KLA Corporation
Combined average of Moat (AI Resilience), Growth, and Valuation scores.
Moat Score
Near-monopoly (~58% process-control share in 2025, up 360bps since 2021) in semiconductor process control and wafer inspection — every advanced node ramp at TSMC, Samsung, and Intel runs through KLA tools. KLA holds the #1 position in process control for advanced wafer-level packaging, where FY2026 results put calendar-2026 revenue on track for ~$1.1B, growing at roughly twice the underlying market.
KLA's moat is the accumulated yield-learning data from decades of inline wafer inspection — a proprietary defect-signature library that only deepens at each new process node:
- Yield-Learning Data Compounds Per Node: Every wafer inspected adds to KLA's defect-signature library. Customers can't replicate this dataset because it's accumulated across the entire industry. New nodes (3nm, 2nm, 1.4nm) extend the lead — defect modes get more subtle and require more inspection sophistication, not less. Process-control share has climbed to ~58%.
- Inline at Every Wafer: KLA process-control tools sit inline in fab production lines — every wafer at TSMC's leading-edge fabs passes through KLA inspection. The transaction-embedding moat is structural: pulling KLA tools requires re-architecting fab flow.
- Advanced Packaging is the New Front: HBM stacks and chiplet integration require an order of magnitude more process control than monolithic chips. Advanced-packaging process control has gone from $635M in calendar 2025 to a guided ~$1.1B in calendar 2026 — past the company's own $1B target, growing at about twice the market, in a TAM that effectively didn't exist 5 years ago.
Moat Verdict
KLA is a clear net beneficiary of AI, and nothing in the FY2026 close moved a moat status. The strongest moats — proprietaryData (yield-learning library, ~58% process-control share), systemOfRecord (yield-management software), regulatoryLockIn (foundry certs + export controls), bundling, businessLogic — are all AI-strengthened or AI-neutral. The one AI-vulnerable moat (learnedInterfaces) is correctly downgraded because operator workflows are being automated. FY2026 evidence reinforces rather than changes the picture: a record $3.66B June quarter, ~$12.5B backlog, demand broadening past the single leading-edge customer, and advanced-packaging process control compounding at roughly twice its market. The 2030 framework (~$26B revenue, 45-47% op margin) underscores structural durability — KLA sits in the ASML/MSCI tier, second only to ASML in semicap structural durability.
80.7 resilient · 68.0 vulnerable · 80/20 = 78.2 · = 78
Open a moat to read its note.
Fab process engineers train on KLA tools, but operator workflows are increasingly automated and AI-assisted. Interface mastery is no longer the primary lock-in.
Yield-management algorithms and defect-signature classifiers encode decades of process learning, but they are built on the defect libraries and yield data already rated strong under proprietaryData; scoring the same asset as vendor logic too double-counts it. Re-rated from strong to intact.
process-control business does not derive moat from public datasets.
Applications engineers embedded with TSMC, Samsung, and Intel are scarce, but the talent moat is bounded — KLA, Applied Materials, and Lam compete for the same engineering pool.
Comprehensive process-control suite — defect inspection (29xx series) + metrology (LS9xxx) + reticle inspection + advanced packaging (Kronos) + yield-management software. Competitors lack the cross-tool data integration.
Decades of accumulated defect-signature libraries, process recipes, and yield-learning data per customer per node. This is the core moat — it compounds with every wafer inspected (process-control share reached ~58% in 2025) and cannot be replicated by competitors.
No regulatory moat of KLA's own. Foundry certification at each node is customer qualification — the switching cost rated under transactionEmbedding — and US BIS export controls restrict KLA's sales into China (management noted on the Q4 FY2026 call that KLA is barred from fab segments where domestic competitors operate): a headwind, not a barrier to rivals. Previously strong.
Yield-learning data flows back from many customer fabs, helping KLA improve algorithms which all customers benefit from. Indirect network effect bounded by the small number of leading-edge customers globally; broadening in FY2026 as demand extended beyond a single leading-edge customer into DRAM, NAND and advanced packaging, where KLA holds the #1 process-control position.
KLA tools sit inline in leading-edge production — every wafer at those fabs passes through inspection — and each tool is qualified into the process at every new node, so replacing it means requalifying the line. The lock shows in share: roughly 58% of process control, more than its next several rivals combined. Previously intact on the ground that inspection is not a payment layer; this framework rates workflow embedding in the production path the same way it rates ASML's scanners.
KLA yield-management software tracks every defect, wafer and lot, but that record is the same yield-learning dataset rated strong under proprietaryData, and fabs keep their own MES as the production record. Re-rated from strong to intact so one asset is not scored three times.
Fabless or equipment franchise whose edge is IP and process know-how, rated elsewhere; manufacturing volume does not give it a unit-cost lead rivals cannot buy.
Buyers are enterprises choosing on switching cost, integration and performance, which the other pillars rate. The name carries reputation, not a price premium it could hold on brand alone.
Combined average of Moat (AI Resilience), Growth, and Valuation scores.
