Layer 2 · KLAC — KLA Corporation
One-line thesis
KLA is the semiconductor process-control monopoly — 85% global share in optical-pattern wafer inspection — and the combination of HBM advanced-packaging inspection pull, the 10-for-1 stock split, and a 21% dividend raise alongside a clean Q3 FY26 beat-and-raise makes this the highest-quality "boring monopoly" trade in the cycle.
What KLA physically does
Every leading-edge fab — TSMC N3 / N2, Samsung 3nm, Intel 18A, SK Hynix HBM lines, Micron Boise HBM — runs hundreds of KLA inspection tools at every step of the wafer manufacturing process. The flagship products are optical-pattern wafer inspectors (the 29xx, 39xx, and 89xx series for unpatterned and patterned defect inspection), e-beam inspection and review tools (eDR-7xxx series), reticle-inspection systems for EUV mask qualification, and the suite of overlay and CD-SEM metrology tools that measure whether the silicon is being printed within sub-nanometer tolerances. In rough numbers: a leading-edge logic fab will install $1-1.5B of KLA equipment per 100k-wafer-month capacity layer, and a leading-edge HBM fab installs roughly $300-500M per HBM ramp step.
The economic logic of inspection is asymmetric. A leading-edge wafer at 3nm carries roughly 100-200 die per wafer, with each die selling for hundreds to thousands of dollars at NVIDIA / Apple / AMD ASPs. A single yield-killing defect — a particle, a pattern collapse, an overlay mis-print — that escapes inspection in early process steps destroys the wafer downstream after another $5-10k of process value has been added. KLA tools sit upstream of every process step that costs more than $1k per wafer to perform, with the explicit purpose of catching defects before that value is added. The throughput requirement is high (a single tool scans hundreds of wafers per hour) and the false-positive rate matters (every false flag stops a wafer for re-inspection, costing throughput).
KLA's near-monopoly is on the optical-pattern wafer inspection segment specifically, where TechInsights and Gartner estimate KLA holds roughly 85% global share — competitors like Hitachi High-Tech, Applied Materials' VeritySEM line, and Onto Innovation operate in adjacent metrology niches but do not directly contest the high-volume in-line inspection franchise. The structural reason for the share is the combined cost of the optical illumination system (deep-UV at 193nm wavelength, multi-pass differential interference contrast imaging), the defect-recognition software trained on 25 years of fab-floor data, and the customer-co-development relationships that lock in the next-generation inspection recipes a node before that node ships.
In advanced packaging specifically — the HBM stacking process, CoWoS through-silicon-via inspection, hybrid-bonding alignment metrology — KLA has been adding direct exposure faster than any other equipment vendor. The acquisition of SPTS in 2018 plus the organic build-out of the ICOS bump-inspection line plus the metrology tools that ship into TSMC CoWoS-L lines collectively add another $1.5-2B of advanced-packaging-specific TAM through 2027. The reticle-inspection franchise (the Teron family for EUV mask qualification) is a related growth line — every high-NA EUV scanner from ASML requires a stack of reticle-inspection tools at the fab to verify mask integrity at the sub-nanometer pattern level, and KLA is the dominant supplier with effectively no merchant competitor at the leading edge.
The financial print
Q3 FY26 (reported April 30 2026, quarter ended March) printed $3.42B of revenue against $3.32B consensus — beat by $100M (+11% YoY) — and non-GAAP EPS of $9.40 also beat. The company raised the FY26 guidance bracket to "high-teens" revenue growth on the strength of AI, advanced packaging, and memory pulls. The Board also announced a 10-for-1 stock split (effective late Q2 FY26) and a 21% dividend hike — twin signals that management views the current revenue trajectory as durable. The stock at $1,892 sits roughly 2% below the May 12 all-time high of $1,939; 1m +12%, 3m +28%, +155% off the 52-week low.
The 10-for-1 split is more than cosmetic. KLA's nominal share price near $1,900 had become a meaningful friction for retail flow, options activity, and inclusion in equal-weight indices. Post-split, the $190-equivalent stock price unlocks retail participation that historically corresponded to a one-time multiple expansion of 100-200 basis points across the comparable semiconductor cohort (LRCX, AVGO did the same in 2024). The dividend raise is the conservative complementary signal — management is committing to a payout ratio increase rather than aggressive buybacks, telegraphing capital-return discipline rather than financial engineering.
