Semiconductor wafer fab equipment (WFE). LRCX / AMAT. Run 16 Sep 2026. Built from Lam's FY2021-FY2026 10-Ks, FY2026 earnings call transcripts and September 2026 conference transcripts (Kun's LRCX sources folder), plus Applied's FY2024-FY2025 10-Ks, FY2025-FY2026 earnings releases and prepared remarks (SEC EDGAR, Applied IR). Events swept through 16 Sep 2026; most recent events checked: Lam at Goldman Sachs Communacopia (10 Sep 2026), Applied board appointment (27 Aug 2026) and dividend declaration (10 Sep 2026). Figures in USD as reported; Lam's fiscal year ends in June, Applied's in October (an 8-month offset, flagged wherever it matters). Not a valuation and not a recommendation.
Three cells carry the next five years.
Applied leads by scale in the first two cells and claims the #1 position in leading-edge logic and DRAM (Q3 FY26 prepared remarks, company claim). But its foundry/logic systems revenue grew only from $13.5B to $13.9B over FY24-25. Lam's served market, meanwhile, is rising faster than it planned at its 2025 Investor Day (Q4 FY26 call). Lam's forward risk is NAND: the $40B upgrade wave finishes by end-2027 and new NAND fabs arrive only from 2H28.
Compared over the same calendar window, the two companies' gross margins are about a point apart, but Lam's EBIT margin is 4 to 6 points higher. The lead comes from lighter operating costs: R&D was 10.2% of sales vs 12.6%, and SG&A 4.9% vs 6.2% (latest fiscal years). Lam spends its R&D on a narrower etch and deposition portfolio, while Applied funds more product lines plus the $5B EPIC R&D center. On gross margin, Lam credits pricing actions, low-cost manufacturing in Malaysia and product mix for its 52.0% June quarter, and guides to the mid-50s (Q4 FY26 call). Applied reached 50.4% in its July quarter after 13 straight quarters of year-on-year gains (Q3 FY26 release). The test of whether this is structural: at Lam's FY24 memory trough its EBIT margin was 28.6%, level with Applied's 28.9%. Today's gap is part structure, part operating leverage.
Lam: a customer and memory cliff, felt through operating leverage.
Applied: China mature-node demand plus fixed costs.
Lam breaks first because its exposure is concentrated in fewer buyers and one memory type. Applied's is spread across a broader base.
These five cells carry essentially all of both companies' revenue. Neither company reports product by region, so the regions attached to each cell are inferred from where its customers' fabs sit. Every score is argued in the tabs below.
| Cell (product x region) | Lam | Applied | Why (one clause, sourced) |
|---|---|---|---|
| Leading-edge foundry/logic, Taiwan + US | 4 | 4 | Lam foundry revenue up sharply with gate-all-around wins (Akara installed base doubling yearly); Applied claims #1 at the leading edge but grew 3% in FY25 (Q4 FY26 call; Applied FY25 10-K) |
| Installed-base services, global | 4 | 3 | CSBG $8.3B, +20% in FY26 and +43% in the June quarter; AGS $6.4B, +3% in FY25, but more than two-thirds subscription (10-Ks; Q3 FY26 remarks) |
| DRAM / HBM, Korea + Taiwan + US | 3 | 4 | Applied DRAM is 26% of a $20.8B systems base vs Lam's ~$3.3B (inferred); Lam gaining via Vector hard mask and Akara wins (Q4 FY26 call) |
| NAND upgrades, Korea + Japan + China multinationals | 5 | 2 | Lam dominant in high-aspect-ratio etch, with NAND systems revenue doubling quarter on quarter; flash is 7% of Applied systems (Q4 FY26 call; Applied FY25 10-K) |
| Mature-node foundry, China domestic | 3 | 3 | Both are fenced off from part of this market by export rules. Lam share of China revenue fell from 43% to 26% in four quarters; Applied steadier at 27-30% (earnings calls and releases) |
Scores are anchored to the exhibits in the three tabs. Scores are per cell and are not summed; the three lenses get their verdicts in the Three Answers above.
