KLA Corporation (KLAC) - Deep Dive Research Report

Prepared 2026-06-21. Fiscal year ends June 30. Most recent reported period: Q3 FY2026 (quarter ended March 31, 2026), reported April 29, 2026.

1. What the Company Does

Every advanced computer chip is built by stacking dozens to over a hundred patterned layers onto a silicon wafer, each layer printed at dimensions far smaller than a wavelength of visible light. At those scales, a speck of dust, a misaligned pattern, or a film a few atoms too thick can kill the chip. A modern leading-edge wafer can carry hundreds of billions of transistors, and a single fab can lose tens of millions of dollars if a process drifts out of spec for even a few hours. The problem KLA solves is simple to state and brutally hard to do: find the defects, measure the dimensions, and tell the chipmaker exactly where and why their process is going wrong, fast enough to fix it before the scrap piles up.

KLA makes the inspection and measurement machines, plus the software and services, that sit between every major step of chip manufacturing and act as the fab's eyes and quality-control brain. This discipline is called "process control." When TSMC prints a layer, a KLA tool scans the wafer for particles and pattern defects; another KLA tool measures whether the linewidth and the overlay between layers are within tolerance; KLA software correlates all of it to pinpoint which process tool or step caused a yield excursion. The chipmaker uses that information to ramp a new process to high yield faster and to keep a mature process from quietly bleeding money. KLA does not make the chips and does not deposit or etch the main circuit layers (that is Applied Materials, Lam Research, and ASML's domain) - it is the referee that tells everyone else whether their step worked.

The company was built by merging the two pioneers of this niche. KLA Instruments was founded in 1975 by Ken Levy and Bob Anderson to automate photomask defect detection, then moved into wafer inspection. Tencor was founded in 1976 by Karel Urbanek and John Schwabacher to measure semiconductor films and detect contamination with laser scanning. The two merged in 1997 to form KLA-Tencor, the first single-source supplier of chip process diagnostics. The company dropped "Tencor" and became KLA Corporation in 2019, the same year it closed the $3.4 billion acquisition of Israel's Orbotech, which pushed KLA into printed-circuit-board and display inspection and, through Orbotech-owned SPTS Technologies in Wales, into etch and deposition equipment for advanced packaging. KLA is headquartered in Milpitas, California, and employs roughly 15,000 people.

The reason process control is such a good business is that it is a small slice of fab spending that controls the outcome of the entire rest of the spend. A leading-edge fab might allocate around 15% of its equipment budget to process control, but the yield that process control protects determines the return on the other 85%. That asymmetry, combined with the fact that KLA has owned the hardest parts of the problem (optical and e-beam inspection at the bleeding edge) for decades, is why the company holds the kind of share that almost no other equipment category exhibits.

Management frames the business around what they call "the most demanding and complex inspection and metrology challenges" in the industry - the steps where being second-best simply does not get you qualified.

2. Business Segments

KLA reports in three segments. The first dwarfs the other two, but the smaller two are the company's growth options and the reason it touches advanced packaging and PCBs.

Semiconductor Process Control (~85-88% of revenue)

This is the heart of KLA and the source of its dominance. It contains the full portfolio of inspection, metrology, and yield-management software used by IC, wafer, and reticle/mask makers from R&D through high-volume production. The product families include optical wafer-defect inspection (brightfield and darkfield), e-beam inspection and review, reticle (photomask) inspection, optical critical-dimension and overlay metrology, film and thin-film measurement, wafer-shape and stress metrology, and the software (RAPID, Klarity, and related yield-analysis tools) that ties defect and measurement data back to specific process steps.

The core capability here is roughly half a century of accumulated physics, optics, algorithms, and reference data that let a KLA tool find a defect a few nanometers across on a wafer moving past at high speed, and distinguish a killer defect from harmless noise. That sensitivity-versus-throughput tradeoff is the entire game, and it took decades of co-development with the leading chipmakers to get right. Every new process node creates new defect types that KLA has to learn to catch, and because KLA sees defects across the whole industry's leading edge, it has a data and learning advantage that a new entrant cannot buy.

This segment exists as the company's identity because it is where the moat lives. KLA holds well over half of the total process-control market, more than 85% of optical wafer inspection, and over 80% of reticle inspection. Its closest competitor in inspection and metrology, Applied Materials, holds under 10% of that combined market. End markets are foundry/logic (around 82% of process-control systems revenue) and memory (around 18%, split heavily toward DRAM over NAND). This is the margin engine, the cash cow, and the strategic core all at once.

