Dear Readers,

Welcome to this week's Byte.

Every AI headline is about the chips - Nvidia's processors, TSMC's factories, data centers the size of small towns. But nobody in those headlines asks the prior question: who makes the machines that make the chips?

A modern chip factory (a "fab") costs $20 billion or more, and most of that money does not buy the building. It buys equipment - hundreds of extraordinarily precise machines that lay down materials one atomic layer at a time, carve patterns finer than a virus, and inspect every wafer for defects. The industry calls this spending WFE, wafer fab equipment - the total spent each year on the machines that build chips.

By now, you probably know me well enough to know that I'm always looking for quality compounders that have the potential to become multi-baggers over time - and that I have a soft spot for monopolies and oligopolies, the toll booths of an industry. WFE is a toll-booth industry. Three American companies - Applied Materials (AMAT), KLA Corporation (KLAC) and Lam Research (LRCX) - stand at its center, and each rules a different stretch of the road.

Today I will do something different: compare three companies in one Byte, side by side - what each one does, what protects it, whose balance sheet is strongest, and what each stock is worth on cautious math. And I will tell you which one I think is best - and why, for now, I am not buying any of them.


So, What Do These Companies Actually Do?

Think of building a chip as constructing a skyscraper with dozens of floors, where each floor must be built, carved, and inspected with near-perfect precision. Three kinds of work repeat, floor after floor:

·      Deposition - laying down ultra-thin layers of material, the concrete of each floor.

·      Etch - carving the patterns into those layers, cutting the doorways and wiring channels.

·      Process control - inspecting every floor for defects before the next one goes up, because a flaw buried on floor three ruins the whole building by floor thirty.

Applied Materials is the general contractor's supply house: it sells equipment for nearly every step - deposition, etch, inspection, and the advanced packaging that now stitches finished chips together - plus equipment for making displays. It is the largest of the three, with about $31 billion of revenue over the last twelve months and a market cap of ~$310B.

Lam Research is the specialist in etch and deposition, and above all in memory chips - the DRAM and NAND that store data. When memory makers spend, Lam feels it first and most. Revenue over the last twelve months: about $26 billion with ~$299B in market cap.

KLA Corp does one thing: process control. It makes the inspection and measurement tools that find defects - think of it as the building inspector with X-ray eyes, except fabs cannot legally or practically ship good chips without passing its inspection. It is the smallest of the three, about $14.5 billion of revenue with a market cap of ~$206B and, as we will see, the most profitable per dollar of sales.

The setting matters. Both KLA's and Lam's managements told investors in the last two weeks that WFE spending should reach roughly the mid-$150 billion in 2026 and exceed $200 billion in 2027 - the industry's biggest step-up in years, driven by AI. The toll booths are about to see a lot more traffic.


What Moats Do They Have?

When I evaluate a company, I don't just want to know whether revenue is growing. I want to understand what protects it from competitors. These three have moats of the same family - but not the same strength.

1. KLA - the near-monopoly

Process control is the deepest moat of the three. KLA holds the majority of its market, and at the most advanced chip nodes (the smallest, newest generations of chips) there is effectively no second source. A fab that is producing defective wafers loses millions of dollars a day; it does not experiment with a cheaper inspector. Decades of accumulated know-how about how chips fail, an installed base of machines in every major fab, and service engineers stationed around the world make this position self-reinforcing: the more wafers KLA inspects, the more failure data it owns, and the better its tools get. The financial fingerprint of a monopoly is in its margins: a gross margin of about 61–63% (of every $100 of sales, about $62 remains after the direct cost of making its products - a figure most hardware companies can only dream of) and, last quarter, a record operating margin of 46.1%.

2. Lam Research - the memory stronghold

Lam's etch and deposition tools are woven into how memory chips are physically built - modern memory stacks over a hundred layers, and each layer must be etched with extreme precision. Once a chipmaker has qualified Lam's tools for a process (qualification takes months of testing), switching suppliers means re-proving the entire recipe. Lam's second moat is its installed base: thousands of machines already in fabs generate a growing services business - spare parts, upgrades, maintenance - that keeps paying even in years when nobody builds a new fab.

