Semiconductor capital equipment. LRCX / ASML. Run 14 September 2026. Built from Lam Research Forms 10-K for fiscal 2024, 2025 and 2026, ASML Forms 20-F for 2024 and 2025, both companies' 2026 quarterly filings, the Lam Q4 FY2026 earnings call and the Lam presentation at Goldman Sachs Communacopia on 10 September 2026. Events swept through 14 September 2026; most recent events checked: Lam at Goldman Sachs Communacopia, 10 September 2026, and ASML's Q2 2026 results and raised full-year guidance, 15 July 2026. Lam figures in US dollars, ASML figures in euros; all ratios are percentages of each company's own revenue and are therefore free of any exchange-rate effect, and no figure in this page is converted between currencies. Not a valuation and not a recommendation.
The mechanism is served-available-market expansion inside a pool that is itself growing, and it is arithmetic rather than ambition. Lam's share of addressable wafer fab equipment spending is roughly 36 to 36.5% today against a high-30s target, and management says it is getting there faster than the plan set at its 2025 Investor Day; every point of that on a 2026 equipment pool guided to the low USD 150bn range is about USD 1.5bn of revenue Lam could not previously address. The content is named and datable: gate-all-around adds roughly USD 1bn of addressable market per 100,000 wafer starts per month of new capacity and backside power adds a similar amount, addressable market per NAND wafer doubles from the 128-layer node to 500-plus layers, and the through-silicon via step for high-bandwidth memory is a Lam process.
ASML's forward mechanism is narrower by construction, which is the price of a monopoly on one step: more EUV in more fabs, with advanced logic EUV spending growing 10 to 20% a year to 2030 and DRAM 15 to 25%. Its own long-range frame has already been overtaken. The EUR 43bn to 45bn guided for 2026 meets the EUR 44bn low end of ASML's published 2030 scenario four years early, which implies roughly a 4% compound rate from here to the EUR 52bn moderate case (inferred: arithmetic on ASML's own published scenarios and its 15 July 2026 guidance). The cells carrying Lam's next five years are etch and deposition into leading-edge foundry in Taiwan and the United States, DRAM and high-bandwidth memory in Korea, and the 500-layer NAND transition. The cells carrying ASML's are EUV into advanced logic in Taiwan and DRAM in Korea, and both are constrained by how many systems it can build rather than by how many customers want.
ASML converts growth into margin through mix rather than volume, and the proof sits in its own technology table: unit shipments fell from 583 systems in 2024 to 535 in 2025 while system revenue rose 11.6%, because an EUV system carries an average selling price of roughly EUR 242m against about EUR 12.8m for a KrF tool (inferred: value divided by units, ASML FY2025 20-F). That is what pricing power looks like when it is structural. The customer must buy the expensive tool to get the node, so realised price rises without a list-price negotiation. Three-year averages put ASML ahead on gross margin, 51.8% against 48.8%, and on operating margin, 33.1% against 32.0%.
The direction of travel belongs to Lam, and this is the most valuable finding of the run: the growth winner and the margin winner are converging rather than diverging, which is unusual. Lam's gross margin rose 320bps over three years against ASML's 150bps, narrowing the gap from 4.0 points to 2.3, and its operating margin moved 670bps against 180bps. Lam's mechanism is self-help that has not run out, and it names all three parts: volume still shifting into its low-cost Malaysia plant, new differentiated products carrying more content each year, and pricing actions. It has raised its long-term target to a mid-50s gross margin and a mid-40s operating margin, from 50.5% and 35.3% in fiscal 2026. ASML's mechanism is further EUV mix plus a service line at 25.1% of sales that nobody else is permitted to perform, targeting 56 to 60% gross margin by 2030 from 52.8% in 2025. ASML stays ahead on level; Lam closes most of the distance.