Q4 FY26 prints July 30. Consensus sits at approximately $3.55B of revenue and $9.80-10.00 of non-GAAP EPS. The setup is constructive — KLAC has rerated to the high-quality compounder multiple tier (35-37x forward) and any beat-raise extends that.
Customer mix
TSMC is the largest single customer at approximately 25-30% of revenue, with Samsung Foundry, SK Hynix, Micron, Intel, and the Chinese foundries (SMIC, YMTC pre-restrictions) collectively contributing another 50%. The remaining 20-25% is the broader memory cohort (Kioxia, Western Digital, Samsung Memory) plus the analog / specialty foundries (Tower, Vanguard, GlobalFoundries). The China exposure is real but declining — KLAC has been the most affected by export controls of any non-EUV equipment vendor — but the offset is the advanced-packaging-specific pull at TSMC and the HBM expansion at SK Hynix / Micron, both of which are running ahead of plan.
The geography split is the more useful disaggregation. Roughly 38-42% of revenue is Taiwan (principally TSMC), 18-22% is Korea (Samsung Foundry, Samsung Memory, SK Hynix), 12-15% is China (declining from over 25% pre-restrictions), 12-15% is North America (Intel, Micron, GlobalFoundries), and the balance is Japan, Europe, and Southeast Asia. The Taiwan and Korea weighting puts KLAC squarely in the centre of the AI silicon manufacturing geography — every leading-edge AI silicon die ultimately passes through KLA inspection at TSMC, and every HBM stack passes through KLA inspection at SK Hynix or Samsung Memory or Micron Boise.
Competitive context
There is essentially no head-on competitor at the optical-pattern wafer inspection tier. Applied Materials' inspection portfolio (the VeritySEM / e-beam lines) competes in adjacent segments but does not directly contest the high-volume optical inspection tier. Hitachi High-Tech is meaningfully present in Japan-based customers (Kioxia, Renesas) but has not won materially at TSMC, Samsung, or SK Hynix. Onto Innovation, which is also in this Tier A list, is in the bump-inspection / advanced-packaging adjacent niche and is more an ecosystem partner than a competitor at the leading-edge wafer-inspection tier.
The structural defence is the volume-of-defect-recognition-data moat. KLA tools generate hundreds of millions of defect images per quarter; the recipes shipped to new fabs are pre-trained on this aggregated data and ship at production-ready maturity from day one. A competitor with even a technically superior point-tool cannot replicate that data moat in less than five years.
Terminal risk
The terminal risk is twofold. First, China export restriction tightening — KLAC has historically derived roughly 25% of revenue from China and the marginal restrictions on advanced-process equipment have already shaved that line; further tightening (which is broadly the policy direction under the current US administration) cuts a few hundred million of revenue per round. Second, the long-tail risk that as AI logic moves to 2nm and below, a fundamentally different inspection paradigm (multi-beam e-beam, or X-ray) displaces optical-pattern inspection — Hitachi and TEL have been investing in multi-beam e-beam for a decade, and a credible technology transition is the only scenario where KLAC's monopoly compresses.
Bull case
The base case is FY26 revenue of $13.5-14.0B (high-teens growth on $11.7B FY25), FY27 revenue of $15.5-16.5B, non-GAAP EPS rising from $36 in FY26 to $42-44 in FY27. At 38x forward — premium to the equipment peers but consistent with KLA's monopoly franchise — that prints the stock at $1,600-1,670 pre-split, $160-167 post-split — broadly flat to spot but with capital return underpinning. The base case is not where the alpha is.
The upside case is HBM packaging inspection ramping faster than the FY27 base assumes — the SK Hynix and Micron HBM4 lines collectively add roughly $500-700M of KLA TAM through 2026-27, and if KLA captures full share (likely, given the customer relationships) that revenue is upside to consensus. CoWoS-L and CoWoS-R inspection at TSMC adds another $300-500M. The split-driven retail flow plus broader equal-weight index buying creates a 5-10% multiple expansion. Aggregate upside case: stock at $2,200-2,400 in 2027 (or $220-240 post-split), +16-27% from spot.