Moat Score
Near-monopoly (~58% process-control share in 2025, up 360bps since 2021) in semiconductor process control and wafer inspection — every advanced node ramp at TSMC, Samsung, and Intel runs through KLA tools. KLA holds the #1 position in process control for advanced wafer-level packaging, where FY2026 results put calendar-2026 revenue on track for ~$1.1B, growing at roughly twice the underlying market.
Growth Score
Q4 FY2026 (reported July 28, 2026) revenue $3.66B — a record, +15% YoY and +7% sequentially — with non-GAAP gross margin 62.4%, non-GAAP operating margin 43.7%, and non-GAAP EPS $1.05 against a $1.00 consensus. FY2026 closed at $13.58B revenue (+11.7% YoY), $3.66 GAAP EPS and $3.77B free cash flow. The September quarter is guided to $4.0B ±$200M (+25% YoY) with non-GAAP EPS $1.16 ±$0.10 — a step up, not a continuation. Backlog exits FY2026 at ~$12.5B, management raised its 2026 wafer-equipment market view to the low $150B range, and the industry view for 2027 is ~$190B. Advanced-packaging process control is guided to ~$1.1B in calendar 2026 from $635M in 2025. The offsetting facts: memory pricing is a >100bp gross-margin headwind management expects to persist through 2027, and the 2027 acceleration rests on a market forecast rather than on backlog.
Valuation Score
The June review caught KLA at the top. Shares peaked at ~$302 on June 30, fell through July with the rest of semicap, and dropped a further 6% into the Q4 print before recovering to ~$184 — a ~39% drawdown from the high, even though the quarter beat and the September guide raised the run-rate 25% YoY. The forward multiple has gone from ~53× to ~37×, which puts the price between the bear ($130) and base ($215) cases and restores a margin of safety the June file explicitly said was gone.
The Yield-Learning Moat
KLA's moat is the accumulated yield-learning data from decades of inline wafer inspection — a proprietary defect-signature library that only deepens at each new process node:
- Yield-Learning Data Compounds Per Node: Every wafer inspected adds to KLA's defect-signature library. Customers can't replicate this dataset because it's accumulated across the entire industry. New nodes (3nm, 2nm, 1.4nm) extend the lead — defect modes get more subtle and require more inspection sophistication, not less. Process-control share has climbed to ~58%.
- Inline at Every Wafer: KLA process-control tools sit inline in fab production lines — every wafer at TSMC's leading-edge fabs passes through KLA inspection. The transaction-embedding moat is structural: pulling KLA tools requires re-architecting fab flow.
- Advanced Packaging is the New Front: HBM stacks and chiplet integration require an order of magnitude more process control than monolithic chips. Advanced-packaging process control has gone from $635M in calendar 2025 to a guided ~$1.1B in calendar 2026 — past the company's own $1B target, growing at about twice the market, in a TAM that effectively didn't exist 5 years ago.
Moat Verdict
KLA is a clear net beneficiary of AI, and nothing in the FY2026 close moved a moat status. The strongest moats — proprietaryData (yield-learning library, ~58% process-control share), systemOfRecord (yield-management software), regulatoryLockIn (foundry certs + export controls), bundling, businessLogic — are all AI-strengthened or AI-neutral. The one AI-vulnerable moat (learnedInterfaces) is correctly downgraded because operator workflows are being automated. FY2026 evidence reinforces rather than changes the picture: a record $3.66B June quarter, ~$12.5B backlog, demand broadening past the single leading-edge customer, and advanced-packaging process control compounding at roughly twice its market. The 2030 framework (~$26B revenue, 45-47% op margin) underscores structural durability — KLA sits in the ASML/MSCI tier, second only to ASML in semicap structural durability.
80.7 resilient · 68.0 vulnerable · 80/20 = 78.2 · = 78
Open a moat to read its note.
Fab process engineers train on KLA tools, but operator workflows are increasingly automated and AI-assisted. Interface mastery is no longer the primary lock-in.
Yield-management algorithms and defect-signature classifiers encode decades of process learning, but they are built on the defect libraries and yield data already rated strong under proprietaryData; scoring the same asset as vendor logic too double-counts it. Re-rated from strong to intact.
process-control business does not derive moat from public datasets.
Applications engineers embedded with TSMC, Samsung, and Intel are scarce, but the talent moat is bounded — KLA, Applied Materials, and Lam compete for the same engineering pool.
Comprehensive process-control suite — defect inspection (29xx series) + metrology (LS9xxx) + reticle inspection + advanced packaging (Kronos) + yield-management software. Competitors lack the cross-tool data integration.
Decades of accumulated defect-signature libraries, process recipes, and yield-learning data per customer per node. This is the core moat — it compounds with every wafer inspected (process-control share reached ~58% in 2025) and cannot be replicated by competitors.