Lam's structural exposure is that roughly 64% of its revenue is systems sold into a contested market at parity pricing with no customer pre-payment cushion, with deferred revenue of USD 2.43bn that fell year on year. When equipment spending turns, Lam faces volume decline and price pressure from Applied Materials, Tokyo Electron and Hitachi in the same quarter, because a customer cutting capex can also shop the remaining budget. ASML's exposure is the opposite shape. It is insulated quarter to quarter by about EUR 18.2bn of contract liabilities customers have already paid, but its revenue turns on very few decisions: one customer was 23.9% of 2025 sales, up from 16.6% a year earlier, and the two largest were 38.0%. A single customer deferring a node does more damage to ASML in one stroke than a general slowdown does to Lam.
Behind that sits the sharper exposure. Carl Zeiss SMT is ASML's sole supplier of lenses, mirrors, illuminators and collectors, manufactured only at Oberkochen and Wetzlar, and ASML's own 20-F states that if Zeiss could not maintain production over a prolonged period ASML would effectively cease to be able to conduct its business. The exclusivity that keeps competitors out also caps ASML's output at whatever Zeiss can build, which is why the 2027 capacity plan of plus 30% is a Zeiss question rather than an ASML one. Lam has no counterparty of that criticality and runs factories in seven countries that management describes as deliberately not interrelated, so one region's demand shock does not close all of them.
These six cells carry roughly 94% of Lam's revenue and 97% of ASML's; every score is argued in the tabs below. Cells are scored on position, not size, so a small cell can score 5 and a large one 2.
| Cell (product x region) | Lam | ASML | Why (one clause, sourced) |
|---|---|---|---|
| Leading-edge foundry and logic capex, Taiwan | 4 | 5 | Lam at a record Taiwan quarter, 27% of June-2026 revenue, winning 2nm and 3nm with Akara and new ALD but sharing wallet with Applied Materials and Tokyo Electron; ASML indispensable, Taiwan EUR 8,337.9m or 25.5% of 2025 sales (LRCX Q4 FY2026 call; ASML FY2025 20-F) |
| Leading-edge foundry and logic capex, United States | 4 | 4 | Lam's US share described by its CFO as consistent with its global share as TSMC, Samsung and Intel build out, supported by a USD 3bn lab and engineering programme; ASML US sales EUR 4,089.1m, 12.5% of 2025 (Goldman Communacopia, 10 Sep 2026; ASML FY2025 20-F) |
| DRAM and high-bandwidth memory capex, Korea | 4 | 5 | Lam strong across 1-alpha, 1-beta and 1-gamma nodes and owns the through-silicon via step for HBM, Korea 19% of fiscal 2026; ASML sees DRAM EUV spending compounding 15 to 25% to 2030, Korea EUR 8,159.6m or 25.0% of sales (LRCX FY2026 10-K and Q4 call; ASML FY2025 20-F) |
| NAND capex, Korea, Japan and United States | 5 | 1 | Lam dominant: the USD 40bn conversion cycle, addressable market per wafer doubling from 128 to 500-plus layers, and NAND revenue more than doubling sequentially in the June-2026 quarter; NAND is not an EUV market and ASML's memory exposure is DRAM-weighted (LRCX Q4 FY2026 call; ASML FY2025 20-F end-use table) |
| Mature and mainstream capex, China | 3 | 2 | Lam holds high share where it is permitted to sell, China 34% of fiscal 2026, but domestic Chinese customers declined in the June quarter while multinationals in China grew; ASML China fell from 36.1% to 29.1% of sales and KrF system value fell 54.5% in two years (LRCX FY2026 10-K and Q4 call; ASML FY2025 20-F) |
| Installed base service and upgrades, global | 4 | 5 | Lam larger and fast, USD 8,347.2m in fiscal 2026 and up 43% year on year in the June quarter, but its own 10-K names third-party spare-parts providers as real competition; ASML EUR 8,193.0m and up 26.2%, and no one else may service an EUV tool (LRCX FY2026 10-K and Q4 call; ASML FY2025 20-F) |
Scores are anchored to the exhibits in the three tabs; a score with no exhibit behind it does not ship. Scores are per cell and are not summed, because adding a NAND score to a China score would manufacture precision that does not exist. The three lenses get their verdicts in the Three Answers above, not here.
