Most people actually have a lot of myths about the semiconductor industry. Here are some of the most common and not so correct concepts:
Hardware only has one-time income
Hardware cannot be recycled and reused
Because I did mention in section 3-2 of my book “The Rules of Super Growth Stocks Investing” that hardware has only one-time income, which is more risky than the software industry. Moreover, semiconductor investment, especially the manufacture of chips, is starting at tens of billions of dollars at the current price.
The field of hardware is huge
But everyone can’t not see the wood for the trees. The field of the hardware industry is too big: terminal electronic devices, chips, household appliances… and countless industries are all areas of hardware. The saying “hardware only has one-time income” is still hold true, and chips are no different. The question is whether the products of individual companies are competitive, if there is a strong moat, the substitution is low; orders will still be endless, because customers have few options. All devices is composit of semiconductor investment.
Inventory pressure on hardware
There are many kinds of hardware
This is very similar to the above-mentioned “hardware only has one-time income.” The reason is simple, most of semiconductor investment will be phased out soon. Both are the natural fate of the hardware industry and cannot be avoided. But most people have forgotten one thing, the products of competitive companies are usually in short supply, and customers are chasing you. This type of company usually has the right to set prices and can also control the amount of shipments. This is Buffett’s favorite type of company-please see my list in my book “The Rules of Super Growth Stocks Investing”, section 2-1 illustrated.
Semiconductor companies have power to raise prices
The best example is that in August 2021, TSMC (ticker: TSM) took advantage of the global semiconductor chip shortage and announced an immediate 20% price increase for all products. Do any customers dare to say that I’m looking for another home? No, because the customer has no choice.
Another example is Broadcom’s (ticker: AVGO) in the second quarter of 2021. During the quarterly earning conference on revenue release, some analysts questioned that Broadcom’s business growth seemed to be less than expected when global chips were out of stock; To everyone’s surprise, Broadcom CEO Tan Hock Eeng retorted brazenly: “We did it deliberately.” Why did Tan Hock Eeng dare to say that, because the supply is controlled by Broadcom. To put it bluntly, customers need Broadcom and want Broadcom.
Chip designer has no factory
Foundry was formed in recent years
Becasue capital expense is huge for semiconductor investment, most chip designers have a semiconductor manufacturing plant. However, the current semiconductor supply chain is formed by the development of the last thirty years. Twenty or thirty years ago, almost all chip vendors named after the name had their own chip factories, unless the company was very small.
Fewer companies have wafer fabs
In addition to the cost of semiconductor investment, now the situation is the other way around. Only a few chip designers have their own wafer factories–Intel (ticker: INTC) and Texas Instruments (ticker: TXN) basically make their own chips. Many special-purpose semiconductor companies usually have their own factories, such as Analog Devices (US stock code: ADI), ON Semiconductor (US stock code: ON), Infineon (US stock code: IFNNY), Bosch (Bosch), STMicroelectronics (US: STM). Memory makers such as Micron and Samsung have also been producing their own chips.Companies with fabs make a lot of money
Also, most of the semiconductor vendors, which usually have their own semiconductor factories, are not as well-known as Intel or Texas Instruments, but have wide moats.
Chips must use most advanced manufacturing processes
Very few chips require advanced processes
Critical and well-known chips must use the latest and most advanced manufacturing processes: this is a very wrong concept, commercial mobile phones and high-speed transmission chips, because the consumer market is fiercely competitive, and the chip iterations are rapid, this sentence is true. However, in most fields except mobile phones and high-speed computing (i.e., CPU and GPU for computer), this is not the case at all ( the main central processing units of mobile phones and high-speed computing require advanced manufacturing processes based on many factors, other chips do not need to kill chickens at all. The device has hundreds or thousands of chips at least).
Features are what customers want
But for most semiconductor market, in the defense industry, general electronic devices, automobiles, analogs and other fields; stable functions, supply, product reliability are more than everything, coupled with cost or other various considerations (there are many factors to consider), seldom use advanced instead, they will use very mature (outdated in the eyes of ordinary people) processes that have been used for many years– For example, automobile chips are not using the latest manufacturing process, because it is not necessary, and the function is much more important than the manufacturing process. The new manufacturing process is very expensive and not cost-effective at all.
Most of the chips we use are mature processes
In short, don’t be brainwashed by the misconception that “the vast majority” of chips in the world use mature manufacturing processes. What most people think of as common sense has no basis at all.
Please refer to my previous article “Where are the Chips Used“, which lists where the chips produced around the world are used.
Only chip use advanced process make money
Only the chip designers who use advanced process chips can make a lot of money: this is not right, all of Texas Instruments’ chips are produced by its own chip factory, and many of the processes used are manufacturing techniques invented 10 years ago, but Texas Instruments is very profitable. The net profit margin of Texas Instruments is the highest among all large chip vendors, even better than TSMC.
