q2 FY2026 revenue was $7.91B (+11% YoY) with gross margin of 49.9% and operating margin of 31.9%; Q3 FY2026 guidance was $8.95B ± $0.5B revenue and $3.36 ± $0.20 non-GAAP EPS.

AMAT is a foundational picks-and-shovels vendor in the AI infrastructure build-out. Physical AI inference at the edge requires leading-edge chips (GPUs, custom ASICs, edge processors) which require state-of-the-art wafer fab tools. AI computing infrastructure (datacenter GPUs, HBM, interconnect ASICs) drives ~80% of AMAT's growth in FY2026. The company's primary customers (TSMC, Samsung, SK hynix, Intel, Micron) are all building next-generation process nodes specifically to enable AI and high-bandwidth memory products. AMAT's capacity constraints and lead times are real — supply chain readiness is explicitly noted as a management priority.

AMAT sits at the front end of the Physical AI supply chain: the more the world builds leading-edge logic, HBM, and advanced packaging capacity for AI compute, the more indispensable AMAT's toolset becomes. The moat is qualification-driven and capacity-constrained, not software-driven.

AMAT is a broad-based semiconductor equipment and services franchise with high operating margins, strong cash conversion, and meaningful aftermarket/service revenue. The main cyclical and geopolitical risk is China concentration. The investment case appears durable, but it is not immune to WFE downcycles.

• AI-driven revenue growth: CEO stated "rapid global build-out of AI computing infrastructure combined with Applied's strong leadership positions in leading-edge logic, DRAM and advanced packaging provide an exceptionally strong foundation for sustained, multi-year revenue and profit growth." (Q2 2026 earnings, 2026-05-14)
• Supply chain tightness: CFO noted "As the largest process equipment company in the fastest growing markets, our top priority is ensuring we have the operational and supply chain readiness to support our customers' growth. We have increased our build plan, inventory positions and logistics capacity." (Q2 2026 earnings, 2026-05-14)
• Expected FY2026 growth: Management guided that "our semiconductor equipment business [to] grow more than 30 percent in calendar 2026," vs flat 2025. This is driven entirely by AI chip demand acceleration.
• Segment performance: Semiconductor Systems (AMAT's core equipment business) grew 10% YoY revenue to $5.97B, with 54.7% gross margin and 35.1% operating margin. Foundry/logic (67% of seg revenue) is the lead growth driver due to leading-edge AI logic transitions.
• Technology partnerships: Announced major EPIC Center co-development partnerships with TSMC, Samsung, SK hynix, Micron, and three universities (ASU, RPI, Stanford) focused on "energy-efficient" AI chip technologies—explicitly positioning AMAT as a co-developer, not a mere vendor.
• Backlog/build acceleration: Explicit language about increased "build plan, inventory positions and logistics capacity" signals AMAT is capacity-constrained and allocating production to fulfill customer demand growth.
• Advanced packaging acquisition: Agreed to acquire NEXX (from ASMPT) to expand large-area advanced packaging deposition portfolio. This is directly enabling 3D chiplet stacks for AI accelerators — a critical Physical AI bottleneck.
• Customer design wins: Introduced new Precision Selective Nitride PECVD and Trillium ALD systems specifically for "3D Gate-All-Around transistors for the world's most advanced logic chips" and "complex metal gate stacks that optimize transistors for a wide range of AI computing applications."

• FY2025 revenue of $29.02B and 2025 margins near 49% gross / 28.6% operating show a very large, profitable equipment platform rather than a niche tool vendor.
• The Q2 FY2026 earnings release said semiconductor equipment business growth should exceed 30% in calendar 2026, explicitly tying demand to the "rapid global build-out of AI computing infrastructure" and to leading-edge logic, DRAM, and advanced packaging.
• Management also said it had increased build plan, inventory positions, and logistics capacity to support customer growth, which reads like supply-chain tightness rather than slack demand.

W1 — Switching-cost asymmetry (the qualification moat, dollarized).

Parameters: leading-edge logic fab = 100,000 wafer-starts/month (WSPM) [assumed, TSMC N3-class scale]; leading-edge wafer ASP = $8,000 [assumed, conservative — N3/N2 wafers price higher]. Monthly output value = 100,000 × $8,000 = $800M/month. Requalifying a switched deposition/etch module = 12–24 months [inferred, industry-standard fab qualification]. A 3% yield loss during 18 months of requalification = 0.03 × $800M × 18 = $432M at-risk wafer value — versus a ~$10M deposition tool [inferred, Endura/Producer class]. Switching-cost asymmetry ≈ 43× the tool price. No fab switches a qualified AMAT module for price; the risk/reward is structurally one-sided.

W2 — Cost-of-moat vs. moat-defended (R&D intensity).

FY2025 R&D = $3.57B on $28.37B revenue = 12.6% (Q2 FY2026 R&D $1.027B = 13.0% of revenue) [XBRL companyfacts, CIK 0000006951]. The moat it defends: FY2025 operating income $8.29B (29.2% OM), now running $3.08B/quarter at 33.7% GAAP OM. Cost-of-moat ≈ 12.6% of revenue defends a ~30–34% operating margin — a ~2.4–2.7× moat ROI, and the OM is expanding (13 consecutive quarters of YoY GM expansion, per CFO).

W3 — DRAM/HBM acceleration (the mix shift).