Normalisation first, because the two filings do not line up. Lam's fiscal year ends in late June and ASML's in December, a six-month offset that puts Lam's window half a year later in the same cycle; this is stated wherever a growth rate appears. Lam's reported split of systems against customer-support revenue maps onto ASML's split of net system sales against installed base management, and both are shown that way. On end markets Lam reports foundry, memory and logic separately while ASML reports only logic and memory, so foundry and logic-IDM are read together where the two are compared. Currency is handled by not converting anything: every ratio is a percentage of each company's own revenue, and absolute figures stay in the currency they were reported in.
The butterfly below is the whole cell analysis in one picture. Neither company discloses revenue at the level of product multiplied by geography, so five of the six rows are inferred by applying each company's disclosed product split to its disclosed geographic split. That inference assumes product mix is uniform across regions, which it is not: Lam's China revenue skews toward mature-node Reliant tools and ASML's skews toward DUV rather than EUV. The bars are therefore right about rank and roughly right about size, and the caveat matters most in the China row, where both companies' true mix is more mature than the average.
| Cell | Lam, % of total | Lam growth | ASML, % of total | ASML growth | Margin signal and leader |
|---|---|---|---|---|---|
| Service and upgrades, global | 35.9% ($8,347.2m, 10-K) | +20.2% | 25.1% (EUR 8,193.0m, 20-F) | +26.2% | ASML leads: uncontested on EUV tools. Lam faces third-party spares competition named in its own 10-K |
| Systems x China | 21.8% (inferred) | +26.0% (region, inferred) | 21.8% (inferred) | -6.6% (region, 20-F) | Lam leads on direction. Both restricted; ASML's China share fell 7.0 points in one year, KrF value down 54.5% over two |
| Systems x Taiwan | 14.1% (inferred) | +45.9% (region, inferred) | 19.1% (inferred) | +91.5% (region, 20-F) | ASML leads: EUV cannot be substituted at 2nm. Lam's Taiwan quarter was a dollar record at 27% of revenue |
| Systems x Korea | 12.2% (inferred) | +8.8% (region, inferred) | 18.7% (inferred) | +27.3% (region, 20-F) | ASML leads on DRAM EUV adoption; Lam owns the through-silicon via step for high-bandwidth memory |
| Systems x United States | 4.5% (inferred) | +26.0% (region, inferred) | 9.4% (inferred) | -9.6% (region, 20-F) | Even. Both follow TSMC, Samsung and Intel onshore capacity; ASML's 2025 dip is delivery timing, not share |
| Systems x Japan | 5.8% (inferred) | +13.4% (region, inferred) | 3.2% (inferred) | +22.9% (region, 20-F) | Lam larger here on NAND and mature-node exposure; small cell for both |
| All other regions | 5.7% | ND | 2.7% | ND | Southeast Asia, Singapore, EMEA, Netherlands; immaterial to the call |
Lam's engine is intensity, not capacity. Every move into the third dimension adds process steps that are etch and deposition steps: gate-all-around is a 3D gate that must be selectively etched and conformally deposited, backside power is a 3D interconnect, high-bandwidth memory stacks are bonded through silicon vias that Lam drills and fills, and 500-layer NAND is the same problem repeated vertically. Management quantifies it the same way each time, roughly USD 1bn of addressable market per 100,000 wafer starts per month for gate-all-around and a similar figure for backside power, which is why its share of equipment spending can rise while its share of any single application stays flat. All of it is organic; the last transaction of consequence was the Novellus combination in 2012, and the CFO stated on 10 September 2026 that large acquisitions are not part of the strategy.