How the labels were matched.
| Cell | Lam FY26 $M (% of total) | Growth | Applied FY25 $M (% of total) | Growth | Margin signal / leader |
|---|---|---|---|---|---|
| Foundry/logic systems | ~9,130 (39%) inferred | ~+37% | 13,935 (49%) | +2.9% | Applied by scale; Semi Systems op. margin 35.5% (FY25 10-K). Lam gaining. |
| Installed-base services | 8,347 (36%) | +20.2% | 6,385 (23%) | +2.6% | Lam larger and faster; AGS op. margin 28.1%, lower than Applied's systems margin (FY25 10-K); Lam margin ND |
| DRAM / HBM systems | ~3,330 (14%) inferred | ND | 5,407 (19%) | -3.0% | Applied leads |
| NAND systems | ~2,430 (10%) inferred | ND | 1,456 (5%) | +83% | Lam leads |
| Display and other | 0 | - | 1,185 (4%) | ND | Applied only; loss-making Corporate & Other line |
| Latest quarter, total revenue | 6,720 (Jun-26) | +30% YoY | 9,115 (Jul-26) | +25% YoY | Next guide: Lam $8.1B (+52% calc) vs Applied $10.25B (+51%) |
Regions: Lam FY26 China 34%, Taiwan 22%, Korea 19%, Japan 9%, US 7%, SE Asia 6%, Europe 3% (Lam FY2026 10-K). Applied FY25 China 30%, Taiwan 24%, Korea 20%, US 11%, Japan 8%, SE Asia 4%, Europe 3% (Applied FY25 10-K). Latest-quarter China share: Lam 26% (June 2026), Applied 28% (July 2026).
Leading-edge foundry/logic. Moving from FinFET to gate-all-around roughly doubles the number of surface-treatment steps, and adds selective etch and atomic-layer deposition steps (Lam Q4 FY26 call). Lam's Akara etch platform and Argos surface-treatment tool won tool-of-record positions (the tool a chipmaker qualifies for a given process step) at 2nm. Applied cites differentiated positions in deposition, materials modification, conductor etch and eBeam, and hosts TSMC and Samsung at its EPIC R&D center (Applied Q2/Q3 FY26 remarks; EPIC release, Feb 2026). Both grow organically here; Applied's $120M NEXX acquisition (closed June 2026) is packaging and immaterial.
DRAM/HBM. Memory makers expect to add 20-25% DRAM capacity a year, but are limited by how fast cleanroom space comes online (Lam Citi, 9 Sep 2026). Applied has the bigger DRAM base; Lam's DRAM revenue is growing from a smaller base by carrying logic process know-how into DRAM. NAND. Lam says its content per NAND wafer doubles from 128 to 500+ layers, but the spend is conversions: raw wafer starts are down 20% from peak, and new fabs come from 2H28 (Citi). Advanced packaging. Both guide more than 70% growth in 2026.
Both sell direct to a handful of chipmakers, and a tool qualified for a given process step is hard to displace. Lam's 10-K says it may struggle to sell to a customer that has qualified a competitor's tool. Pricing therefore shows up in new products and upgrades rather than list increases. Lam named "pricing actions" as a driver of its 52% June-quarter gross margin and ranks new-product mix first, factory location second and direct pricing third (Goldman Sachs, 10 Sep 2026). Applied shows the same power more slowly: 13 straight quarters of year-on-year gross margin gains.
| Dimension | Lam | Applied |
|---|---|---|
| Footprint | Oregon, California, Ohio, Malaysia (second factory ramping), Taiwan, Korea, Austria (Q4 FY26 call) | "Distributed" US, Singapore, Japan, China, Korea, Taiwan, Israel, plus contract manufacturers (FY25 10-K) |
| Inventory posture | $4.3B, 3.0x turns, best in about 5 years (June 2026) | $5.9B FY25, about 148 days (about 2.5x, calc); $6.6B July 2026 |
| Single-source language | Supply chain and outsource-provider disruption risk factor (FY26 10-K) | "Some key parts" from a single qualified supplier (FY25 10-K) |
| Capacity move | Malaysia phase 2; densifying existing factories | $500M Singapore expansion; aims to double quarterly system output by 2028 |
In etch, Lam names Applied, TEL and Hitachi as its main rivals; in deposition, Applied, with ASM and Wonik IPS in ALD/PECVD (the atomic-layer and plasma-enhanced deposition methods) (Lam FY2026 10-K). Independent share data (TechInsights, Gartner) was not accessible (ND). One secondary source, The Information Network (May 2026, paywalled, unverified), says Applied and TEL lost share in 2025 while Lam gained. Lam's own claim of three straight years of outgrowing WFE points the same way.
