KLA Corporation
Capital-equipment + recurring service/software: sells inspection & metrology systems (~75-80% of rev) to chipmakers, then earns a high-margin, multi-year service annuity (install base growth 13-15% CAGR target) plus reticle/software.
The thesis on this name
State of AI Compute
The resolution to 'monopoly-position-vs-mid-pack-rate' — KLA has BOTH: near-chokepoint process-control position (yield-ramp single-sourcing) WITH the best rate in WFE (62.2% GM, asset-light) at a discount to ASML; verify confirmed it as the best risk-adjusted name in the semicap cluster and the only one with quality AND a non-heroic multiple, so it leads th…
State of AI Compute
Long KLAC over 12-24 months: own the ~60%-share, single-sourced process-control monopoly with sector-best 62% GM and 30%+ FCF margin as AI-driven WFE and advanced-packaging inspection intensity compounds — but enter on a starter and scale on pullbacks, because the structural-quality edge is consensus and the current ~…
State of AI Compute
~60%-share process-control monopoly, but a verified ~50x peak multiple near a WFE peak; haircut to a 2% starter with a MANDATORY concurrent cyclical tail hedge (put spread / paired short vs a memory/China-levered semica…
State of the Memory Supercycle
The process-control near-monopoly (~55%+ of inspection/metrology) and the purest 'complexity tax' on memory. As DRAM moves to 1c/High-NA EUV and HBM stacks more TSV/hybrid-bonded die, yield-critical inspection steps multiply — KLA is guided to gain ~100bps to ~13% of WFE in 2026 (BofA). Higher process complexity structurally raises KLA's dollar content per wafer, so it benefits even if total wafer-starts plateau. Best-in-class WFE margins and a moat (defect-detection physics + decades of recipe data) that is the hardest to disrupt of any layer in this stack. The toll-booth that gets paid more precisely when the cycle's leading-edge transitions are hardest.
State of the Memory Supercycle
The process-control monopoly that gets MORE indispensable as nodes shrink and HBM yields matter — KLA gains ~100bps of WFE share to ~13% in 2026; the highest-margin, lowest-cyclicality toll-booth.
State of the Memory Supercycle
Process-control near-monopoly, highest margins, gains ~100bps WFE share in 2026; gives back least in a wobble.
Earnings, margins, COGS & capex
KLA is a high-margin, high-FCF franchise riding AI-driven wafer-fab-equipment (WFE) intensity: FY25 revenue grew 24% to $12.16B and quarterly revenue has climbed every quarter through Q3 FY26 ($3.42B, +11.5% YoY). Non-GAAP gross margin sits at sector-best ~62% and operating margin ~42-43%, converting to ~31% TTM FCF margin. Growth is led by advanced-packaging inspection (~$1B in 2026, +50%+) and a 13-15%-CAGR service annuity, partly offset by a ~100bp DRAM-cost gross-margin headwind and a structurally shrinking China mix.
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~39¢ is cost of goods and ~19¢ operating expense, leaving ~43¢ of operating profit (~35¢ net).
Revenue trend
Margins
flat-to-up; ~62% sustained, +45bp vs guide
stable/strong
up
strong; dips to ~18% in working-capital-heavy quarters
COGS structure
COGS is dominated by purchased materials/components for complex optical, e-beam and X-ray inspection systems (lenses, sensors, photonics, precision stages), plus skilled manufacturing labor and a growing field-service organization. Mix is the key swing factor: higher service and software content lifts margin, while elevated DRAM/component costs are a current ~100bp gross-margin headwind through CY2026; tariffs could add up to ~$350M cost / ~100bp headwind risk.
Capex
Low and asset-light: ~$340M capex in FY25 (~2.8% of revenue), funding R&D labs, cleanroom/test capacity and IT — KLA outsources much fabrication, so incremental WFE demand drops to FCF rather than capex [fact].