ASML's engine is adoption, and it is capacity-limited rather than demand-limited. It recognised 48 EUV systems in 2025, up from 46 in 2024, and plans to add 30% to a 2026 low-numerical-aperture capacity of about 65 systems for 2027, with a further 30% under investigation for 2028. High numerical aperture is real but early: four EXE units in 2025 for EUR 1,156.9m, with more than 400,000 wafers run by customers by year end and the EXE:5200B demonstrated to full specification at a customer site. Its growth therefore depends on building more of the same thing faster, not on finding new things to sell.
Neither company has an intermediary. Both sell direct to a handful of integrated device manufacturers and foundries, so nothing sits between list price and net price except the negotiation itself, and route to market does not separate them. What separates them is how each raises realised price. ASML does it through mix without touching a list price: unit shipments fell 8.2% in 2025 while system revenue rose 11.6%, because the customer moving to the next node must buy a EUR 242m EUV tool instead of a EUR 12.8m KrF tool. Lam does it by putting more value into each new tool and then charging for it, which requires a new product every year. Lam names pricing explicitly as one of three drivers of its record June-2026 gross margin, alongside operational and scale efficiency and favourable mix, and declines to discuss like-for-like pricing, which is itself informative.
This is the one lens where Lam is clearly the safer business. ASML's optics come from Carl Zeiss SMT alone, under an exclusive arrangement, from two sites in Germany, and its own filing states that a prolonged Zeiss production failure would end ASML's ability to conduct business. ASML has bound the risk as tightly as it can, holding 24.9% of Carl Zeiss SMT Holding, funding Zeiss High-NA research and making non-interest-bearing advances against future optical column deliveries, but binding a single point of failure closer does not make it two points. Lam single-sources certain components and sub-assemblies, is qualifying alternates and carries deliberate excess inventory, and manufactures in Oregon, California, Ohio, Malaysia, Taiwan, Korea and Austria; its CEO describes those sites as not interrelated, so a demand or disruption shock in one does not propagate to all.
The market around the duel is not symmetrical. Lam's 10-K names Applied Materials as its primary competitor in dielectric and metals deposition and in etch, ASM International and Wonik IPS in atomic layer and plasma-enhanced deposition, Hitachi and Tokyo Electron in etch, and Screen, Semes and Tokyo Electron in wet clean; Lam claims leadership in conductor etch on the strength of more than 40,000 installed chambers worldwide. ASML's 20-F has no equivalent passage, because there is nobody to name in EUV. Both are share-donors in China to indigenous toolmakers, and for the same reason: export controls create a protected domestic market. Lam's filing makes the point bluntly, stating that the restrictions provide an advantage to international competitors not subject to them.
The NAND cell is Lam's sharpest. The USD 40bn conversion opportunity that has driven record upgrade revenue is now expected to be largely complete by the end of 2027, pulled forward from a multi-year expectation, and upgrades were the largest sequential contributor to Lam's record June-2026 support quarter. The offset is that conversions destroy raw wafer capacity, which management says is down about 20% from peak, so greenfield NAND should follow, but the timing gap between the two is a real hole. ASML's sharpest cell risk is Taiwan: 25.5% of 2025 sales, growing 91.5% in one year, concentrated in a customer base management does not name and in a jurisdiction whose political status its own risk factors call unique. And ASML's DUV-into-China cell is already in visible decline, with KrF system value down from EUR 2,202.5m in 2023 to EUR 1,001.3m in 2025.
Qualification lock, durability medium. Lam's own 10-K states the mechanism: once a manufacturer qualifies a tool for an application and node it generally keeps that selection for the application's life, and Lam may have difficulty selling where a competitor was qualified first. Proof of scale: more than 40,000 conductor etch chambers installed worldwide. The lock is real but cuts both ways, and has to be re-won at every node.
Accumulated process capability, durability medium to high. Forty years plus the Novellus combination, expressed today in Akara conductor etch with DirectDrive, Vantex, Lam Cryo 3.0, the ALD range, dry photoresist and panel-level advanced packaging, where 510mm by 515mm systems have shipped into development programmes. Money alone does not replicate this, but a determined competitor with time can.
