Applied Materials: America's Biggest Semiconductor Equipment Maker
Author’s note: The video is below if you want to watch it first.
Applied Materials is America's biggest semiconductor equipment manufacturer. They are a R&D leader and without the work they do, we would not have the sophisticated chips we have today.
Just as the number of foundries capable of fabbing at the leading edge is shrinking, the set of potential manufacturers capable of making leading edge chip making tools is shrinking too. The two or three who are left are some of the very best in the business.
In this video, we will continue our exploration of the chip making supply chain and do a deep dive into this unheralded but vital company.
Origins
Applied Materials was founded in 1967 by Michael A McNeilly and others, several years before the invention of the microprocessor. At its founding, the company had just five employees and $100,000 in the bank.
Back then, semiconductor companies researched and developed their own manufacturing equipment. So Applied started out by supplying the components for that equipment, most of which focusing on a type of technology called chemical vapor deposition. More on what that is later.
Applied quickly moved beyond just providing simple components. They began assembling and selling their own machines to semiconductor makers like Fairchild, IBM, TI, and Intel. They came out with some of the first commercial systems for certain deposition uses and techniques.
This came at a time of great secular growth in the industry, with many new entrants in the market. Such a gold rush meant fast growth for the companies selling the pickaxes.
Applied Materials went public in 1972, just five years after its founding. Their products had 6.5% share of the wafer fabrication market.
Downturn and Rebirth
But what comes up must come down. And the 1970s saw a brutal bear market in the entire semiconductor industry. Applied found itself in a crisis. In 1975, the company recorded a 45% decline in sales. Desperate, the company grew bloated and unwieldy due to a series of unwise acquisitions and joint ventures in the name of diversification.
Looking for new leadership, Applied's board brought in James Morgan as President and CEO. An Indiana farm boy, he had both venture capital and business operation experience.
It is interesting to note that Morgan was not a trained semiconductor engineer. But critically, he did have a reputation for turnarounds. He turned around a few businesses while working at Textron - a diversified industrial conglomerate. And he would eventually lead Applied for nearly three decades.
At the time of Morgan’s arrival, Applied had a dangerous amount of debt and risked falling into bankruptcy. To satisfy the banks, Morgan slimmed down the lineup and jettisoned several businesses.
For instance, Applied used to make their own silicon. Morgan ditched that, correctly reasoning that it would be too capital intensive to compete in.
The new Applied Materials would be ultra-focused on semiconductor manufacturing. They wanted to refocus on their businesses - mostly chemical vapor deposition but a few other technologies too - and be world class at just those. First or second place or leave it for someone else.
Chemical Vapor Deposition
It is here that I should pause and explain what chemical vapor deposition actually is. At its most general, deposition is used for placing very thin layers on things.
You need a variety of different deposition techniques to create today's modern semiconductors. The type of deposition depends on the type of material or technique used to create which structure.
Chemical vapor deposition is named as such because it relies on a chemical reaction happening between certain ingredients to deposit a thin layer of stuff - sometimes metal but could be other things - onto a substrate. So you start with a substrate.
You pump some gases - referred to as the precursors - and they react. Can be either with each other or with the substrate itself.
The result is a very thin, very evenly applied layer of a chemical on your substrate.
CVD's big advantages are that the resulting layer is pretty even across a variety of shapes, very pure, and in some cases will not require as much of a vacuum as other deposition processes.
The disadvantage of course is that you are dealing with and trying to manage a chemical reaction. With many other types of depositions, you either want to avoid reactions or keep them exceedingly simple. With CVD, a chemical reaction is required. Controlling this is challenging.
I also want to add that CVD precursors are frequently toxic, corrosive, or even explosive. Their byproducts are also often toxic, corrosive, or even explosive. But, no biggie right? The complexity of such a process partly explains how Applied attained its current position.
Corporate Strategy
At the start, Applied's customers were also its competitors. Because like I said, companies such as TI and IBM had their own internal equipment groups with their own internally developed processes.
Applied's big challenge was to prove that their stuff can do it better. Because of how long this equipment is used - years and sometimes decades - customers will naturally favor their internal equipment groups, either because of control or inertia issues. So Applied had their work cut out for them.
Applied first started gaining deep expertise in semiconductor manufacturing processes by working with customers who could not afford their own process lab. Those customers needed a way to figure out certain very complicated processes like epitaxy.
Epitaxy is a very complicated process used to grow crystal layers on a crystalline substrate. It is a variant of chemical vapor deposition - an early Applied core expertise.