Latest earnings
Beat, and beat cleanly on quality. Versus the guidance issued with the March quarter: revenue $3,657.6M against $3.575B ±$200M (above the midpoint); non-GAAP GM 62.4% against 61.75% ±1.00%; non-GAAP EPS $1.05 against a $0.987 midpoint (the $9.87 ±$1.00 guide, pre-split). Versus the street, non-GAAP EPS $1.05 beat a ~$1.00 Zacks consensus by ~5% and revenue beat by ~1.3–1.7% (source summaries differ slightly on the revenue consensus, ~$3.60–3.61B). Unlike the prior quarter, nothing about the beat is below-the-line: the $482.8M of incremental revenue carried through at a higher gross margin than guided.
Raised in substance, as of 7/28/2026. Q1 FY27 (September quarter): revenue $4.0B ±$200M (+9.4% sequential off a record quarter); GAAP gross margin 61.6% ±1.0%; non-GAAP gross margin 62.5% ±1.0%; GAAP diluted EPS $1.14 ±$0.10; non-GAAP diluted EPS $1.16 ±$0.10. Market frame raised on the call: CY2026 WFE to ~$150B (~mid-20% growth over ~$120B in CY2025), up from the prior '>$140B' frame; CY2027 industry consensus cited at ~$190B. Advanced packaging revenue guided to ~$1.1B in CY2026, up more than 70% YoY, versus ~$1B / +50%+ previously. Gross-margin framing was widened to a 60–65% long-run range, with the ~100bp memory-component cost headwind now expected to persist through 2027. KLA does not guide operating margin or full-year revenue.
- Q4 revenue
- $3,657.6M, +15.2% YoY — an acceleration from +11.5% in Q3 FY26
- FY26 revenue
- $13,579.5M, +11.7% vs FY25's $12,156.2M (record)
- Non-GAAP gross margin
- 62.4% in Q4, above the 61.75% guide midpoint; 62.5% ±1.0% guided for Q1 FY27
- Non-GAAP operating margin
- 43.7% in Q4, on $682.0M of non-GAAP operating expenses
- Free cash flow
- $817.1M in Q4 (22.3% margin, -23% YoY on a $586M receivables build); $3,767.1M FY26 (27.7% margin)
- Capital returned
- $876.3M in Q4 ($571.0M buyback + $305.3M dividends); $3,347.6M for FY26
- Service revenue
- $820.4M in Q4, +16.8% YoY, 22% of total revenue
- Semiconductor Process Control segment
- $3,256.8M in Q4 (89% of revenue), +13.2% YoY
- PCB and Component Inspection segment
- $241.1M in Q4, +56.5% YoY — the fastest-growing reportable segment (distinct from the ~$1.1B CY26 advanced-packaging figure, which spans products in several segments)
- Regional mix (Q4)
- Taiwan 31%, China 26%, North America 18%, Korea 10%, Japan 6%, Europe 5%, rest of Asia 4% (earnings presentation)
- Q1 FY27 revenue guide
- $4.0B ±$200M, +9.4% sequential
- Balance sheet
- $4,902.4M cash and investments vs $5,887.4M debt at Jun 30 2026 — ~$985M net debt
Growth drivers
- AI-driven WFE expansion — management raised CY2026 WFE outlook to >$140B and signaled 2027 > 2026
- Advanced-packaging inspection/metrology — ~$1B revenue in CY2026, nearly doubling from $635M (HBM, 2.5D/3D, chiplets)
- Process-control intensity rises at each node (gate-all-around, backside power, EUV/High-NA) — more inspection steps per wafer
- Service annuity off a growing, sticky installed base — 13-15% long-term CAGR target, +16% YoY in Q3 FY26
- Market-share gains in patterned-wafer/optical inspection (~7x nearest competitor; +360bp share since 2021)
- Leading-edge logic + HBM/DRAM capacity buildout by TSMC, Samsung, SK Hynix, Intel, Micron
Reported financials — SEC EDGAR
Audited GAAP figures pulled from SEC filings · latest filing 2026-08-06. The audited primary-source spine — not financial advice.