Low-cost manufacturing footprint, durability medium. Seven countries, with the Malaysia shift since 2023 named by management as the structural step-up in gross margin. Replicable with capital.
Sole source for EUV, durability high. The strongest single fact in this duel is a negative: ASML's 20-F contains no competitor passage for EUV, and frames competitive risk as whether its own High-NA programme and its suppliers make technical progress, not whether a rival wins a socket. No leading-edge node can be built without it.
Supplier-layer exclusivity, durability high. Carl Zeiss SMT supplies optical columns to ASML and to nobody else, and ASML buys them from nobody else. A well-funded entrant cannot buy the optics at any price, because they are not for sale. Roughly two decades and a part-owned supplier ecosystem stand between an entrant and a working system.
Installed base that only ASML can service, durability high. Installed base management revenue of EUR 8,193.0m in 2025, up 26.2%, growing with the EUV fleet and with customer productivity upgrades, with no third-party parts market to erode it.
Concentration runs against ASML. In fiscal 2026 Lam's four largest customers were 16%, 15%, 12% and 12% of revenue, about 55% combined with no single customer above 16%, a broadening from fiscal 2025 when two customers were 17% and 15% and fiscal 2024 when one was 17%. ASML's largest customer was EUR 7,796.7m or 23.9% of 2025 sales, up sharply from 16.6% in 2024, with the two largest at 38.0%. On raw dependence, therefore, Lam is the safer of the two, and moving in the safer direction while ASML moves the other way.
Control of terms runs the other way, and decisively. ASML held EUR 18,196.2m of contract liabilities at the end of 2025, up from EUR 16,266.5m: customers financing ASML's working capital in exchange for a delivery slot. Lam's equivalent deferred revenue balance was USD 2.43bn and fell USD 250m year on year, which the 10-K attributes mainly to lower customer down payments. That single contrast is the clearest available proof of who sets the terms in each relationship. The second proof is realised price: ASML shipped fewer systems in 2025 and collected more money, while Lam's margin expansion came from a combination of pricing, plant mix and product content that it has to reassemble every year.
ASML states the position itself, in its own risk factors: Carl Zeiss SMT is the sole supplier of lenses, mirrors, illuminators, collectors and other critical optics under an exclusive arrangement, capable of producing them only at Oberkochen and Wetzlar, and if Zeiss terminated the relationship or could not maintain production over a prolonged period ASML would effectively cease to be able to conduct its business. The relationship is managed as tightly as a relationship can be, run on a stated principle of two companies and one business, with ASML holding 24.9% of Carl Zeiss SMT Holding, funding Zeiss High-NA research of EUR 22.5m in 2025 and advancing non-interest-bearing payments to secure deliveries. Lam single-sources certain components and sub-assemblies, names none as existential, is qualifying alternates, holds excess inventory against the risk, and spreads production across seven countries that its CEO describes as deliberately uncorrelated so that the same demand event does not hit all of them.
On pass-through the two are level, and this sub-lens does not separate them. Lam absorbed aluminium and steel tariff costs in fiscal 2026 and still expanded gross margin 180 basis points. ASML expanded 150 basis points in 2025 and guides to a further 120 to 320 basis points for 2026. Neither has a cost pass-through problem.