No regulatory moat of KLA's own. Foundry certification at each node is customer qualification — the switching cost rated under transactionEmbedding — and US BIS export controls restrict KLA's sales into China (management noted on the Q4 FY2026 call that KLA is barred from fab segments where domestic competitors operate): a headwind, not a barrier to rivals. Previously strong.
Yield-learning data flows back from many customer fabs, helping KLA improve algorithms which all customers benefit from. Indirect network effect bounded by the small number of leading-edge customers globally; broadening in FY2026 as demand extended beyond a single leading-edge customer into DRAM, NAND and advanced packaging, where KLA holds the #1 process-control position.
KLA tools sit inline in leading-edge production — every wafer at those fabs passes through inspection — and each tool is qualified into the process at every new node, so replacing it means requalifying the line. The lock shows in share: roughly 58% of process control, more than its next several rivals combined. Previously intact on the ground that inspection is not a payment layer; this framework rates workflow embedding in the production path the same way it rates ASML's scanners.
KLA yield-management software tracks every defect, wafer and lot, but that record is the same yield-learning dataset rated strong under proprietaryData, and fabs keep their own MES as the production record. Re-rated from strong to intact so one asset is not scored three times.
Fabless or equipment franchise whose edge is IP and process know-how, rated elsewhere; manufacturing volume does not give it a unit-cost lead rivals cannot buy.
Buyers are enterprises choosing on switching cost, integration and performance, which the other pillars rate. The name carries reputation, not a price premium it could hold on brand alone.
Growth Analysis
Growth Drivers
Key Risk
The ~$190B CY2027 wafer-equipment market that underwrites the acceleration is a forecast, not orders — the ~$12.5B backlog covers roughly three quarters at the guided $4.0B run-rate. If hyperscaler capex digestion holds 2027 WFE near the 2026 low-$150B level instead of growing ~26%, the September-quarter guide is the cycle peak and revenue flattens through FY2027 with the memory-pricing headwind still in the gross margin.
Score Derivation
80.3 base + 4.0 trajectory − 5 risk = 79
Base 80 (14-17% blended CAGR — the +24.6% Q1 FY2027 guide decayed toward the 13-17% 2030 framework, not the +11.7% FY2026 actual and not the guide itself) + 4 trajectory (3 of 3 drivers accelerating) + 0 margin (record 62.4% gross margin and 43.7% operating margin, but a >100bp memory-pricing headwind through 2027 offsets the operating leverage) − 5 risk (the ~$190B CY2027 wafer-equipment market is an expectation, not backlog) = 79
Price Scenarios (12–24 Months)
Valuation Multiples
| Trailing P/E (GAAP) | ~50× |
| Forward P/E (NTM) | ~37× |
| PEG Ratio | ~1.8× |
| Price / Sales (NTM) | ~14× |
| Price / FCF | ~64× |
| Dividend | $2.30/sh |
The de-rating did the work the earnings did not. Forward P/E fell from ~53× to ~37× while FY2027 estimates rose, so KLA is no longer the outlier it was in June — ASML trades near ~48×, and Lam and Applied sit in the low-to-high 30s on recent data. PEG of ~1.8× is a normal price for a business with this moat. What is still rich is the cash-flow multiple: ~64× FY2026 free cash flow assumes the AI capex cycle keeps running, and FCF conversion has to catch up with reported earnings for the base case to hold.
Approximate figures as of July 2026.
Where We Are vs Targets
Loading live price…
CY2027 wafer-equipment spend flattens near the 2026 low-$150B level instead of the ~$190B consensus, memory pricing stays a >100bp gross-margin headwind, and the multiple reverts toward KLA's pre-AI band — ~26× FY2027 EPS.
- CY2027 WFE lands flat against 2026 rather than growing ~26%; the $4.0B September guide proves the cycle peak
- The ~$12.5B backlog (roughly three quarters of cover) converts without replenishment and FY2028 revenue is flat to down
- Memory pricing pressure persists through FY2027, holding gross margin near 62% against the 60-65% incremental model
CY2027 WFE reaches roughly $190B, KLA compounds at 14-17%, advanced packaging scales past the ~$1.1B calendar-2026 level, and the multiple holds ~43× FY2027 / ~35× FY2028 EPS.
- FY2027 revenue ~$16.9B (+24%) off the guided $4.0B September run-rate; non-GAAP EPS ~$4.98
- Advanced packaging keeps compounding above market as HBM4 and chiplet stacking raise process-control intensity per wafer
- Operating margin grinds from 43.7% toward the 45-47% 2030 target; FY2028 EPS ~$6.2
AI capex runs longer than the usual two-year semicap cycle, 2027 WFE overshoots $190B, memory pricing recovers, and the multiple re-rates back toward ~48× FY2028 EPS.
- CY2027 WFE exceeds $190B and 2028 grows again on hyperscaler ASIC programmes and sovereign fab buildouts
- Advanced packaging passes $2B as HBM4/HBM5 stacking and panel-level packaging expand the process-control TAM
- Memory pricing normalises and gross margin clears 64% on the 60-65% incremental model; FY2028 EPS ~$6.5-7.0