Specialty Semiconductor Process (~5% of revenue)

This segment, built largely on SPTS Technologies (acquired via Orbotech), provides etch and deposition systems for advanced packaging and specialty devices: MEMS, RF filters, power devices (including silicon carbide and gallium nitride), photonics, and the through-silicon-via and redistribution-layer steps that stitch multiple chips into one package. The Newport, Wales facility is the center of gravity. Unlike the core segment, this is process equipment (it physically builds device structures), not inspection.

It exists separately because the technology, the customer base (often OSATs and specialty foundries rather than leading-edge logic fabs), and the economics differ from process control. Its strategic value has risen sharply with the AI era: advanced packaging - stacking HBM memory onto logic, building CoWoS-style 2.5D and 3D modules - has become a bottleneck for AI accelerators, and KLA sells both the specialty process tools and the inspection that goes with packaging. Competitors here include Applied Materials, Lam Research, and smaller specialty players.

PCB and Component Inspection (~7% of revenue)

This is the Orbotech-derived business: automated optical inspection and direct-imaging systems for printed circuit boards, advanced IC substrates, and electronic components. As AI servers drive demand for high-layer-count substrates and complex PCBs, this segment has picked up packaging-related volume. It is the most cyclical and lowest-margin of the three and is the least strategically central, but it gives KLA a foothold in the packaging and substrate supply chain that complements the specialty-process and core inspection businesses. KLA exited the separate Display inspection business to concentrate on semiconductor-adjacent markets.

3. Products and Business Detail

The product catalogue is best understood by where each tool sits in the fab flow.

Patterned and unpatterned wafer inspection. KLA's brightfield optical inspectors (the long-running 29xx/39xx families) scan a printed wafer for pattern defects; darkfield inspectors (Puma and related lines) catch particles and scattering defects at high throughput. The Surfscan line inspects bare, unpatterned wafers for the silicon-wafer makers and for monitoring fab cleanliness. At the hardest nodes, optical alone is not enough, so KLA's e-beam inspection and review tools (the eDR/eSL families) use electron beams for higher resolution at lower throughput, used to find and classify the smallest killer defects and to verify what optical tools flag.

Reticle inspection. The photomask is the master template; a single defect on it repeats onto every wafer. KLA's reticle inspection tools (Teron family) protect the mask, both at the mask shop and in the fab, and KLA holds the dominant share here. This is one of the most defensible niches in the whole company because the cost of a bad mask is catastrophic and qualification cycles are long.

Metrology. KLA measures critical dimensions (how wide a feature is), overlay (how precisely one layer aligns to the one below), and film properties (thickness, composition, stress). The Archer overlay tools and the SpectraShape/optical-CD systems are the workhorses. As nodes shrink and structures go vertical (gate-all-around transistors, 3D NAND stacks of hundreds of layers), the number of measurements per wafer rises, which is the structural tailwind management calls "process-control intensity."

Specialty process tools. Through SPTS, KLA sells deposition (PVD, CVD) and etch systems tuned for packaging interconnects, MEMS, RF, and compound-semiconductor power devices.

PCB and component AOI. Orbotech direct-imaging and inspection systems for boards and substrates.

Software and services. Klarity and related yield-analysis software turn raw defect/metrology data into root-cause insight. Services - installed-base support, spare parts, upgrades, and contracts - is roughly a quarter of revenue and runs around $775 million a quarter, with management citing more than 50 consecutive quarters of year-over-year growth. Services is the recurring, counter-cyclical ballast on top of the lumpy systems business.

The manufacturing model is concentrated, high-mix, and low-volume: these are multi-million-dollar precision instruments built in limited numbers, with core operations in California, Israel (Orbotech), and Wales (SPTS). The constraint is not factory floor space but the optics, sensors, precision stages, and software calibration that go into each tool. Geographically, KLA sells almost entirely outside the US (roughly 89% of revenue is international), with China, Taiwan, and Korea the three largest destinations because that is where leading-edge logic and memory capacity is concentrated.