| Cell | Route control | Pocket price | Continuity | Outcome, share and margin | Confirming KPI |
|---|---|---|---|---|---|
| EUV lithography x Taiwan and Korea | Direct, both | ASML far above: about EUR 242m per system, no alternative (inferred from 20-F technology table) | ASML constrained by Zeiss capacity, not by demand | ASML share flat at effectively all of it, margin up; 2026 gross margin guide 54 to 56%. Lam is in the same fabs at a fraction of the tool value | EUV units recognised in 2026 against about 65 units of capacity |
| Etch and deposition x Taiwan, Korea and China | Direct, both | Lam at parity with Applied Materials and Tokyo Electron, won on cost of ownership and new product content | Lam's seven-country footprint is the advantage; ASML absent from this cell | Lam share up: addressable market moving from 36 to 36.5% of equipment spending toward the high 30s faster than its 2025 plan; margin up, June-2026 gross margin 52% | Lam revenue growth against wafer fab equipment growth; a third consecutive year of outperformance confirms |
| Installed base service and upgrades, global | Direct, both | Lam at parity, competing with third-party spare-parts providers named in its own 10-K; ASML above, since nobody else may service an EUV tool | Both secure | Lam grows off a larger base, USD 8.35bn and up 43% year on year in the June quarter, at contested margin; ASML grows faster in percentage terms, up 26.2%, at uncontested margin | Lam support revenue growth against installed-base growth, versus ASML installed base management growth |
Cause one, absence of a second source. Impact: major, favouring ASML. Lam faces named competitors in every market; ASML faces none in EUV. To close it Lam would have to make an entire application category single-source, which the physics of etch and deposition does not permit, because many chemistries and chamber designs can remove or add material while only one optical system currently resolves at 13.5 nanometres. Not closable at any price or in any timeframe.
Cause two, customer financing of working capital. Impact: moderate, favouring ASML. EUR 18.2bn of contract liabilities against USD 2.43bn of deferred revenue that is falling. To close it Lam would need customers to pre-fund tools they can buy elsewhere. Not closable while a second source exists, which makes this a symptom of cause one rather than an independent gap.
Cause three, breadth of the addressable opportunity. Impact: major, favouring Lam. Lam's addressable market is 36 to 36.5% of an equipment pool guided to the low USD 150bn range for 2026 and is widening through gate-all-around, backside power, 500-layer NAND and advanced packaging. Lithography is a narrower slice, and ASML's own 2030 revenue ceiling of EUR 44bn to 60bn has been reached at the bottom end in 2026. ASML cannot widen its slice by winning adjacent process steps the way Lam can, and its 20-F frames growth as more EUV in more fabs rather than as more steps. Closable only by ASML entering non-lithography equipment, which it has not signalled and which would put it against Lam, Applied Materials and Tokyo Electron on their ground.
The reconciliation here is unusually clean. ASML files a Form 20-F with US GAAP primary statements presented by function, exactly as Lam presents its 10-K, so there is no by-nature to by-function mapping to construct and no ratio has to be marked not disclosed. Three caveats are stated once and carried everywhere. Lam's fiscal 2026 ended 28 June 2026 while ASML's 2025 ended 31 December 2025, a six-month offset that places Lam's window half a year later in the same up-cycle, so every Lam column reflects a slightly hotter demand environment than the ASML column beside it. Lam reports in US dollars and ASML in euros, but all five metrics are ratios to each company's own revenue and are therefore free of exchange-rate effects; nothing here is converted. Lam's fiscal 2024 included company-itemised restructuring of USD 43.4m in cost of goods sold and USD 18.2m in operating expenses, left inside the functional lines as reported, worth about 30 basis points on that year's gross margin; there was none in fiscal 2025 or 2026 and none at ASML in any of the three years.
