Lam Research Corp · Moat analysis

Report previewLRCX.USReport date Last updated
← Back to reports
Report contents
01

Industry landscape

Where the company competes: an oligopoly of five, with etch as Lam's home turf

Lam Research builds the tools that shape and coat a silicon wafer. Its equipment sits in the middle of the semiconductor supply chain: upstream are precision component and subsystem makers (ceramics, RF generators, vacuum pumps, valves), and downstream are the roughly two dozen companies that own leading-edge fabs. Lam integrates components into process tools, sells them for $1 million to $10 million or more per unit depending on configuration, then collects recurring revenue from spare parts, upgrades and service on the installed base. The company reports two revenue lines: Systems, and Customer Support Business Group, which has supplied roughly a third of revenue in recent years. Total revenue came to $17.2 billion in fiscal 2025, and has ranged between about $14.6 billion and $17.4 billion in each of the last five fiscal years, per the company's annual reports.

The submarket Lam actually competes in is narrower than "wafer fab equipment." It sells three families of process steps: plasma etch, which cuts patterns into deposited films; thin-film deposition, chiefly atomic layer deposition, plasma-enhanced CVD and electrochemical deposition; and single-wafer cleaning, which removes residue between steps. Lam does not sell lithography scanners, the largest single slice of the fab tool budget, and it does not sell process control or metrology. Those exclusions matter because they set the boundaries of the arena where Lam's position is actually contestable, and they explain why its share of total fab spending understates its leverage over the steps it does supply.

Geographically the market is global in the sense that every leading-edge fab buys from the same short supplier list and qualifies tools against the same roadmaps, but demand is concentrated by country. Korea, Taiwan, China, Japan and the United States account for the overwhelming majority of tool purchases, and a single customer's capex plan can move Lam's quarterly revenue by several hundred million dollars. The customer base is the binding constraint on granularity here: Lam's 10-K discloses individual customers that exceed 10% of revenue, and the list has consistently been drawn from Samsung, TSMC, SK Hynix, Micron and Intel. Whether those five are memory makers or logic foundries changes the mix of process steps Lam sells, so the "market" Lam faces is as much a function of who is building fabs in a given year as of how many competitors it has.

Barriers are high, and the head count confirms it

The practical test is whether the leaders in each step can be counted on one hand. In etch they can: Lam, Tokyo Electron and Applied Materials supply nearly all merchant etch capacity, and Lam is generally described as the largest of the three. Deposition is a four-firm race, with Applied Materials, Tokyo Electron, ASM International and Lam; cleaning is three, with Tokyo Electron, Screen Holdings and Lam. Across the whole fab tool market, industry estimates put the five largest suppliers at roughly 70% of spending, which is about as concentrated as capital-goods markets get.

Process stepMain rivalsLam's standing
Plasma etchTokyo Electron, Applied MaterialsLargest of three; the core franchise
Thin-film deposition (ALD, PECVD, ECD)Applied Materials, Tokyo Electron, ASM InternationalStrong in ALD and electrochemical deposition; weaker in the broadest deposition slice
Single-wafer cleanTokyo Electron, Screen HoldingsOne of three credible suppliers

Entry barriers are high for reasons that show up in the income statement. Lam has spent above $1.5 billion a year on research and development, and a new etch chamber has to be qualified at a customer's process of record, a cycle that runs many months and often involves joint development. Once a tool is written into a production flow at a given node, replacing it means re-qualifying the step, which is why share tends to move at node transitions and not in between. The service network reinforces this: spare-parts availability and field engineering response times are part of what fabs are buying.

What that structure does not fix is cyclicality. A three-firm market still sees revenue halve when memory capex stops, as it did in fiscal 2024, when revenue fell to $14.9 billion from $17.4 billion. Concentration among suppliers does nothing about concentration among buyers, and it does not stop a customer from dual-sourcing a step to keep pricing honest.

The open question is China. Lam's China revenue has swung widely as US export controls (the Bureau of Industry and Security rules of October 2022 and subsequent tightening) reshaped which fabs can buy which tools, and domestic suppliers such as NAURA and AMEC are being funded to fill the gap in etch and deposition. So far the evidence only shows those vendors winning business inside China at trailing nodes. Whether they hold share once Chinese fabs push toward leading-edge logic is not yet established, and that, more than any move by Applied or Tokyo Electron, is what would change this assessment.

02

Competitive advantage

Does the company have a moat? Yes, and the returns make the case better than the share data

Moat rating: Strong. Lam Research supplies three steps in wafer fabrication: plasma etch, thin-film deposition and single-wafer cleaning. Each of those steps has three or four credible merchant suppliers. That structure has supported after-tax returns on invested capital far above any plausible cost of capital for a decade, including a year when revenue fell 14.5%. The profitability test does most of the work in this rating; the market-position evidence is real but less quantitative than for a company that publishes share.