4. Customers

KLA's customers are the world's chipmakers, and the list is short because leading-edge manufacturing is concentrated. The biggest accounts are TSMC, Samsung, and Intel in logic/foundry; SK Hynix, Samsung, and Micron in memory; GlobalFoundries and others at trailing edge; and the OSATs (ASE, Amkor) plus specialty foundries for the packaging and specialty-process tools. The silicon-wafer makers and mask shops buy the bare-wafer and reticle inspectors.

The buying decision is made by a chipmaker's process-integration and yield-engineering organizations, not procurement alone. The criteria are sensitivity (can the tool find the defect that matters), throughput (can it do so fast enough to be economic), and the quality of the data and software that turn a defect map into an actionable fix. The sales cycle is long and deeply technical: KLA co-develops next-generation inspection capability with the leading customers years ahead of a node ramping, so the tool is effectively designed into the process before high-volume orders arrive.

Switching costs are unusually high. Once a KLA tool is qualified into a process flow, the customer's yield models, defect libraries, and engineering workflows are built around its output. Swapping to a competitor's inspector means requalifying the entire monitoring scheme and risking a yield blind spot during the transition - rarely worth it when yield is the customer's single largest economic lever. That lock-in, plus KLA's data advantage from seeing the whole industry's leading edge, is why share has actually expanded rather than eroded.

Concentration is real and cuts both ways. A handful of leading-edge logic and memory customers drive a large share of systems revenue, so a capex pause at TSMC or a memory downturn hits KLA directly. But that concentration is also a reflection of quality: KLA wins precisely because the most demanding customers cannot afford to be wrong on inspection. Revenue is a mix of lumpy systems sales (tied to customer capex timing) and the steadier services stream, which smooths the cycle and improves predictability.

5. Competitive Landscape

Process control is the most concentrated category in all of semiconductor equipment, and KLA sits at the top of it. The structure is best described as one dominant generalist (KLA) surrounded by specialists who win specific niches.

The nearest broad competitor is Applied Materials, which has an inspection-and-metrology business but holds under 10% of that combined market and, per independent market trackers, lost share to KLA in 2025. Onto Innovation (formed from Rudolph Technologies and Nanometrics) is a focused metrology and packaging-inspection player that competes hard in specific metrology and advanced-packaging niches. Nova is an Israeli metrology specialist strong in optical CD and materials metrology. Camtek competes in packaging and PCB inspection. Lasertec of Japan owns the actinic EUV mask-inspection niche where KLA is comparatively weaker. Hitachi High-Tech competes in e-beam metrology and review. ASML, primarily a lithography company, also fields e-beam metrology and computational tools that overlap at the edges. Carl Zeiss and SCREEN appear in adjacent optics and process niches.

KLA wins on breadth, sensitivity at the bleeding edge, and the software/data flywheel: it sees more of the industry's hardest defects than anyone, which sharpens its algorithms, which wins the next node. It loses, or is simply absent, in a few defined pockets - actinic EUV mask inspection (Lasertec) and certain specialty metrology lines where Nova or Onto have a better-fit tool. The barriers to entry are among the highest in capital equipment: decades of optics and algorithm IP, deep co-development relationships, an enormous installed base feeding a high-margin services annuity, and qualification lock-in that makes customers reluctant to switch even when a competitor closes a technical gap. A new entrant cannot simply build a better inspector; it has to also be trusted, qualified, and supported across a customer's global fleet.

(Market caps are rough peer-size references only and move continuously.)

6. Industry

KLA's demand is driven by semiconductor capital spending, and specifically by the leading edge of it. The ultimate drivers today are AI infrastructure (accelerators and the HBM memory stacked beside them), high-performance computing, and the broad digitization that keeps adding silicon content to everything. Within that, KLA benefits from a second-order tailwind it calls process-control intensity: as transistors go 3D (gate-all-around), as memory stacks climb past hundreds of layers, and as chips get assembled by advanced packaging, the number of inspection and measurement steps per wafer rises faster than the number of wafers, so KLA's served market grows faster than overall wafer-fab equipment spending.

Wafer-fab equipment is a large, multi-tens-of-billions-of-dollars annual market, and process control is roughly 15% of it; independent trackers and SemiAnalysis-style bottom-up wafer-fab models forecast continued growth through 2027 led by advanced logic and HBM. The advanced-packaging slice specifically has gone from a niche to a fast-growth pocket: KLA's own advanced-packaging-related revenue rose from around $500 million in 2024 to over $925 million in 2025 and is guided toward roughly $1 billion in 2026.