Foundry/logic: Lam is more exposed to mature-node China orders, which fell sharply in the June quarter; Applied to TSMC's pacing, since its largest customer was 19% of FY25 revenue. DRAM: Applied is more exposed to a memory capex pause, because DRAM is its biggest growth cell. NAND: Lam is far more exposed; the conversion wave ends in 2027. Services: a utilization drop hits spares for both, but Applied's subscription contracts cushion AGS.
1. Installed base in etch. More than 40,000 conductor etch chambers are installed (Q4 FY26 call), feeding spares, upgrades and learning cycles. Durability: High.
2. High-aspect-ratio etch and 3D know-how. Built on NAND (Cryo 3.0) and now carried into gate-all-around logic and DRAM through Akara, Argos and Vector wins. Durability: High, because it takes years of process learning, not money.
3. Packaging. Leader in through-silicon-via (TSV) etch and electroplating. Durability: Medium.
1. Breadth. Positions in CMP, ion implant, epitaxy and eBeam inspection, where Lam is absent, enable co-optimized multi-step solutions (FY25 10-K). Durability: High.
2. EPIC co-development. A $5B R&D center with TSMC, Samsung, SK hynix and Micron as partners gives Applied access to customers' roadmaps early. Durability: Medium-High, since a rival could fund one but not quickly win the same partners.
3. Connected installed base. More than 37,000 chambers are connected to AIx software. Durability: Medium.
Lam: four customers were 16%, 15%, 12% and 12% of FY26 revenue (55%), up from one customer at 17% in FY24 (Lam FY2026 10-K). Its 10-K warns that large customers can negotiate "decreased pricing". Applied: two customers were 19% and 15% of FY25 revenue (34%), and more than two-thirds of service revenue is subscription (Applied FY25 10-K; Q4 FY25 remarks). Both have customers pre-committing: Lam says tool discussions now run into 2027 and beyond, and Applied's FY25 backlog was $15.0B.
Lam: factories in several regions serving different demand, which Lam says lets it "get what we need when we need it and also at the prices" (Q4 FY26 call). Tariffs weighed on gross margin in the September 2025 quarter (Q1 FY26 call). Applied: a distributed model with contract manufacturers and some single-supplier parts, and now under Commerce Department (BIS) audits and annual certifications after the February 2026 settlement. Pass-through looks similar: both lifted gross margin through the 2025 tariff period.
| Cell | Route control | Pocket price | Continuity | Outcome (share / margin) | Confirming KPI |
|---|---|---|---|---|---|
| Leading-edge foundry/logic, TW+US | Both direct; tool-of-record per step | Parity (inferred); Lam premium on new tools | Both high | Lam share up / margin up; Applied share flat / margin up | Lam foundry systems $ growth vs Applied foundry-logic $ growth, CY2027 |
| DRAM/HBM, KR+TW+US | Both direct | Parity (inferred) | Both high | Applied holds lead; Lam share up from smaller base | Lam DRAM % of systems at 25% or more |
| Installed-base services, global | Both captive to own installed base; third-party spares compete | Lam premium in upgrades; Applied subscription lock-in | Both high | Lam grows faster; Applied steadier | CSBG growth minus AGS growth, CY2027 |
The tension: Applied has the stronger overall power position but the weaker growth cells. Lam is growing fastest precisely where its customers are most concentrated, so its growth depends on a few buyers it cannot control.
1. Breadth vs depth. Applied spans the flow; Lam dominates the steps that are multiplying. Impact: Major. Neither can close the gap within five years; building an etch franchise takes a decade of tool-of-record wins, not money. 2. Customer concentration. Impact: Moderate. Lam can only dilute it as DRAM and logic customers broaden, over 2-3 years. 3. Manufacturing cost footprint. Impact: Moderate. Applied is closing this with Singapore and its 2028 output plan, over 2-3 years.
