Lam Research sells etch and deposition equipment used to make memory and logic chips, so its orders follow chipmakers' capital spending. The general method is in the stock courses.
Tools Follow Wafers
Lam Research sells the etch and deposition tools that fabs use to make leading-edge and memory devices. It is a wafer-fab-equipment cyclical, not a GPU company. China export rules, NAND/DRAM capex, and foundry node transitions are three different clocks. If you mash them into “AI needs chips,” you will misread a memory-capex pause.
Size LRCX as a high-beta equipment cyclical with export-gap risk. Not NVDA-lite.
1. Etch tools and the memory cycle
Lam’s identity is process-step intensity: more 3D NAND layers, more etch steps, more dollars per wafer. That identity made the 2010s NAND buildout a fortune and the NAND gluts a drawdown. The physics did not change because a GPU shortage made memory look strategic.
China became a large slice of WFE demand and then a regulatory object. A license regime can cut a region without cutting global bit demand. Both the cut and the reroute show up in the guide. Read the 10-K geographic note like an operating manual.
Services, spares, and upgrades are the ballast when new-tool shipments freeze. Ballast is real. It is not a software multiple. Do not pay one unless mix has actually changed.
2. What the modern company sells
Systems vs customer-support revenue. Memory vs foundry/logic end markets. The supplement’s WFE commentary is the weather system. A foundry node delay and a NAND capex cut can offset or stack. Stacking is how 40% drawdowns happen.
Customer concentration among a handful of memory makers and foundries is the NVIDIA problem in tool form. One pause hurts. Two is a cycle.
Peers (Applied, Tokyo Electron, KLA, ASML-class lithography) are complements and share-takers depending on the step. Lam does not own the whole wafer. It owns specific steps that get more intense as devices get 3D. Micron is one of Lam's customers; when memory makers cut spending, Lam's orders fall too, usually with a lag.
3. Why it still compounds — and what stops it
Process know-how at etch/dep and installed-base services are the moat. The moat does not repeal a capex winter. It keeps Lam alive through one.
What stops an upcycle: memory overbuild, a foundry digestion year, or an export rule that the market had not discounted.
LRCX can print record tools and still be a sell if forward WFE is rolling. Forward WFE is the object.
4. How traders actually use the ticker
LRCX is a gap name with SOX/SMH beta. Implied moves into earnings are large. Size as equipment, not as a foundry, not as NVIDIA. LRCX reacts to memory pricing, chipmaker capex plans and export-control news on China, so size with the position size calculator for headline risk. Export rule changes have moved the whole equipment group on the stock scanner.
Illustration: $50,000 in the account, $500 at risk, a $8 stop below $105, roughly 62 shares. Write it into the written trading plan, and if you are really trading the memory cycle, the risk rules still set the size.
Pairs vs other WFE only with a relative step or China-mix view.
5. Mistakes, limits, takeaways
Mistakes: AI-slogan thesis; ignoring China mix; marrying LRCX at a WFE peak. Revenue by customer segment and by region, especially the China share, shows where the risk sits. The statements course explains these disclosures; What Is Stock Trading? covers the fundamentals.
Export rules and spot capex change often, so check the current ones before you size. Educational only.
Worth remembering
- WFE is a cycle with a 3D kicker.
- China is mix and a license regime.
- Services are the ballast that holds revenue up when new-tool shipments freeze.
- Size for a violent print.
Lam sells the tools for every chip boom and feels every capex cut first. Educational only. Not tax, legal, or a recommendation to buy, sell, or hold LRCX.