Revenue — annual (GAAP)
Margins & balance sheet — FY’26
Bull & bear
The June quarter turned the AI-WFE thesis from an outlook into a print: revenue accelerated to +15.2%, gross margin beat its own guide at 62.4% non-GAAP, and management raised both the market (CY26 WFE to ~$150B) and the fastest-growing line (advanced packaging to ~$1.1B, +70%). A ~62%-gross-margin, ~44%-non-GAAP-operating-margin, 2.8%-capex franchise pointed at $4.0B quarters is the cleanest toll on leading-edge silicon anyone lists — and it now trades at roughly a third less than it did five weeks ago on numbers that got better, not worse.
- Growth accelerated instead of rolling over: Q4 revenue $3,657.6M, +15.2% YoY against +11.5% the prior quarter, and the September guide of $4.0B ±$200M is another +9.4% sequential off a record — the shape of a cycle mid-ascent, not a peak (Q4 FY26 release)
- Margin held above its own target under a live cost headwind: 62.4% non-GAAP gross margin beat the 61.75% guide, 43.7% non-GAAP operating margin, and Q1 FY27 is guided to 62.5% ±1.0% — with the ~100bp memory-component drag already inside those numbers
- The demand frame was raised twice: CY2026 WFE to ~$150B (from '>$140B' in June, ~mid-20% growth over ~$120B in CY2025), and management cited an industry-consensus ~$190B for CY2027 — the top-down variable that drives everything else moved up
- Advanced packaging upgraded to ~$1.1B in CY2026, up more than 70% YoY (from ~$1B / +50%+ previously); the PCB and Component Inspection segment grew 56.5% YoY to $241.1M — the HBM/2.5D/3D attach is showing up in reported revenue, not just in slides
- The annuity keeps compounding: service revenue $820.4M, +16.8% YoY, now 22% of sales — recurring, high-margin, and tied to an installed base that only gets larger with each WFE dollar
- Capital structure and returns are a non-issue: ~$985M net debt against $4.90B of cash and investments, $3,347.6M returned in FY26 (~89% of free cash flow) with $876.3M in Q4 alone, all funded on 2.8% capex intensity
The de-rating already happened — the stock is ~29% below our June mark at ~$182.82 and the forward multiple is ~33.6x rather than ~54x — so 'valuation' is no longer the first objection. What is left is the harder one: this is a WFE-cyclical whose customer base is five fabs, whose second-largest region (China at 26% of the quarter) is a policy variable, and whose cost headwind management just extended by a year. The market took 10.8% out of the stock on a beat-and-raise, which is worth respecting.
- The memory-component cost headwind got LONGER, not smaller: management now expects the ~100bp gross-margin drag to persist through 2027, versus the prior 'through CY2026' framing — and the long-run margin language widened to a 60–65% range rather than a ~62% target
- China was 26% of June-quarter revenue, UP from 24.3% in the March quarter — on the visible datapoints the China mix is not currently shrinking, which leaves more revenue exposed to export-control and license-renewal decisions we do not control
- The market disagreed with the print: shares fell 10.8% on Jul 29 (the first session after the release) despite a beat on revenue, gross margin and EPS, and are ~29% below our June mark — when a clean beat gets sold, the sell-side model was already ahead of the guide
- Cash conversion weakened underneath the income statement: Q4 free cash flow fell 23% YoY to $817.1M on a $586M receivables build and a $423M other-assets build, and FY26 FCF of $3,767.1M was essentially flat against FY25's $3,746.6M while net income rose 19% — growth is currently being funded out of working capital
- Concentration and cyclicality are unchanged by a good quarter: 89% of revenue is one segment, a handful of fabs set the order book, and FY24's -7% revenue year is the reminder that WFE gives back; nothing in this print insures against an AI-capex digestion phase
- Even after the de-rating the multiple is not cheap for a cyclical: ~50x trailing GAAP / ~48.6x trailing non-GAAP earnings and ~33.6x forward at $182.82 — that still requires the CY27 ~$190B WFE consensus to substantially land
What it is worth
Forward P/E + FCF/DCF cross-check on an AI-WFE cyclical-growth franchise
~$170-200
AI-capex digestion / China loss compresses both estimates and the multiple toward the high-30s P/E
~$255-290
mid-50s forward P/E holds on continued AI-WFE growth and buybacks (near current price)
~$320-360
WFE compounds past $140B, advanced packaging sustains 50%+, margins hold ~62%, ~30x forward on rising EPS
Trades ~54x forward / ~73x trailing P/E and ~$338B cap after a ~2x six-month run — a premium, cycle-peak-sensitive multiple; estimate, not a target.