Q3 FY2026 DRAM revenue = Semiconductor Systems $7.04B × 26% = $1.830B, vs. Q3 FY2025 $5.564B × 22% = $1.224B → +$606M, +49.5% YoY. DRAM rose from 22% → 26% of the tools segment in one year — the HBM/3D-stacking ramp (AMAT's Epi/CMP/PECVD/ECD/metrology are the HBM-enabling toolset). This is the single clearest quantitative confirmation of the AI-memory bottleneck thesis.

W4 — Forward guide implies re-acceleration.

Q4 FY2026 revenue guide $10.25B ± $500M = +37% YoY implied at midpoint (vs. +25% in Q3, +11% in Q2). Calendar-2026 Semiconductor Systems growth raised to >30% (from ">30%" prior, now "further raising"), with management guiding "another strong growth year in 2027."

W1 — China mix compression vs US doubling (the export-control residual, now measured).

Q3 China $2,506M / 28% vs $2,548M / 35% YoY (−$42M, −2% dollars, −7 pp mix). Q3 US $1,367M / 15% vs $683M / 9% (+$684M, +100%). Taiwan $2,025M / 22% (+10% YoY); Korea $1,521M / 17% (+31% YoY). Asia Pacific is still 80% of Q3 ($7,265M). The bottleneck is not “China goes to zero”: 28% of a record quarter is still $2.5B. The mix shift is the thesis-relevant number — incremental dollars are coming from US/Korea/Taiwan leading-edge, not China trailing-edge. Source: 10-Q geographic table. [measured]

W2 — Top-2 customer dollarization (unnamed, 9M).

Two customers = ~20% and ~14% of 9M revenue = ~$4.81B + $3.37B = $8.17B of $24.04B (34%). No other customer >10%. This is buyer concentration risk, not AMAT pricing power: the 10-K/10-Q concentration table is the opposite of toll-road Q2 (which asks what share of the buyer’s input AMAT is). Use it as a shock-absorber test, not a moat proof. A 10% cut at the 20% account is ~$481M of 9M revenue (~$160M/quarter) — ~1.8% of Q3. Painful, not thesis-breaking, unless that account is also the DRAM/HBM node owner. [measured dollars; inferred identity]

W3 — Contract-liability coverage (prepaid demand).

Contract liabilities $3,271M vs $2,566M at FYE 2025-10-26 (+$705M, +27.5%). Coverage vs Q3 revenue = 3271/9115 = 0.359× a quarter ≈ 1.08 months of revenue prepaid. 9M recognized ~$1.8B of the FYE liability, so the $705M net increase is new unsatisfied performance obligations, not a stale deferred pile. This is the 10-Q’s demand-visibility number; it is not backlog (AMAT does not disclose remaining performance obligations as a single RPO line). [measured]

W4 — Receivable balloon vs revenue (the disconfirming working-capital check).

AR $7,691M vs $5,185M at FYE (+$2,506M, +48.3%) while 9M revenue is only +11.4%. Q3 DSO ≈ 76.8 days vs 71.9 days on the Jul-2025 comparable AR of $5,772M (+4.9 days). Contract assets actually fell ($227M vs $281M), so the balloon is billed receivables, not unbilled technical-signoff. Read: shipments are running ahead of cash conversion. Not a thesis break at 77 DSO, but the threshold to watch is DSO > 85 or AR growth remaining ≥3× revenue growth next quarter. [measured / derived]

W5 — DRAM mix (10-Q confirmation of EX-99.1).

SS DRAM 26% of $7,040M = $1,830M vs 22% of $5,564M = $1,224M → +$606M, +49.5% YoY. NAND 7% vs 9% is flat-to-down in dollars ($493M vs $501M). Foundry-logic dollars still grew +22.9% even as mix slipped 69%→67%. MD&A: foundry/logic spending rose on “stronger demand for leading-edge manufacturing technologies,” 9M partially offset by “lower customer demand for trailing-edge logic systems”; memory rose on “DRAM technology transitions.” That is the HBM/GAA intensity claim in the company’s own words. [measured mix; MD&A qualitative]

W6 — Cost-of-moat vs moat-defended (9M, 10-Q).

9M R&D $3,055M / $24,037M = 12.7% of revenue. 9M operating income $7,429M = 30.9% OM (includes the $253M BIS charge). Cost-of-moat ÷ operating margin ≈ 12.7 / 30.9 = 0.41×, or ~2.4× operating-profit dollars per R&D dollar. Q3 alone is cleaner: R&D $1,100M (12.1%) vs OM 33.7% with no BIS charge in the quarter. SS Q3 GM 55.3% vs 53.4% (+190 bps); MD&A attributes GM expansion to higher revenue, higher ASPs, lower material/manufacturing cost, and (Q3) favorable mix — the pricing-power tell. [measured]

W7 — Packaging bolt-on is not the P&L.

9M cash paid for acquisitions, net of cash acquired = $262M (vs $29M prior year). Against SS Q3 annualized run-rate $7,040M × 4 = $28.16B, $262M = 0.93% of SS run-rate. Even taking the reported NEXX consideration at $120M, that is 0.43% of SS run-rate. Packaging adjacency is real; it is not how AMAT captures the bottleneck this year. [measured cash; NEXX $120M is third-party ASMPT disclosure, not restated in this 10-Q]

All primary figures below are from [AMAT 10-K FY2024] unless noted. All should be re-validated on live EDGAR before use.

$19.9B ÷ $95B = 20.9%

$6,100,000,000 ÷ 45,000 = $135,600 per system per year