KLA sits at a chokepoint in the global chip supply chain. It does not make chips, but almost no leading-edge wafer reaches volume production without passing through KLA inspection and metrology repeatedly. That position makes it geopolitically exposed: process-control tools are squarely inside US export-control regimes, and China is both a major customer and a policy flashpoint. China fell from around 41% of KLA revenue in fiscal 2024 to roughly 30% in fiscal 2025 as export controls and a normalizing China capex wave took hold.

The industry is cyclical - memory in particular swings hard, and even logic capex pauses between build-outs. KLA dampens this with its services annuity (recurring, installed-base-driven, growing through downturns) and with its leading-edge mix, which holds up better than trailing-edge commodity capacity. The structural tailwind (AI, rising process-control intensity) is currently overwhelming the cyclical noise, but the cyclicality has not been repealed.

7. Growth Triggers

All points below are drawn from the six most recent earnings calls.

• Sequential revenue growth expected throughout calendar 2026, with 2027 growth exceeding 2026. Management stated demand visibility is unusually strong, with customers showing urgency to secure capacity. (Q3 FY2026 call, April 29, 2026)

  "We expect to continue to see quarter-to-quarter revenue growth throughout 2026... customers are exhibiting a higher level of urgency around securing capacity." - Rick Wallace (Q3 FY2026 call, April 29, 2026)

• Advanced-packaging revenue guided to roughly $1 billion in calendar 2026, up from ~$635 million in 2025, more than 50% growth, driven by HBM and 2.5D/3D AI packaging. (Q3 FY2026 call, April 29, 2026)

  "Advanced packaging... will grow from approximately $635 million in 2025 to approximately $1 billion in 2026." - Bren Higgins (Q3 FY2026 call, April 29, 2026)

• Advanced-packaging revenue raised twice during fiscal 2025 - from $850 million to over $925 million for calendar 2025, versus ~$500 million in 2024 - as CoWoS and HBM demand ramped. (Q4 FY2025 call, July 31, 2025; repeated/raised across Q1-Q3 FY2026)

• Services revenue guided to a ~13-15% CAGR through 2030, extending a streak of more than 50 consecutive quarters of year-over-year service growth. (Q3 FY2026 call, April 29, 2026; theme repeated Q4 FY2025 and Q1 FY2026)

• Leading-edge foundry/logic and HBM investment cited as the primary systems demand engine, with AI described as a core driver across customer segments. (Q1 FY2026 call, October 29, 2025; Q2 FY2026 call, January 29, 2026)

• Broad-based greenfield fab construction expected to lift 2027 above 2026, as new fabs across multiple customers begin equipping. (Q3 FY2026 call, April 29, 2026)

• New $7 billion share-repurchase authorization and a 17th consecutive annual dividend increase, signaling management's confidence in forward free cash flow. (Q3 FY2026 call, April 29, 2026)

8. Key Risks

China export controls and revenue concentration. China went from roughly 41% of revenue in FY2024 to about 30% in FY2025, and management disclosed that China customer revenue fell year over year in the December 2025 quarter "primarily due to the effects of U.S. export controls." The mechanism: tightening rules can wall off a large customer base overnight, and the policy direction is unpredictable. This is a high-probability, moderate-to-significant drag that is already partly playing out, with tail risk of a sharper cut.

Customer concentration at the leading edge. A small number of logic and memory customers (TSMC, Samsung, Intel, SK Hynix, Micron) drive a large share of systems revenue. A capex pause, a delayed node, or a memory glut at any one of them flows almost directly into KLA orders. This is structural and unavoidable given how concentrated leading-edge manufacturing is.

Cyclicality, especially memory. Systems revenue is tied to customer capex, which swings. Memory is the most volatile end market; a DRAM or NAND downturn would cut orders quickly. The services annuity and the leading-edge logic mix cushion this, but do not eliminate it.

Margin drag from input costs. Management flagged that elevated DRAM chip costs are creating roughly a 100-basis-point gross-margin headwind expected to persist through calendar 2026 - a reminder that KLA is also a buyer of semiconductors and is exposed to the same supply tightness that drives its top line.