Semiconductor companies struggled with the consistency and reliability of their self-assembled machinery for these processes. Applied worked extremely closely with them, developing both the equipment and the processes together at the same time. They gained critical deep expertise while doing so.
Eventually, Applied found that they could offer turn-key equipment that is both rigorously tested and representative of best practices. This naturally sold very well. And as Applied grew, the available resources they had on hand to get even better at their specialities grew too. Each new customer that Applied acquired helps further that advantage.
Manufacturers like TI eventually jettisoned their internal equipment arms because those divisions could not match the scale of capital and expertise that Applied can bring to their domain. What internal division can afford to spend billions on R&D just for chemical vapor deposition?
The dollars spent on research and development form a core part of that advantage. The ability to invest more than anyone else can afford on researching the very best of these niche processes - that is why customers choose Applied.
Does this flywheel of specialization, investment, and collective expertise sound familiar to you? It should. It's the same way TSMC grew to become the best in the world at semiconductor manufacturing. Same as TSMC split the semiconductor world into foundries and fab-less, Applied helped to split the manufacturing world into equipment makers and foundries.
Globalization
In my video about Tokyo Electron, I talked about how the Japanese semiconductor equipment maker globalized out of its original Japanese market to become the giant it is today.
Applied followed that same globalizing path. They are based in the Silicon Valley but quickly expanded abroad. In 1977, they set up Applied Materials, Europe.
In 1979, they became one of the first American companies to enter the Japanese semiconductor market. Then after that China, Singapore, Korea, and so on.
Building these customer service and account management teams locally helped make sure that the company can continue to collaborate extremely closely with their customers.
Globalization meant more than just getting into a market first. It also meant sticking with it even when things got tough. In the early 80s, a lot of Applied's competitors were pulling out of Europe. But Applied stayed with it despite a few years of losses and emerged from the downturn with a stronger market position.
Such grit helped them retain their customers and cement their loyalty. For instance, Japanese customers were especially worried about their American suppliers cutting back on support and capability every time the market turned. It took time and strong relationship building to overcome those concerns, but it paid off.
A Breakthrough Product
Going into the 1980s, Applied was a big company but they were just one of many in their niche. That all changed with a breakthrough product that would take them into the industry's top tiers: the Precision 5000.
In the late 1970s, semiconductor manufacturers started looking for a new way to perform etch operations. Features kept shrinking. The leading edge at the time was 3 micrometers. At that size, the current state of the art etching technique - high pressure plasma etching - was unreliable and introduced critical design flaws.
Applied Materials saw an opportunity for a low pressure plasma etching machine. They licensed a promising technology from Bell Labs and gathered a small team of their best dudes to rapidly develop a rough concept machine in less than a year - the 8100 Plasma Etching Machine.
The 8100 was a hit, but its hasty product development cycle led to substantial flaws. The manufacturing team was not consulted during the development process, leading to scaling issues. And installed machines ran into many troubles early on.
The next-generation 8300 was introduced to solve these issues as well as other early design flaws the 8100 had.
One of the early issues that the 8100 had was that people had to semi-manually load wafers into the machine. Since dirty, disgusting humans were involved here, it was not ultra-clean and dust particles occasionally contaminated things.
Applied saw the opportunity to integrate what their etch machine did with their current line of CVD products. To do it all in one automated, ultra-clean environment. Two machines in one. The company bet big on this market.
The Precision 5000, as it would eventually be called, has multiple chambers. One for CVD and one for plasma etching. Inside, the machine looks like a hexagon. You have the multiple ultra-clean chambers along the edges.
Then at the center you have the unprocessed wafers coming up from below on an elevator. A robot arm then picks up the individual wafers and puts them into the relevant chamber. It is definitely innovative.
Unlike the 8100 and 8300, which handled wafers in batches, the Precision 5000 handled wafers one by one: A single wafer design. Doing it this way offered more accuracy and precision while at the same time making it easier to maintain and operate.
The 5000 was a massive hit. The first of its kind, a multi-chamber, multi-process machine that actually worked well enough to be used in production.
One P5000 tool in the Oak Hill fab in Austin, Texas worked for over 20 years since its installation in 1991. Over its two decade tenure, the machine has processed over 4.4 million wafers.
The P5000 changed the industry so much so that the tool was inducted into the Smithsonian's permanent collection of Information Age Technology.
Applied Materials had never offered anything relating to etch processes before the 8100. The field had over 50 market participants when Applied first released the 8100 in 1981. But with the superior technical capability of the 8100, 8300 and the P5000, Applied eventually won the space handily and the market rapidly consolidated.