I have compiled a detailed table in Sections 2-4 of the book “Super Growth Stock Investing Rules” comparing all major semiconductor companies and large U.S. stock companies. Among the super-large listed companies, Texas Instruments only loses to Visa (tickers: V) and Mastercard (ticker: MA), which is higher than that of pure software companies such as Meta (ticker: META) and Microsoft (ticker: MSFT). Much more (for details, please refer to the table I put together for you in Sections 2-4 of the book “The Rules of Investing in Super Growth Stocks”).
Semiconductors investment is unbeatable
Semiconductors are important
Semiconductors and chips design are strategic industries, and semiconductor investment in this field is teflon or foolproof: No one can deny the first half of this sentence in recent years. But investors have forgotten that the competitiveness of no industry or company can last for a long time. As far as the semiconductor industry is concerned, the chip is a typical business cycle industry, especially general-purpose memory and panels (panels are also a member of the semiconductor industry). I have analyzed this for you in section 4-4 of the book “The Rules of Super Growth Stocks Investing”.
Semiconductors are bulk commodities
Thanks to the COVID-19 epidemic and the impact of Trump’s embargo on China, the semiconductor industry has been out for more than three years. again in the future. The semiconductor industry has gone long for three years, and as countries are building factories and increasing production, there will soon be overcapacity because semiconductor investment increased recently. Semiconductors are a modern commodity; this has happened repeatedly before and will happen in the future. Bulk materials will definitely be greatly affected by the economic cycle.
Readers get confused by semiconductor investment
What is the reason why these leading brands in the market have been sprinting for advanced processes? Should there be market demand, will it continue to invest?
There are market demands, and the crux of them is “whether or not to use” high-end processes.
Products suitable for advanced manufacturing processes
The CPU on a mobile phone often has more than 10 billion transistors (Apple’s M1 has 16 billion, A15 has 15 billion), if instead of 5nm, the 28nm process is used (I am assuming), then the entire iPhone 13 mobile phone box will not be able to hold this one. CPU, iPhone 13 will be as big as iPad. Because each major process evolution will add another 25-40% (depending on the company, and many different factors) that can accommodate transistors in the same space.
As a result, a gigantic battery required as well, which makes the phone bigger and heavier. But why is M1 so complicated? Because of functional needs, it is reasonable for users to expect that the main functions of each generation of mobile phones should be improved compared to the previous generation, otherwise no one will have to change mobile phones. Each generation of functional improvements means that these new functions must be realized by the number of newly added transistors.
The same is true for computer CPUs and GPUs. Because of the intensive competition, especially the space and battery requirements of laptops and mobile phones, manufacturers must adopt advanced manufacturing processes. Otherwise, your laptops will be as big as TV.
This is why TSMC’s advanced manufacturing processes are occupied by mobile phone manufacturers such as Apple (ticker: AAPL) and Qualcomm (ticker: QCOM), or Nvidia (ticker: NVDA) and AMD (ticker: AMD), which requires high-speed computing computer CPU and GPU.
Products not suitable for advanced manufacturing processes
However, the iPhone and laptop are high-priced products, and form factor size is small, so it is reasonable to “need to use” advanced process. But the TV, refrigerator, and car in your home have the same motherboard, as well as countless chips and CPUs (they are called MCU on these devices). If the main chip above uses a 5nm process, of course, the space can be reduced and the power consumption can be reduced. However, the functions of the chips on TVs, refrigerators, and automobiles are actually not much different from the functions of ten years ago and now. No upgrade required.
Also, these things don’t have battery charging issues. And they have a lot of space. TVs, refrigerators, and chips on cars that spend a lot of money on high-end processes are wasteful, and they will increase the cost of chips several times, which is not worthwhile at all. We don’t need to upgrade TVs, refrigerators, cars every year; and the functions of these things have not changed for decades, and the chip does not deserve to be redesigned.
Samsung itself makes TVs, refrigerators, and many chips for automakers, but these types of chips were never intended to be produced using advanced processes. Because the chips in TVs, refrigerators, and cars are made of 5nm or 28nm, users don’t feel the difference, but if they pay 10% more, they don’t want to, because the functions are the same, they pay more 10% is unreasonable.
A matter of semiconductor investment
Keep in mind — semiconductor investment is very expensive, it’s a matter of cost at all.
- The content of this site is the author’s personal opinions and is for reference only. I am not responsible for the correctness, opinions, and immediacy of the content and information of the article. Readers must make their own judgments.
- I shall not be liable for any damages or other legal liabilities for the direct or indirect losses caused by the readers’ direct or indirect reliance on and reference to the information on this site, or all the responsibilities arising therefrom, as a result of any investment behavior.