Both report costs by function under US GAAP, so the five ratios compare cleanly. Lam FY24-26 (ending June) is set against Applied FY23-25 (ending October). Applied's last year ends eight months before Lam's, so Applied's column misses most of the 2026 upswing. The dumbbells in Answer 2 show the calendar-aligned check. Treatment of one-off lines:
| % of sales, 3y avg | Lam FY24-26 | Applied FY23-25 | Gap | What drives it |
|---|---|---|---|---|
| COGS | 51.2 | 52.4 | -1.2 | Lam's Malaysia scale and pricing in FY26; the two were level in FY25 |
| R&D | 11.5 | 12.1 | -0.6 | Narrower portfolio; Lam higher in its FY24 trough |
| SG&A | 5.4 | 6.3 | -0.9 | Lighter selling model and fewer product lines; Lam lower in all 3 years |
| Gross margin | 48.8 | 47.6 | +1.2 | Mirror of COGS |
| EBIT margin | 32.0 | 29.0 | +3.0 | Operating leverage in the FY25-26 upswing |
| Lam (FY) | COGS | R&D | SG&A | GM | EBIT | Applied (FY) | COGS | R&D | SG&A | GM | EBIT |
|---|---|---|---|---|---|---|---|---|---|---|---|
| FY24 | 52.7 | 12.8 | 5.8 | 47.3 | 28.6 | FY23 | 53.3 | 11.7 | 6.1 | 46.7 | 28.9 |
| FY25 | 51.3 | 11.4 | 5.3 | 48.7 | 32.0 | FY24 | 52.5 | 11.9 | 6.6 | 47.5 | 28.9 |
| FY26 | 49.5 | 10.2 | 4.9 | 50.5 | 35.3 | FY25 | 51.3 | 12.6 | 6.2 | 48.7 | 29.2 |
Raw $M. Lam FY24/25/26: revenue 14,905 / 18,436 / 23,233; COGS 7,853 / 9,457 / 11,507; R&D 1,902 / 2,096 / 2,376; SG&A 868 / 982 / 1,150; operating income 4,264 / 5,901 / 8,200. Applied FY23/24/25: revenue 26,517 / 27,176 / 28,368; COGS 14,133 / 14,279 / 14,560; R&D 3,102 / 3,233 / 3,570; SG&A 1,628 / 1,797 / 1,768; operating income 7,654 / 7,867 / 8,289.
The most persistent gap is SG&A. Lam spends 0.3 to 1.3 points less of sales on it in every year. That fits Lam's narrower product range sold to the same few customers, and it is structural. The EBIT gap is bigger but not persistent: at Lam's FY24 memory trough its EBIT margin (28.6%) sat just below Applied's (28.9%), then widened to 6 points as volume returned. Applied's R&D intensity went up in FY25 as revenue grew, which reflects a choice to fund breadth and EPIC. The cross-check against the power map is partly contradictory. Applied's stronger overall power does not show up as a higher gross margin, and the two gross margins were identical in FY25. So breadth protects Applied's share, not its price. Lam's cost edge is real but cyclical, which supports the "Lam breaks first" answer.
| Metric | Threshold | By when | If it hits, it favors | Where published |
|---|---|---|---|---|
| Lam served market as % of WFE | 37% or higher for CY2027 | Jan 2028 | Lam | Lam earnings call / investor events |
| Non-GAAP gross margin, latest quarter | Lam 53% or more, Applied 51% or less | Jun 2027 quarter | Lam | Quarterly releases |
| Services growth: CSBG minus AGS | Lam ahead by 5 pts or more for CY2027 | Jan-Feb 2028 | Lam | Lam call; Applied release |
| Lam NAND share of systems revenue | Below 12% for 2 quarters | Through Dec 2027 | Applied | Lam earnings calls |
| Applied Semi Systems growth vs WFE | Faster than WFE in CY2026 and CY2027 | Nov 2027 | Applied | Applied Q4 releases and remarks |