Concentration and sustained excess returns are the two tests that matter because a moat is, by definition, above-normal returns protected by something that keeps entrants out. A tight supplier list that persists, alongside leaders who keep earning through the cycle, is evidence of that protection. High profits that appear in a boom and vanish in a bust point to execution, which rivals can copy.

Lam passes the profitability test without much argument. After-tax ROIC was about 60% in fiscal 2026 on year-end invested capital and about 70% on average invested capital, an estimate built from the company's reported operating income, tax rate, cash and current liabilities. On the year-end basis the figure stayed between roughly 44% and 63% in each of the five fiscal years through June 2026, and above 37% in every year since fiscal 2018. With an effective tax rate near 12% and almost no debt, the spread over a 10% hurdle stays wide on any reasonable definition of the capital base.

The downcycle is where the case becomes convincing. Revenue fell 14.5% in fiscal 2024, to $14.9 billion, and gross margin rose to 47.3% from 44.6% the year before, per the company's annual report. Net margin slipped less than a point. Customers cut capex; they did not extract price. A three-firm market without entry barriers tends to behave differently in a trough, when idle capacity gets used to buy share.

Market position: fairly strong, held by qualification more than by share

Market position: Fairly Strong. Lam calls itself a leader in each of its three process areas in its fiscal 2026 10-K, and it is the largest of the three merchant etch suppliers. It does not disclose market share, and no consistent third-party series covering all three steps across five years is publicly available, so the rating rests on the structure of each step, the customer list and the mechanism that keeps tools in place.

The framework's rule of thumb for a strong position is roughly 40% share plus a clear number-one rank. Lam does not publish a number, so the test can only be applied where it is clearly the largest of three, in etch; in deposition and clean it is one of three or four. That mix supports Fairly Strong rather than Strong.

What can be checked is stability. Lam's 10-K names the customers above 10% of revenue each year, and the list has stayed around Samsung, TSMC, SK Hynix, Micron and Intel. Four customers accounted for 55% of fiscal 2026 revenue. None of those accounts has publicly switched its primary etch supplier at a leading-edge node, the event that would show up first if share were eroding. That is a proxy, not a share series, and it would not catch a drift of a few points inside a node.

Profitability: strong, and it held through the trough
Fiscal yearRevenue (USD bn)Gross marginOperating marginNet marginROEROIC (est.)
202217.2345.7%31.2%26.7%74.8%52%
202317.4344.6%29.7%25.9%62.3%49%
202414.9147.3%28.6%25.7%45.7%44%
202518.4448.7%32.0%29.1%58.2%63%
202623.2350.5%35.3%31.3%65.1%60%

ROIC estimated as after-tax operating income divided by year-end total assets less current liabilities and cash.

Revenue compounded but not in a line: $17.4 billion in fiscal 2023, $14.9 billion in fiscal 2024, then $18.4 billion and $23.2 billion. Gross margin climbed 5.9 percentage points over the five years and net margin 4.6 points. The company's September 2026 quarter continued the trend, with revenue of $8.22 billion, gross margin of 52.3% and net margin of 33.5%.

Return on equity overstates the edge, because buybacks have shrunk book equity to $12.5 billion against $7.3 billion of net income. ROIC is the better measure and carries its own asterisk: invested capital is thin at $12.0 billion because Lam outsources much of its manufacturing and holds little net PP&E. Scale of profit against a small asset base, not asset intensity, is what produces the number. Applied Materials and Tokyo Electron report under segment structures that do not yield a comparable ROIC, so the cleaner peer comparison is Lam against its own cycle, and there the stability is the point: a 14.5% revenue decline cost about 5 points of ROIC.

Overall rating: strong, with price protected more than volume

Rated Strong. The profitability evidence is unambiguous and persistent, and the supplier structure of each served step is tight enough to explain why. The rating would be higher on market position if Lam published a share series showing stability within 2 percentage points over five years; the proxy evidence supports stability but does not measure it.

Two limits are worth naming. Buyer concentration means four customers can move revenue by hundreds of millions of dollars on a capex decision. And the moat protects the price of a qualified tool and the service annuity attached to it, not the volume of tools shipped. Revenue can halve in a memory downturn with the moat intact, which is a distinction that matters for anyone treating the franchise as a bond substitute.

Where the advantage comes from and how long it can last: qualification lock-in, paid for by the installed base

Two sources carry the advantage: customer captivity that comes from qualifying a tool into a production flow, and economies of scale in research and field service that the installed base pays for. A third possible source, a supply-side cost advantage, is thin. Durability then turns on two questions: whether process changes reduce the number of steps Lam sells, and whether Chinese suppliers can qualify at leading-edge nodes.

Customer captivity: fairly strong, and renewed at every node

Customer captivity: Fairly Strong. The mechanism starts with how a fab picks a tool. A process engineer selects a chamber for a specific step at a specific node, runs wafers, matches results against the surrounding steps and writes the choice into the process of record. That takes months of scarce engineering time and carries real yield risk. Lam's 10-K states that once a chipmaker qualifies a tool for a step on a node, it "generally maintains that selection" for that node.