"Elevated DRAM costs... represent approximately a 100 basis point headwind to gross margin... we expect this to persist through calendar 2026." - management commentary (Q3 FY2026 call, April 29, 2026)

AI-cycle dependence. The current growth surge leans heavily on AI infrastructure and HBM. If AI capex expectations reset, the advanced-packaging and leading-edge demand that is driving above-trend growth could decelerate faster than the broader WFE market. This is a lower-probability-but-high-impact scenario given how much of the optimism in the recent calls rests on AI urgency.

Niche competitive gaps. In actinic EUV mask inspection (Lasertec) and certain specialty metrology lines, KLA is not the leader. If those niches grow in importance and KLA cannot close the gap, it cedes share in pockets even as it dominates the whole.

9. Walk the Talk

The six calls used, oldest to newest: Q2 FY2025 (January 30, 2025), Q3 FY2025 (April 30, 2025), Q4 FY2025 (July 31, 2025), Q1 FY2026 (October 29, 2025), Q2 FY2026 (January 29, 2026), and Q3 FY2026 (April 29, 2026). The most recent is within 90 days of today.

The throughline across these six quarters is a management team that set conservative, specific targets and then repeatedly raised them - the pattern of an under-promiser rather than a hype machine. The clearest test case is advanced packaging. In the December 2024 quarter (Q2 FY2025), management was navigating fresh export controls and guided cautiously, with the March quarter set at roughly $3.0 billion. They beat it. By Q4 FY2025 (July 2025) they put a hard number on advanced packaging: "We now expect advanced packaging systems related revenue to exceed $925 million in calendar 2025, up from our previous estimate of $850 million last quarter." That was itself an upward revision of an earlier $850 million figure, against roughly $500 million the year before. The number was not a one-off boast; it was raised in steps across consecutive calls, and by Q3 FY2026 the framing had moved on to "approximately $1 billion in 2026." When a company keeps having to revise its own forecast upward by similar increments, that is the signature of guidance set below the line.

The services commitment is the second clean track record. Through every one of these six calls, management reaffirmed that services was growing year over year - they cited the streak passing 50 consecutive quarters - and they progressively firmed up the long-term frame into a "13-15% CAGR through 2030." This is a promise that is easy to check quarter to quarter, and it held in each of the six.

On the topline cadence, the Q3 FY2026 guide for the June quarter was about $3.575 billion, with management saying explicitly they expect "quarter-to-quarter revenue growth throughout 2026." That sequential-growth claim is consistent with the actual path the prior quarters traced (roughly $3.0B → $3.2B → $3.3B → $3.4B across the window), so the guidance has matched the realized trajectory rather than overshooting it.

Where management has been appropriately honest rather than promotional is on headwinds. They did not bury the China export-control hit; they quantified the year-over-year China decline and attributed it plainly to policy. They flagged the ~100 basis point DRAM-cost margin drag and told investors it would persist through 2026 rather than hand-waving it away. Nothing in the six calls reads as a quietly dropped promise; the targets that were set were either met or raised.

Assessment: this is management that does what it says and tends to set the bar low enough to clear it. The credibility read is high - consistently accurate and modestly conservative, with honest disclosure of the negatives.

10. Shareholder Friendliness Index

KLA pays a steadily rising dividend and is a consistent, sizeable repurchaser of its own stock. The quarterly dividend has climbed through consecutive annual increases - from $1.45 (late 2023) to $1.70 (2024) to $1.90 (2025), marking the 16th consecutive annual raise in 2025 - and management announced a 17th consecutive annual increase alongside the Q3 FY2026 results (April 2026). Dividends have compounded at roughly 15% annually since the program began in 2006. This is a clearly growing, well-covered payout, not a token one.

On buybacks, KLA repurchases shares every year and the share count is genuinely shrinking, not just offsetting option dilution. The board authorized a $5 billion repurchase program in 2025 and added a new $7 billion authorization in 2026. In the March 2026 quarter alone, KLA returned $875 million ($626 million of buybacks plus $249 million of dividends), and roughly $3.2 billion over the trailing twelve months. Diluted shares outstanding have fallen on the order of 3% per year, from around 140 million (FY2022) toward roughly 133 million (FY2025) - a real retirement of stock, not flat dilution-offset. (Buyback figures: MoatMap recorded none in the trailing ~90-day window, which for US issuers reflects that aggregate open-market repurchases are disclosed in 10-Q/10-K filings rather than as individual insider transactions; the multi-year program figures above come from KLA's earnings releases and capital-return disclosures.)