| % of sales, 3-year average | Lam Research | ASML | Gap | What drives it |
|---|---|---|---|---|
| Cost of goods sold | 51.2 | 48.2 | +3.0 | ASML sells a single-source tool at roughly EUR 242m; Lam sells contested tools at parity pricing |
| Research and development | 11.5 | 14.7 | -3.2 | ASML alone funds the EUV physics programme with no second source to share it; Lam spreads research across many applications on a base that grew 26% last year |
| Selling, general and administrative | 5.4 | 4.0 | +1.4 | Lam supports more than 40,000 installed chambers across seven manufacturing countries and many applications; ASML ships 535 systems a year to a handful of customers |
| Gross margin | 48.8 | 51.8 | -3.0 | The mirror of cost of goods sold, and the cleanest single reading of the power gap |
| Operating margin | 32.0 | 33.1 | -1.1 | Lam's lighter research spend almost entirely offsets its weaker gross margin |
| Lam Research, % of revenue | COGS | R&D | SG&A | Gross margin | Operating margin |
|---|---|---|---|---|---|
| Fiscal 2024 (revenue $14,905.4m) | 52.7 | 12.8 | 5.8 | 47.3 | 28.6 |
| Fiscal 2025 (revenue $18,435.6m) | 51.3 | 11.4 | 5.3 | 48.7 | 32.0 |
| Fiscal 2026 (revenue $23,232.7m) | 49.5 | 10.2 | 4.9 | 50.5 | 35.3 |
| ASML, % of revenue | Cost of sales | R&D | SG&A | Gross margin | Operating margin |
|---|---|---|---|---|---|
| 2023 (net sales EUR 27,558.5m) | 48.7 | 14.4 | 4.0 | 51.3 | 32.8 |
| 2024 (net sales EUR 28,262.9m) | 48.7 | 15.2 | 4.1 | 51.3 | 31.9 |
| 2025 (net sales EUR 32,667.3m) | 47.2 | 14.4 | 3.9 | 52.8 | 34.6 |
The most persistent gap is research and development as a percentage of sales, and it runs in Lam's favour: 14.7% for ASML against 11.5% for Lam, present in all three years, never narrowing below 2.6 points and widening to 3.8 points in the middle year. It is structural rather than a one-off, and the mechanism comes straight from the power map. ASML is the only company in the world funding extreme ultraviolet lithography, so it carries the entire cost of extending Moore's Law on one revenue line, with Zeiss High-NA funding on top of that, while Lam amortises its research across etch, deposition, clean, upgrades and packaging on a base that grew 26% last year. Strategically this says ASML buys its monopoly every year and Lam rents its position application by application, and both are paying about the right price for what they get.
The gross-margin gap runs the other way and confirms the power map rather than contradicting it: the company with no second source earns 51.8% against 48.8%, which is the expected result and would be a warning sign if it were absent. What is not expected is the direction of travel. Lam's gross margin rose 320 basis points across the three years while ASML's rose 150, so the gap narrowed from 4.0 points to 2.3, and Lam's operating margin moved 670 basis points against ASML's 180. Lam attributes its improvement to three named things: volume shifted into its low-cost Malaysia plant since 2023, new differentiated products, and pricing actions. Each has further to run, and the company has raised its long-term targets to a mid-50s gross margin and a mid-40s operating margin accordingly. On the bottom line the two are already nearly level at 32.0% against 33.1% average operating margin, and Lam's June-2026 quarter at 38.4% with a 39.5% September guide will likely erase that 1.1-point average gap within the year. The cost engine therefore agrees with the power map on rank and disagrees with it on momentum, which is the finding that keeps this duel from being a formality.
| Metric | Threshold | By when | If it hits, it favours | Where published |
|---|---|---|---|---|
| Lam addressable market as a percentage of wafer fab equipment spending | 38% or better, from 36 to 36.5% | Fiscal 2027 close, June 2027 | Lam Research | Quarterly earnings call commentary and Investor Day |
| Lam gross margin, quarterly non-GAAP | 53% or better, from 52.0% | June 2027 quarter | Lam Research | Quarterly earnings release |
| ASML EUV units recognised, calendar 2026 | 60 units or more, from 48 in 2025 | FY2026 Form 20-F, February 2027 | ASML | 20-F net system sales per technology note |
| ASML largest customer as a share of net sales | 25% or below, from 23.9% | FY2026 Form 20-F, February 2027 | ASML if at or below; Lam if above | 20-F risk factors |
| ASML 2027 revenue guidance | EUR 50bn or more | Q4 2026 results, late January 2027 | ASML if at or above; Lam if EUR 46bn or below | Quarterly results release and 6-K |