That gives Lam two kinds of pricing power. The first is on the tool itself, where the customer is comparing a $5 million chamber against the cost of a failed qualification. The second, and the more valuable one, is on what follows: spare parts, service contracts and upgrades. Customer support revenue was $8.35 billion in fiscal 2026, about 36% of total revenue, per the 10-K. That annuity is billed against tools already in the field, so it does not disappear when new orders slow.

The margin record shows the lock-in has commercial value. In fiscal 2024, with revenue down 14.5%, gross margin rose. A supplier whose customers could credibly re-bid every chamber would have seen the opposite.

The limit is that captivity is per node, not permanent. Every node transition is a fresh evaluation, and that is when share moves in this industry. Captivity also varies by step: etch chambers are yield-critical and change slowly, while single-wafer clean is closer to a commodity with three suppliers and lower switching pain. Buyer power caps the whole thing. With four customers at 55% of revenue, each one can dual-source a step to keep Lam honest, and some do. Fairly Strong, not Strong, reflects that ceiling.

Scale: strong, and the installed base funds research through the trough

Scale: Strong. The fixed cost that matters most is R&D. Lam spent $1.90 billion on research in fiscal 2024, the year revenue fell 14.5%, the company reported in its 10-K; spending rose while sales shrank. That is the mechanism in one line. A toolmaker that cuts R&D in a downturn arrives at the next node with an uncompetitive chamber and loses the qualification. Lam funds research from the service annuity on its installed base when systems revenue is weak. Smaller rivals without that base have to fund it out of tool sales alone, which is precisely when tool sales are worst.

The second fixed cost is field coverage. Fabs buy uptime, and response time depends on how many qualified process engineers a supplier can put inside a cluster of fabs. A larger installed base supports a denser service footprint at the same cost per tool, which both wins the next tool and defends the last one. Applications engineering is the third: tuning a chamber for a customer's specific stack is a per-customer fixed cost that spreads across every step Lam holds at that account.

The combination with captivity is what makes scale a barrier instead of a convenience. Where switching costs are low, a new entrant can buy its way in with price and grow into scale. Where qualification locks the incumbent in for a node, an entrant has to win at the transition and then hold the following node on its own research budget. That is a slower, more expensive path, and it is why the supplier count in etch has not changed in years. Scale here is relative to the step, though: Lam is not the largest supplier in deposition, and its scale advantage is narrower there than in etch.

Supply-side cost advantage: weak, and mostly a reflection of scale

Supply-side cost advantage: Weak. Lam owns no fabs of consequence and controls no scarce material. It designs chambers and outsources a large share of manufacturing, which keeps capital intensity low and returns high but leaves little proprietary cost position. Process know-how accumulates in etch, yet it diffuses through employee movement and through customers who see the results. What cost edge exists comes from purchasing volume and design reuse across a larger unit base, which is the scale advantage described above wearing a different label.

Durability: the moat holds unless Chinese suppliers climb the node curve

Substitution at the level of the need is not the near-term threat. Fabs still have to deposit and etch films, and the direction of travel raises the number of times they do it: more layers in 3D NAND, gate-all-around transistors, backside power delivery. Management said on its July 2026 earnings call that it expects to outgrow wafer fab equipment spending for a third straight year. That is guidance, not a result, but the served-market math behind it is checkable against the tool intensity of each new architecture.

A real substitution risk runs the other way. If high-NA EUV lithography reduces the number of multi-patterning passes at leading-edge logic, some etch and deposition steps per wafer disappear. That is an inference from how the process flow works, not something the evidence settles; it would show up first as Lam underperforming fab equipment spending at a logic customer.

The more concrete threat is geographic. China was 42% of revenue in fiscal 2024 and 26% in the June 2026 quarter, per the company's disclosures, the swing driven by US export controls that closed part of that market to Lam's tools. Those controls also funded a domestic alternative. NAURA and AMEC are winning business inside China, so far at trailing nodes. If either qualifies an etch chamber at a Chinese leading-edge logic fab, the moat narrows at the edge rather than at the core, and Lam's revenue base shrinks without its margins moving much.

What would change the assessment: two consecutive years of revenue growth below wafer fab equipment spending while ROIC falls under 40%, or the loss of a qualified etch step at a major logic customer at a node transition. Neither has happened. The evidence so far shows a moat that protected price through the last downturn and a service annuity that funds the next node's research. It does not yet show that the moat travels intact into a market split between US-allied and Chinese fabs.

Generated by DeepVal AI. Data and models may contain errors or be outdated. For research only; not personalized investment advice.

Continue the comparison

Continue reading the full report

  1. Comparison and summaryHow Lam Research Corp compares with its closest rivals, and our moat rating

Free to sign up · 20 research points · unlocking this report uses 10

20 points cover 2 full reports. No card needed.

Research reflects its generation date. Preview the first two modules.

More on Lam Research Corp