The structural argument for KLAC carrying a premium multiple to LRCX, AMAT, and the broader equipment cohort is the recurring-service-revenue mix. KLAC's installed-base service revenue runs at approximately 30%+ of total revenue and grows in the high single-digits annually regardless of capex cycle — software updates, defect-recipe libraries, tool calibration services, and the broader customer-support business. This is structurally the closest thing in the equipment sector to a software-like recurring revenue model, and the equity has historically traded at a 20-30% multiple premium to LRCX and AMAT to reflect this. The 38x forward multiple at current levels sits at the high end of the historical premium range but is not at extremes.
Gap / bear case
Three concerns. First, the +155% off 52-week low has stretched the multiple — at 38x forward, KLAC is at the upper end of its 5-year range and a multiple compression back toward 32x on any beat-miss or guide-down prints the stock at $1,500 pre-split. Second, the DRAM cost dynamics that benefit SK Hynix and Micron equity also pressure DRAM-maker capex discipline — if the memory cycle softens in 2027 (which is the consensus path now that it has run hot for two years), the memory equipment book of KLAC takes a haircut. Third, the China-export-restriction overhang remains live — the latest round of export-controls in late 2025 specifically called out inspection equipment as a target category, and continued restrictions through 2026-27 compress KLAC's China book at the same time the broader equipment cohort is benefiting from the AI-cycle re-rate.
The +155% off lows also raises the question of what the marginal incremental buyer's thesis is. The stock has run from the 52-week low principally on AI-cycle multiple expansion plus consensus EPS revisions; a continued re-rate from current levels requires either (a) consensus EPS revisions higher than what the company is currently guiding, or (b) further multiple expansion from already-elevated levels. The 10-for-1 split-driven retail flow is the principal mechanism for option (b); without it, the asymmetric upside from current levels is more modest than the trailing 12-month return would suggest.
Optionality
Four options. Reticle-inspection demand from EUV mask qualification at TSMC, Samsung, and Intel scaling with the high-NA EUV (NXE:5000) installed base — each high-NA tool requires multiple reticle-inspection passes per wafer cycle, and the high-NA install base is expected to grow from approximately 10 systems in 2025 to 40+ by 2027. Hybrid-bonding alignment metrology — KLA has not loudly disclosed its hybrid-bonding franchise but the IP base sitting at the intersection of overlay + CD-SEM + bond-pad inspection is the natural extension. The split-driven retail-flow option: post-10:1 split, the equal-weight index inclusion plus retail derivatives liquidity historically add a one-time 100-200bp multiple expansion that has not yet fully arrived in the stock.
The fourth option is the CoWoS-L and CoWoS-R inspection ramp at TSMC. CoWoS capacity expansion through 2026-27 is the single largest equipment-pull line at TSMC, and KLA captures roughly $300-500M of incremental TAM per CoWoS capacity step. This is upside to consensus FY27 models that have been calibrated to the H2 25 TSMC capex announcement; subsequent capex revisions higher (which the supply chain has been signalling) directly flow to KLA revenue.
The trade
KLAC — BUY 8.5/10. Entry: starter at $1,890-1,900 spot; size up on $1,750-1,800 pullback. Post-split entry (post-late-Q2-FY26) at $180 equivalent. Position size: 2.5-3.5% of NLV as the highest-quality equipment monopoly position. Stop: daily close below $1,650 pre-split / $165 post-split (200-day MA breach). Catalyst date: July 30 Q4 FY26 print; the 10-for-1 split execution date in Q2 FY26; September Apple iPhone launch as a sentiment driver for the broader semi cohort. Trim/exit triggers: high-teens FY26 guide cut to mid-teens; any China-specific tightening removing >$500M annual revenue; multiple expansion above 40x forward without earnings revision. Conviction: the cleanest "boring monopoly" trade in the equipment cohort. The 10:1 split + 21% dividend hike is management telling you the next 24 months of revenue trajectory is durable. Own it in size.