3. Applied Materials - breadth as a moat

AMAT's protection is different: it is everywhere. It sells more kinds of tools than anyone, services the largest installed base in the industry, and increasingly sells outcomes - long-term service agreements - rather than just machines. A fab can buy its whole floor from one supplier who knows how the pieces interact. That breadth is a real advantage, but it is a softer moat than KLA's: in most of its categories AMAT faces a strong rival (Lam in etch, Tokyo Electron in several steps), and its sideline in display equipment ties part of the company to a slower, more cyclical market.

The chart below shows KLAC as the clear winner, as its operating margin sits significantly higher than the other two:


Now Let's Look at the Numbers

All three reported into the teeth of last week's chip selloff, and the stocks of all three sit well below their recent averages. Prices below are from Monday, November 2.


*KLA split its stock 10-for-1 in June 2026, so its per-share price is not comparable to the others - ratios are.

(Free cash flow is the cash a business keeps after paying all its bills and investments; return on invested capital measures how much profit a company earns on each dollar invested in it - above 15% is excellent, and all three clear it easily.)

Two things stand out. First, all three balance sheets are close to spotless. Lam's is the strongest - $3.7 billion more cash than debt, and its entire debt equals about seven months of its free cash flow. KLA is the only one with slightly more debt than cash, and even that is mostly an accounting shadow of decades of buybacks shrinking its book equity; on cash flow, it is nearly as clean as the others. When I test leverage, this is the pair I use: can the company service its debt, and how fast could it repay it? All three pass both tests with room to spare.

Second, the growth records are not equally steady. These are cyclical companies - their customers' spending swings - and honesty about that is the whole game here. Lam's revenue fell 14.5% in its fiscal 2024. KLA's fell 6.5% the same year. AMAT never fell, but it spent three recent years nearly flat (+2.8%, +2.5%, +4.4%). None of them is a straight line, and anyone who buys them must accept the down years as part of the ticket.

The recent prints, though, show the cycle's up-leg in full force:

KLA reported on October 28: revenue of $4.14 billion, up 29% from a year ago, record operating margins, and it raised its December-quarter guide to about $4.7 billion.

Lam reported on October 21: revenue of $8.22 billion, up 54%, record margins and record free cash flow of $3.24 billion in a single quarter, up 103% YOY, and guided December to about $8.6 billion in revenue.

AMAT reports next, on November 12 - analysts polled by S&P Global expect its just-ended fiscal year to show revenue of about $34 billion, up roughly 21%, with earnings per share of about $12.79. Its print is the one still to be proven.

Management quality is high across the board - all three are run by long-tenured insiders (Lam's Tim Archer since 2012, AMAT's Gary Dickerson since 2013, KLA's Rick Wallace since 2006) who return cash relentlessly: dividend growth streaks of 11, 8 and 16 years respectively, plus steady buybacks that shrink the share count a little every year.


Forecast and Valuation

Now, let's look at what the numbers could tell us about each stock's potential from here. A great business can still be a bad investment if you pay too much for it.

Analysts polled by S&P Global are enthusiastic about all three: average price targets of roughly $628 for AMAT (+63%), $218 for KLA (+39%), and $364 for Lam (+56%). Treat those as the optimistic case - they are measured off selloff-depressed prices, and they assume the 2027 equipment boom arrives on schedule.

My own exercise is deliberately more cautious, and I run all three companies through the same yardstick so the comparison is fair: I take analysts' expected earnings per share, grow them at a haircut rate - slower than analysts expect - out to fiscal 2029 in every case, and apply a price-to-earnings multiple (what the market might pay per dollar of yearly profit) a notch below what these stocks command today. (Their fiscal years end in different months - June for KLA and Lam, October for Applied Materials - but fiscal 2029 is the same destination for all three.)

Here are all three, valued on the same fiscal-2029 basis:

This is not a prediction. It is simply a valuation exercise based on a set of assumptions - and with equipment companies, the biggest assumption of all is the cycle, which none of us controls.


What Could Go Wrong?

Like every investment, these come with real risks.

