Applied Materials (AMAT) has been making semiconductor manufacturing equipment since 1967. In 2020, a pandemic year, the stock has risen 35%, using the closing price of $84.27 on December 1, 2020. Its 52-week high and intra-day high, $84.63, is a record. This article will touch on why Applied Materials has had a good run so far this year, but the focus will be on the technological inflection point that is driving this success and is highly likely to drive the stock price significantly higher over the remainder of the decade.
To help readers appreciate this thesis, I will first recap the old paradigm for growth in the semiconductor equipment manufacturing industry. Then, I will give an overview of the inflection drivers and the specific equipment Applied Materials has created to capture value during this transition. Before reaching a conclusion, I will briefly review fiscal Q4 results and the dividend, as well as the acquisition that is underway.
History of the Silicon Shrink Cycle
The first transistor was built in 1947. The first experimental integrated circuit was made in 1959. Since then, at fairly regular intervals, the practical number of transistors per IC has grown. Two factors limit the number of transistors on an IC: the size of the transistors and the size of the chip. In 1965, Gordon Moore of Intel (INTC) observed that the density of transistors was doubling about every two years. In 1986, the first RAM chip was produced capable of holding one megabit of information. Shrinks of feature size are known as process nodes, now stated in nm, or nanometers. The most advanced commercial nodes today are producing samples at 5 nm, and 7 nm chips are readily available, including AMD (AMD) CPU and graphics chips.
Not to detract from the engineering and science skills needed, but in some ways, the story of semiconductor equipment manufacturing from about 1965 to about 2015 was humdrum. The patterns made on the base die, that would form transistors, other features, and connectors, needed to shrink about every two years. For the most part, that meant shrinking patterning, which is achieved with lithography. Semiconductor manufacturing equipment and materials had to handle these ever-shrinking patterns on photoresist, then remove the appropriate parts of the masks with high precision to build the structures of the transistors and their connections. Methods changed occasionally over time, but the basic idea was always clear: shrink.
Inflection: New problems as physical thresholds for shrinkage are approached
The problems the semiconductor manufacturing industry is now confronting began to appear at earlier shrink nodes, but 7 nm node patterning difficulties began to show what will be needed going forward. These mainly have to do with the relationship of the wavelength of radiation, the size of silicon and other atoms used, and the changing properties of features at that scale. Keep in mind that a single defect may ruin an entire chip. If a chip has a large surface area and a billion or more transistors, quality control becomes a serious issue. Chips require multiple processes, but the masks for each process need to align perfectly. The lithography has to be perfect, the masks need to be perfect, the etching and deposition need to be perfect. Visible light waves are too coarse to use, but using extreme ultraviolet or EUV lithography presents its own difficulties. Traditional materials may not work well at this level of precision. Failure means low yields of usable ICs. New materials may be needed to prevent voltage leaks and keep down power requirements in the finished product. To all that, add in the need to move to 3D structures from the traditional 2D structures.
New Applied Materials technologies, materials, and equipment
These new problems have created the possibility of new solutions. Applied Materials has been researching the problems and experimenting with solutions for years now. The overall picture is that the foundries that make ICs, at least the cutting-edge ones, are going to need new equipment that goes well beyond the old Moore’s Law 2-year shrink cycle. This means higher revenues for the companies that can supply solutions.
CEO Gary Dickerson spoke to these solutions at the Applied Materials fiscal Q4 conference call. Some are already being sold to customers, others are still in labs. He has been talking about preparing for this inflection point for several years.
“As the benefits of traditional 2D Moore’s Law scaling slow down, leading companies are describing how the industry is transitioning to a new playbook to drive Performance, Power, Area-Cost and Time-to-Market of new devices. This PPACt playbook includes new architectures, new structures, new materials, new ways to shrink geometries and new packaging technology.”
Metal deposition technology is critical to the performance of interconnections. Applied Materials’ metal deposition revenues grew 42% in fiscal 2020 to $2.2 billion. Its equipment comes in three deposition types. PVD, or physical vapor deposition, sputters materials onto a chip. CVD, or chemical vapor deposition, uses chemicals in a chamber that react with the substrate of the chip. ALD, or atomic layer deposition, is a process where gases are used to deposit a single, pure layer of atoms. Since a wide variety of elements must be deposited, including silicon nitride, copper, or tungsten, and the work must be very precise, it is important to optimize the technology for the various structures required. For instance, in CVD, Applied Materials sells the Endura Volta Selective W CVD for tungsten deposition.
Etching equipment is in transition, becoming increasingly reliant on analysis of structures and materials. High-k metal gate technology is needed to drive input/output speeds for DRAM. Centris Sym3 etch is an example of new technology that addresses the need for better precision at smaller process nodes. Among other benefits, it improves conductance by removing etch materials from the work chamber so they cannot redeposit on the device structures.
In patterning, Applied Materials has a new product combining a novel mask material with an etch solution for that material. This ability to integrate solutions is a selling point for new Applied products.
Inspection is a particularly bright spot for the company. It is selling a new optical inspection system and new e-beam products that are seeing strong initial adoption. Applied Materials claims these have dramatically higher resolution than any other e-beam equipment on the market. This is critical to see data about the structures being created in order to modify the structures to get maximum yields.
Even packaging technology needs to innovate as chips become more complex. Applied Materials is already the leader in advanced wafer-level packaging. It is working on first fully integrated solution for hybrid bonding to connect two chips together in die form and enable shorter interconnects, resulting in a 4x increase in Input/Output density.
Applied Materials also sees the increased use of AI as a driver for more complex ICs, used in greater numbers, driving semiconductor manufacturing equipment sales. Notably, the current use of neural network AIs is very power-intensive. That will drive the need for improved computational performance per watt. This plays into the company’s strengths.
Fiscal Q4 2020 results
Applied Materials’ fiscal Q4 2020 ended on October 25, 2020. Revenue in the quarter was $4.69 billion, up 25% y/y. GAAP net income was $1.13 billion, or $1.23 per share. Non-GAAP net income was $1.15 billion, or $1.25 per share. Cash from operations was $1.32 billion. The company returned $200 million in cash to shareholders in the form of dividends and repurchased $50 million worth of shares. In the past, Applied Materials has devoted more cash to share repurchases, but it is currently conserving cash for the Kokusai acquisition. At the end of the quarter, the cash and equivalents balance was $5.35 billion.
The Kokusai acquisition is expected to clear regulatory hurdles by the end of 2021. Kokusai makes batch processing systems for semiconductor manufacturing. Applied Materials will pay $2.2 billion in cash.
Dividend and share buybacks
The dividend is currently $0.22 per quarter, which works out to a yield of about 1.07%. The ex-dividend date was November 18. The dividend yield used to be better, but has drifted lower due to the increase in the stock price. With a P/E ratio (TTM, GAAP) of 21, the company could afford both a higher dividend and higher stock buybacks. Likely the board will reassess this after the acquisition is completed and integrated.
Caveats and Conclusion
This article has focused on Applied Materials. There are many competitors in the industry, and considerable competition in the segments within the industry. In addition, a serious recession or depression could impact demand for the entire industry.
I believe Applied Materials is right about its being an upward reflection point in the semiconductor equipment manufacturing industry. I expect demand for chip complexity to continue to increase. That means a continued, though perhaps slower, march to smaller nodes. Given the reality of physics at such small dimensions, new materials and types of machinery will be necessary to make these shrinks work. Applied seems well-positioned to capture significant share of the markets opening up. That, in turn, should drive up revenue, profits, and the stock price.
Disclosure: I am/we are long AMAT. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.