SWOT
Strengths
- ~55-60% share of process control (~7x nearest competitor — ~75-80% in patterned-wafer inspection) — effective monopoly on the highest-value inspection steps
- Sector-best margins — ~62% non-GAAP gross, ~42-43% operating, ~31% TTM FCF margin on an asset-light (~3% capex) model
- Sticky, growing service annuity (13-15% CAGR target) off a large installed base smooths WFE cyclicality
- Fortress balance sheet (near net-cash, A-rated) funding ~$875M/qtr of buybacks + 17 straight dividend hikes
Weaknesses
- Cyclical to WFE/memory capex — revenue fell 7% in the FY24 downcycle; not immune to a chip-capex air-pocket
- High China exposure (~33% of FY25 rev) that is structurally shrinking under export controls — lost revenue and license-renewal uncertainty
- ~100bp gross-margin headwind from elevated DRAM/component costs through CY2026; tariff risk up to ~$350M
- Concentrated customer base (TSMC, Samsung, Intel, SK Hynix, Micron) — a few capex decisions swing the quarter
Opportunities
- Advanced packaging (HBM, 2.5D/3D, chiplets) inspection — fastest-growing line, ~$1B CY26 and rising as AI scales
- Rising process-control intensity at gate-all-around, backside-power, High-NA EUV nodes — more steps per wafer
- WFE secular expansion to $140B+ (2026) and higher in 2027 as AI compute buildout broadens
- Software/AI-yield analytics and reticle/mask inspection attach to deepen the per-fab take
Threats
- Tightening U.S. export controls and annual China license renewals — Washington holds a renewal card each cycle
- AI-capex deceleration or a memory glut triggering a WFE downcycle
- Tariffs / trade policy raising COGS or restricting end-markets
- Well-capitalized competitors (AMAT, Onto, Nova, Camtek) plus China indigenization (SMEE, AMEC-adjacent) chipping at lower-end inspection/metrology
Moats, dependencies & bottlenecks
Moats
~7x nearest competitor) Decades of optical/e-beam IP, reference recipes co-developed with fabs, and unmatched defect-sensitivity create a near-monopoly on the hardest inspection steps.
Tools are qualified into a fab's process flow; re-qualifying a competitor risks yield, so incumbents are rarely displaced node-to-node.
13-15% CAGR service revenue off a large, growing base — recurring, high-margin, and deepens customer dependence.
Largest process-control R&D budget and the broadest defect/yield dataset feed AI-analytics that competitors can't easily replicate.
~60% share in a fast-growing segment, but Onto/Camtek/Nova compete harder here than in patterned-wafer inspection.
Dependencies
Revenue tracks WFE; a few customers' capex decisions swing quarters. Concentration + cyclicality.
Current growth is AI-driven; a capex digestion or memory glut would compress revenue and the multiple.
~33% of FY25 rev from China, shrinking; annual license renewals give Washington a recurring lever KLA can't control.
Precision optics, sensors, photonics and DRAM costs drive COGS; ~100bp DRAM headwind through CY2026; tariff risk.
GAA, backside power, High-NA EUV and advanced packaging each add inspection steps — secular tailwind, low near-term risk.