Applied Materials and Tokyo Electron
Applied Materials' most significant competitor is Japan's Tokyo Electron. The two companies are the world's biggest providers of non-lithography semiconductor manufacturing equipment.
Former CEO James Morgan said of Tokyo Electron:
> "We tended to have more innovative products. They've done a pretty good job of engineering, so that's one of the things we keep everyone focused on, that you've got to have outstandingly engineered products."
The two companies keep each other on each other's toes. So it was of great interest to the global industry when the two attempted a merger in 2013.
The resulting company - at the time dubbed Eteris - would have been worth $29 billion and be the world's largest provider of semiconductor manufacturing tools. It would hold 25% market share of the entire industry.
But to both companies' surprise, the United States government rejected the deal, forcing its cancellation in 2015.
It was not that the two companies are in the sophisticated and sensitive semiconductor space. After all, the department did approve Applied's 2011 acquisition of Varian and ASML's 2012 acquisition of Cymer.
And it was not that the two companies had a lot of overlapping products. They do not, despite both offering deposition and etch tools. And where they did overlap, relevant revenue was small and the purchasing decision does not depend on price.
Rather, the DOJ focused hard on the fact that the two companies were the industry's current R&D leaders. Because of their special - one can say unique - expertise and immense resources, they were the companies most likely to create the next generation of chip making tools.
In other words, the US Government felt the merger would have killed *future* competition in the semiconductor tool industry. It's like that movie Terminator.
This view was echoed by the chip makers, who pushed hard against the deal closing. But it is interesting to read this in the context of why Tokyo Electron agreed to a deal - because they thought their ability to survive on their own in the industry was limited.
The Company Today
Anyway, let us take a look at Applied Materials today.
In 2020, Applied made $17 billion in revenue and some $4.1 billion in income before taxes.
This revenue comes out of three segments: Semiconductor Systems, Applied Global Services, and Display and Adjacent Markets.
The Semiconductor Systems segment sells expensive chip making tools for etching, deposition, rapid thermal processing, and more. This is the largest segment by far, making up over 60% of total sales.
Relevant product groups in this segment are the Endura - which can do Physical Vapor Deposition and Chemical Vapor Deposition.
And the Centura, a versatile platform of Chemical Vapor Deposition, etch, and rapid thermal processing. That last one is a process for modifying the properties of deposited films.
The Applied Global Services segment sells solutions for optimizing fab performance. I would say this is more like customer service and product management. They provide spares, upgrades, and automation software. ASML has something similar.
The Display and Adjacent Markets segment provides tools for making TFT-LCD and OLED screens. This might surprise you, but these are semiconductors too and the processes used to make them are quite similar to the work done for microprocessors too. With that being said, it is their smallest segment.
Like with ASML and other semiconductor suppliers, the company sells to a very limited group of customers. Samsung and TSMC provided 36% of their total 2020 sales. In 2019, TSMC and Intel made up 26%.
The company spent over $2.2 billion in 2020 on research, development and engineering. The absolute amount might not seem particularly high compared to other big spenders like Alphabet, Apple, or Huawei - who spend over $10 billion each year.
But considering how focused Applied is, it is quite substantial. The company spends over 12% of their total revenue on R&D.
In comparison, Huawei spends 15.9% of total 2020 revenue on R&D. This is a lot of money relatively and emphasizes just how vital it is for their competitive position.
Conclusion
Today, Applied Materials is valued at over $125 billion. It is about the same size as Sony, Goldman Sachs, and IBM. The company has ridden the same broad macroeconomic wave that has lifted the fortunes of every other chip company.
I think the thing that really struck me is how companies like Applied have been able to leverage their scale and resources to gain extremely strong competitive edges in the industry. The huge amount of R&D spend consistently appears in each of the national champion-type companies that I have profiled on this channel. Companies like Hikvision, LG Energy, and CATL.
Yet at the same time, it is also seems clear to me that management has mixed feelings about this R&D spending. In my video about Broadcom and Qualcomm, I talked about how slashing expenses like R&D leads to jumps in profitability and a healthier stock. But at the possible cost of losing competitive advantage in the future.
It is a precarious tightrope between the short and long term. And I can see why some companies will try to leverage a merger or private equity style cost cutting to make the balancing act easier. But it seems like the government's actions in recent years has made it clear that they want those companies to stay on that tightrope, and to keep inching the industry forward.