Verdict: Returns Capital - a growing 15%-CAGR dividend plus consistent buybacks that are actually shrinking the share count make KLA a clear returner of capital.

11. Insider Activities

Source note: the spine for this section is the injected MoatMap database (US venue, SEC Form 4), which was last scraped 2026-06-19 23:00 UTC - about 41 hours before this report - and is flagged stale, so filings in the most recent two days may be missing. I attempted to cross-check the trailing two weeks against primary SEC/aggregator sources; the open-market picture below should be treated as complete through roughly June 19, 2026.

Over the last 12 months, all recorded insider activity is selling - four open-market sales by three insiders, no purchases.

Buys: None. There were no open-market insider purchases in the window. This is not unusual for a large-cap whose executives are compensated primarily in equity, but it means there is no conviction-buy signal to read here.

Sells - the why: The two CEO sales are identical in size (4,512 shares each, one month apart on the 11th/12th of consecutive months), which is the classic fingerprint of a pre-arranged Rule 10b5-1 trading plan - automated, scheduled diversification rather than a discretionary reaction to the business. The May 11 sales by the CFO-adjacent Chief Accounting Officer (Kirloskar) and a director/officer (Hanley) are small and cluster around the same date, consistent with routine post-vesting diversification or scheduled-plan sales. None of the four is large relative to a senior semiconductor executive's equity-heavy compensation, and none breaks an established pattern. Where the precise reason is not stated in a footnote, the regularity and uniform sizing strongly imply scheduled 10b5-1 selling rather than a signal about the outlook.

Net assessment: Insiders were net sellers over the trailing 12 months, but the activity is light, concentrated in a few people, and bears the hallmarks of scheduled, plan-driven diversification rather than informed bearishness. There is no cluster buying and no first-time CEO purchase to flag, but equally no large or unusual disposal. Read: neutral - routine equity-comp selling with no actionable signal in either direction.

12. Scenarios

Bull case. The AI build-out keeps pulling demand forward, and KLA's structural tailwind - rising process-control intensity - means its served market grows faster than overall fab spending. Gate-all-around logic, 400-plus-layer 3D NAND, and ever-taller HBM stacks each multiply the number of inspection and measurement steps per wafer, and KLA captures most of them because no one can match its sensitivity or its defect-data flywheel. Advanced packaging blows through the $1 billion mark and keeps compounding as every AI accelerator needs more sophisticated 2.5D/3D assembly that KLA both processes (via SPTS) and inspects. The wave of greenfield fabs across the US, Taiwan, Korea, and Japan all equip with KLA tools, lifting 2027 above 2026 as management has flagged. China stabilizes at a smaller but workable share rather than disappearing. Services compounds at the high end of its 13-15% range, giving the model a recurring spine that flattens the cycle. KLA stays the referee that everyone has to pay, and its share inches higher as smaller competitors keep losing ground.

Base case. Management delivers roughly what it has guided: sequential revenue growth through 2026, advanced packaging around $1 billion, services in the mid-teens, and continued share-count reduction plus dividend growth. Process-control intensity keeps the franchise growing through the noise, but the cycle is not abolished - some quarters lean on leading-edge logic, others wait on memory. China remains a managed headwind near 30% of revenue, with export-control news periodically rattling sentiment without breaking the model. KLA's dominance holds, margins stay high with a modest DRAM-cost drag, and the company keeps returning the bulk of its free cash flow. Nothing breaks; nothing dramatically surprises. The stock's story remains "the most defensible niche in semiconductor equipment, riding the AI cycle."

Bear case. AI capex expectations reset, and the advanced-packaging and leading-edge urgency that powered the recent upgrades reverses faster than the broad WFE market. A memory downturn coincides with a logic capex pause, and because KLA's customers are so concentrated, orders fall quickly. Export controls tighten further and wall off more of the China base, turning a managed 30% headwind into a sharper cut with no near-term replacement. In the niches where KLA is not the leader - actinic EUV mask inspection, certain specialty metrology - competitors entrench and the "we win every new node" narrative develops cracks. Input-cost pressure (the DRAM drag) lingers and margins compress just as the topline softens. None of this threatens KLA's existence - the moat is real and the services annuity keeps cash flowing - but it would turn an above-trend grower back into a cyclical that trades with the chip cycle, and the premium narrative built on uninterrupted AI-driven growth would deflate.