Advantages
- Near-monopoly share (~58-60%, ~7x nearest rival) with pricing power on the highest-value inspection steps
- Sector-best profitability: ~62% gross, ~42-43% operating, ~31% TTM FCF margin
- Asset-light model (~3% capex intensity) converts demand to free cash flow
- Sticky, growing service annuity (13-15% CAGR) that damps WFE cyclicality
- Fortress balance sheet (near net-cash, A-rated) funding aggressive buybacks + a 17-year dividend-growth streak
- Structural exposure to the two fastest-growing semi themes — AI-logic node transitions and advanced packaging/HBM
Weaknesses
- WFE cyclicality — revenue fell 7% in FY24; not a secular-only story
- Heavy, policy-dependent China exposure (~33% FY25) being legislated down
- Premium valuation (~54x fwd P/E) leaves little margin for disappointment
- Customer concentration among a handful of leading-edge fabs
- Near-term margin headwinds from DRAM costs and potential tariffs
- Tougher competition in advanced packaging (Onto/Camtek/Nova) than in its core inspection franchise
Bottlenecks
- Customer fab capex timing — orders cluster around a few fabs' build decisions, so revenue is lumpy and out of KLA's control
- China license-renewal cadence — annual approvals gate a shrinking but still-material revenue stream
- DRAM / component cost inflation — ~100bp gross-margin drag through CY2026; tariff overhang up to ~$350M
- Skilled field-service and applications-engineering headcount to support a rapidly growing installed base
- Industry capacity constraints limit how fast WFE can scale (mgmt: can't jump $140B→$200B overnight)
Top signals & trends
Top signals
The single best top-down read on KLA's revenue trajectory; further raises confirm the AI-capex cycle.
Fastest-growing line and a higher-attach, higher-growth mix; watch whether it sustains >30% growth.
~33% FY25 and falling; new restrictions or denied renewals are the key downside catalyst.
Holding ~62% despite ~100bp DRAM drag is the margin-resilience test; tariffs could add ~100bp.
HBM is a tailwind, but a commodity-DRAM glut would dent WFE and KLA orders.
Multiple compression risk if AI-capex enthusiasm cools — the case for starter-and-scale-on-pullbacks.
Trends
Directly expands KLA's addressable market and order book; mgmt sees 2027 higher still.
More inspection/metrology steps per package; ~$1B CY26 line growing 50%+.
Each adds process-control intensity (more steps/wafer), structurally raising KLA's per-wafer take.
Shrinks a ~33% revenue market and adds annual policy risk; partly offset by ex-China demand.
Long-run competition at the low end, but KLA's leading-edge inspection lead is hard to replicate near-term.
~100bp gross-margin headwind through CY2026; tariffs a potential ~$350M / ~100bp drag.
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Precision optics/lenses for inspection systems (private, German) — critical optical supply.
Lasers and photonics components used in inspection/metrology tools.
Illumination, sensors and photonic subsystems.
Elevated DRAM costs are a current ~100bp gross-margin headwind (component, not end-market).
TSMC (foundry for chips KLA designs into tools) Semiconductors/ASICs inside KLA's own systems.
Largest leading-edge foundry customer — AI-logic node ramps drive process-control demand.
Logic + memory (DRAM/HBM/NAND) capex; major KLA customer (Korea, non-US).
HBM/DRAM leader — advanced-packaging + memory inspection demand (Korea, non-US).
Foundry + leading-edge logic buildout (18A/14A) needs process control.
U.S. DRAM/HBM capacity expansion drives inspection/metrology orders.
Specialty/mature-node foundry customer.
#2 in process control with PDC group, but steadily losing inspection/metrology share to KLA; far broader deposition/etch portfolio.
Most direct pure-play rival in optical metrology, OCD, and advanced-packaging inspection/litho process control.
Israeli metrology specialist (dimensional/materials metrology) competing at advanced nodes; smaller but growing.
Advanced-packaging and compound-semi inspection/metrology; a focused challenger in the fast-growing packaging niche.
Etch/deposition leader, not a direct process-control competitor; a WFE peer competing for fab capex wallet share.
Hitachi (e-beam inspection/CD-SEM) and emerging China-domestic inspection vendors compete at the lower/maturing-node end; non-US names, analysis only.