The cycle turns. WFE spending has fallen hard before - in 2019 and again in 2023 - and equipment stocks fall harder. My valuation exercises above all assume 2027 spending above $200 billion. If AI spending disappoints, that number shrinks, and so does every estimate in this Byte.

China. Lam derived about 35% of last quarter's revenue from China, and management expects that share to fall sharply. Export controls on advanced chip equipment are a live policy lever in Washington; each of the three has meaningful China exposure, and controls can tighten faster than product plans can adjust.

It is one bet, not three. These companies rise and fall on the same industry spending line. Owning all three is not diversification - it is one position in three jerseys. KLA adds a further twist: you pay the richest price of the three (about 25 times next year's expected earnings) precisely because its monopoly is the safest. Safety, bought dearly, can still disappoint.

The AI-financing question. The equipment boom is funded by the enormous borrowing now flowing into data centers. If the financing tightens - as markets worried loudly just last week - fab projects get deferred, and equipment orders are the first thing deferred.


My Investment Thesis

So, which is the best of the best?

KLA is the best business. The purest toll booth: a near-monopoly on the inspection step no fab can skip, the highest margins in the industry, an asset-light model, and the steadiest hand at the top. If I could own only one of these franchises forever, it would be KLA.

Lam Research is the best value today. The strongest balance sheet of the three, the fastest current growth, the highest return on invested capital - priced, after the selloff, a notch cheaper than KLA. What you accept in exchange is the deepest cyclicality: memory spending giveth, and memory spending taketh away.

Applied Materials is third - for now. Nothing is wrong with its balance sheet or its management. But its growth record is the flattest of the three, its moat the broadest but softest, and my cautious math cannot make today's price work. Its November 12 report is its chance to change that ranking - acceleration, proven in print, is what it needs to show.

And now the honest footnote. I am not buying any of the three today. Not because they fail my tests - they pass nearly all of them - but because of my own portfolio: between the semiconductor names and funds I already hold, my exposure to this single industry cycle is full, and two of my AI-infrastructure positions are currently underwater, which is a reminder I did not need that this cycle's down-legs are not gentle. Wanting a stock and buying it are different decisions, and position size is the part of investing I actually control. These three will keep their toll booths. If my exposure ever has room - or the cycle offers one of its periodic panics - I know exactly which booths I would visit first, and in what order.

But if you already hold one of these stocks, the entry question is not the one this selloff asks you. Yours is a holder's question:

Has anything about the business changed - or only the price? And if only the price, would I be adding because the case got better, or simply because the stock got cheaper?

A falling price makes a position cheaper. It does not make the company better, and it does not, by itself, improve the investment. The honest reasons to add are the ones this Byte has walked through: the moat is intact, the balance sheet is intact, and the latest results confirmed the story instead of breaking it - which is exactly what KLA and Lam both did in the last two weeks - and the cautious math now offers a better balance of upside to downside than it did before the fall. Notice what is not on that list: "it has fallen a lot." Lam, for instance, is down nearly half from its June high - and that fact, by itself, tells you nothing about whether its business changed.

That distinction matters most with cyclical companies like these. A selloff in an equipment name is sometimes just weather - the whole industry marked down on fears that have nothing to do with any one company's orders. And sometimes it is the first mile of a downcycle, in a business where revenue fell 6.5% (KLA) and 14.5% (Lam) in the last one. The price chart cannot tell you which kind of selloff you are in. The business evidence can: orders, guidance, margins, cash flow. Add when the evidence says the business is fine and the price finally pays you for the cycle's risk - not merely because red ink makes a holding feel like a bargain.

That, for a holder, is where the difference between a great company and a great investment quietly lives.


With that, I'll rest my case. I hope this Byte has given you a clearer view of the companies that build the builders of the AI age - what you actually own if you hold one of them, and a way to think about your own position the next time the whole industry sells off.

As always, the thoughts shared in this Byte are meant to help you think more clearly about investing, not to tell you what to buy or sell. Every investor's situation is unique, so please do your own due diligence before making any investment decisions.

And I shall see you next Monday.

Until then…

Invest Smart with Pooja :)