Lam Research Corp. (LRCX)

Lam Research manufactures the machines that carve intricate patterns into silicon wafers during the fabrication of computer chips. These are not generic tools; they are highly specialized instruments that embody decades of physics, engineering, and process innovation. The company sits at a critical choke point in the global semiconductor supply chain — without equipment like Lam’s, fabs cannot produce the chips that power everything from phones to data centres to military systems. Its customers are the world’s largest chipmakers: Samsung, TSMC, Intel, Micron, and SK Hynix.

Lam is sometimes called a semiconductor-equipment play, a term that misses the reality. It is closer to a infrastructure business. Chipmakers cannot build fabs without capital equipment; they refresh that equipment every few years as process nodes shrink and yields improve. The result is a recurring, lumpy stream of orders that follows the semiconductor cycle — boom and bust tied to demand for chips themselves. In boom years, foundries expand capacity and Lam’s backlog swells; in downturns, they cut spending and Lam’s revenue compresses. Because the equipment is extremely capital-intensive and sold only to a handful of large customers, Lam’s fortunes are tightly bound to the health of the overall chip industry.

What does Lam actually make?

Lam’s product line is organised around the manufacturing steps chipmakers must perform to carve a chip. The company’s largest segment is called etch, which uses plasma — ionised gas — to remove material from the wafer surface in precise patterns. Etch tools are the workhorses of the fab; a modern foundry might have dozens of them running in parallel, each one producing wafers worth millions of dollars when finished.

The second major segment is deposition — laying down new layers of material (metals, dielectrics, semiconductors) onto the wafer. The third is cleaning — removing byproducts and impurities between process steps. A fourth emerging area is patterning — more exotic techniques like extreme ultraviolet lithography that push the boundaries of what’s possible on silicon.

All of these are sold as complete systems — the equipment plus the software, the spare parts, the installation, the training, and the ongoing service contracts that keep the tools running efficiently. A single piece of Lam equipment can cost tens of millions of dollars. A single order for a large fab might run into hundreds of millions. The contracts typically include five to ten years of spare-parts and maintenance revenue, which is recurring and high-margin.

Capital intensity and the cycle

The semiconductor capital-equipment industry is not forgiving to weak competitors. A fab costs $10–20 billion to build and equip. A chipmaker will not risk that scale of investment on an untested supplier; they require years of proven reliability, extensive on-site support, and deep technical partnership. This creates an extremely high barrier to entry and typically only a handful of vendors can serve each major manufacturing step.

Lam dominates etch so thoroughly that it serves essentially all of the world’s advanced foundries. Its margin on equipment is strong — typically in the mid-40s percent before accounting for share-based compensation and R&D. But the business is cyclical. When chipmakers are confident about future demand, they order aggressively and Lam’s backlog can reach $20 billion or more. When demand softens, orders dry up. Revenue swings can be violent: the company has experienced multiple cycles where a banner year for growth was followed by a severe contraction.

The cycle is exogenous — Lam does not control it. It is driven by broader industry events: memory-chip oversupply, changes in smartphone demand, recessions, geopolitical disruption (like the China export restrictions that have tightened in recent years), and the rhythm of new technology nodes rolling into production. Investors in Lam are not buying a predictable, stable business; they are betting on when the next cycle will turn and how high the company can grow when it does.

Execution and the technical moat

What makes Lam worth owning is not the cyclicality but the franchise inside it. The company has spent decades building tools that chipmakers trust with billion-dollar fabs. That trust is not easily displaced. A new competitor would need to invest $1–2 billion just to develop and prove a new etch system, and even then, a fab would require years of testing before deploying one to a production line worth tens of billions of dollars.

Lam’s engineers have also become deeply integrated with its customers’ roadmaps. The company’s R&D team works alongside fab engineers to solve the physics problems that come with shrinking features to ever-smaller dimensions. As nodes move from 7 nanometers to 5 nanometers to 3 nanometers and below, each transition requires new etch chemistries, new plasma designs, new materials. Lam invests heavily in these next-generation processes before customers even order the tools — a bet that the innovation will prove necessary and that Lam’s solutions will be essential to making that technology work.

The other anchor of Lam’s competitive position is installed base — the thousands of tools already operating in fabs around the world. Every tool in use is a potential customer for upgrades, spare parts, and service. Fabs prefer to refresh their Lam tools with newer Lam tools rather than rip out equipment and start over with a competitor. The installed base is also a moat: every year that Lam tools dominate a foundry, switching costs compound.

Growth levers beyond the cycle

Lam’s long-term narrative depends on several secular tailwinds offsetting the cyclicality. The first is process complexity. As features shrink, the physics becomes harder. Etch techniques that worked at 28 nanometers fail at 3 nanometers. The need to etch through multiple layers, around geometric features, and with sub-angstrom precision drives demand for smarter, more capable equipment. Lam’s tools become more sophisticated — and more valuable.

The second is capacity intensity. New applications — artificial intelligence, automotive chips, edge computing — require millions of wafers produced every year. To feed that demand, the world needs more fabs, and more equipment per fab. A single data-centre AI chip may require as many manufacturing steps as a smartphone chip, and each step needs an etch tool. As AI inference spreads, the number of advanced chips being manufactured grows.

The third is geographic concentration risk and geopolitical response. The world’s most advanced chip manufacturing is concentrated in Taiwan (TSMC), South Korea (Samsung, SK Hynix), and a handful of other locations. Governments in the United States and Europe are now subsidising new fabs to reduce this concentration — the CHIPS Act in the US, billions in European subsidies. Every new fab built in a new geography buys Lam equipment. This is a multi-year programme that should provide a revenue boost in the coming decade, though timing is uncertain.

Risks and pressures

The most obvious risk is downturn. Chipmakers are disciplined about capex and will cut orders sharply if demand for chips soften. Lam has proven it can weather these cycles, but shareholders endure severe drawdowns in revenue and earnings. The company must manage its cost structure carefully during downturns to preserve the scale it will need when the cycle turns.

A second risk is China. Historically, Lam sold significant quantities of equipment to Chinese chipmakers and contract manufacturers. Tightening US export controls on advanced semiconductor tools have curtailed this market. Lam lost access to leading-edge sales in China, compressing the addressable market. This is unlikely to reverse in the medium term and represents a structural headwind.

A third, longer-term risk is technology disruption. Scaling silicon has accelerated for fifty years, but it is slowing. At some point — whether at 2 nanometers or 1 nanometer or below — conventional silicon may hit a physical wall. The industry may shift to alternative materials (gallium nitride, silicon carbide) or entirely new computing paradigms. If scaling stops or radically slows, the need for cutting-edge etch equipment could plateau. Lam’s business assumes that the march toward ever-smaller features continues for years. If it doesn’t, Lam’s growth model breaks.

Investing and tracking Lam

Lam’s stock price is tightly coupled to expectations about the next cycle. When analysts predict a cap-equipment recovery, the shares rally; when they predict a slowdown, they decline. Understanding where Lam trades requires reading the company’s quarterly earnings — particularly guidance on future bookings and backlog — and tracking the semiconductor industry pulse directly: fab utilisation rates, memory-chip pricing, the health of smartphone and cloud demand.

The 10-K filing (SEC CIK 0000707549) breaks revenue by customer and by geography, and details Lam’s order book, which is the single best leading indicator of future revenue. Investors should watch the backlog-to-revenue ratio: a high ratio suggests strong demand ahead; a declining ratio signals a cycle turn may be coming. The company also trades at a cyclical-adjusted multiple that swings with the business — cheap in downturns, expensive in booms — so any entry point should account for where Lam sits in the semiconductor cycle.