Translated from the Headlines （Toutiao 头条新闻）
Forbes published an article by George Calhoun on October 23, analyzing Sino-US relations around semiconductors, and believing that the more leverage the US has, the more bad news for Beijing.
The Center for Security and Emerging Technology once pointed out, “The United States is the leading manufacturer of electronic design automation software, and the United States and the United Kingdom are the leading manufacturers of core intellectual property rights. These interrelated chip design investments are key constraints for China. .”
China’s vulnerability in the semiconductor industry is very serious, and the root cause lies in its lack of technology in key parts of the value chain. This applies not only to the core production stage but the weakness of China’s position also extends to certain highly specialized assistive technologies. In fact, these technologies will be more difficult for China to enter. As many people are aware, Western countries have control over semiconductor technology.
Basic industry sectors, and other
In the production of integrated circuits, there are four basic industrial sectors. Design, manufacture, packaging, and semiconductor manufacturing equipment. China is weak in all these areas. Chinese companies do not have any strategic position in the high-value design, manufacturing, and equipment fields. They only have a moderate market share in the least important and most commoditized field (package), which is not enough to create a meaningful economy. lever.
Therefore, from the perspective of the semiconductor industry structure, China is strategically in a more disadvantaged position.
The missing key part
The core integrated circuit manufacturing chain is embedded in a huge ecosystem and maybe the most complex supply network in the entire global economy. It includes thousands of other companies that provide key inputs for core enterprises: silicon, specialized chemicals and gas and other raw materials, high-value services and software, specialized tools and testing instruments, and even provide specialized commodities. Technologies such as ultra-pure water and ultra-reliable electricity. Of course, like any high-tech industry, there are basic intellectual property assets that must be obtained or authorized.
However, in this network, some components are particularly important. Two auxiliary ports are essential: EDA (Electronic Design Automation) and core IP.
Companies in these areas provide critical input for the design part. They enable fabless IC companies and IDMs to build chip designs to create value for end-users in the entire digital economy, and these values will ultimately be paid for by the entire IC network.
These supporters themselves are creators of very powerful economic value, which shows their key role in the industry. In fact, the leading companies in these two fields are even more capable of creating value than the leaders in the four fields described above.
Among the four main links of the basic integrated circuit value chain, the design stage is the main creative link for the differentiation of integrated circuit functions. The value of the end-user mainly comes from the chip design. The role of manufacturing foundry and packaging is to commercialize these designs. Although this is important, the proportion of economic value generated and obtained is much smaller. This is why fabless IC companies like Nvidia or Qualcomm can account for most of the value created by the entire industry.
However, in order to achieve this goal, the design company relies on two other highlighted market segments to get very professional help: EDA and core IP.
Electronic Design Automation (EDA)
A long time ago, integrated circuits were designed and laid by hand. Now, this situation no longer exists. A modern integrated circuit may consist of billions of components. Obviously, this is too complicated for “hand” design. Design automation means that it is a set of software tools used to automate the design process, capable of managing scale and complexity.
An industry website makes a metaphor like this: Imagine the difference between designing a small house and designing a mile-high skyscraper. For skyscrapers, you need to design complex structures, electrical, plumbing, safety and environmental systems, communication and computer networks, elevators, etc., all of which must work together. This is similar to the dramatic increase in the complexity of electronic products that designers must address today. It is this complexity, a complexity facilitated by the relentless impact of Moore’s Law, that drives the demand for automation. Engineers need to verify their concepts, model and analyze their designs, and identify and solve problems before making production commitments.
Companies that provide EDA technology, such as Cadence, Synopsys, and Ansys, are not as well-known as mainline semiconductor companies in basic market segments. However, although EDA vendors may stay “in the background”, their products are critical to the success of the IC development process.
EDA companies are very powerful value creators. Their price-to-earnings ratio and price-to-book ratio are higher than any major semiconductor industry sector.
Compared with fabless IC companies, the financial market gives EDA companies a 35% stock value premium per dollar of revenue and a 25% premium per dollar of sales. Compared with other parts, the premium is obviously much larger.
Core IP: ARM is above all else
The story of the core IP can be extracted from the story of Arm.
This unusual company-Arm Company, or ARM Co., Ltd., or Arm Holding Company, embodies a clear 21st-century business model, light assets, heavy intellectual property rights, even more, amorphous than IC companies without fabs round. Headquartered in the United Kingdom, it is owned by Japanese venture capitalist SoftBank, and will soon be acquired by one of its own customers, NVIDIA. Some people may say that Arm is the key to the entire semiconductor galaxy or at least deals with “edge” applications, which is part of most digital consumer products.
Arm does not belong to any of the above four core parts. It is a design company, but it does not design chips. On the contrary, it develops and sells (licensed) intellectual property modules-computing architecture, and IC designers such as Qualcomm, Nvidia, and Intel can create their own IC products on this basis.
Arm has achieved amazing success in what it does. The company was established in 1990 with an investment of US$3 million from Apple. Now it has become a leading force in the core intellectual property field of semiconductors, with a wide range of applications. Arm almost monopolizes a number of mass-market products, including the architecture of processors used in mobile phones, smartwatches, tablets, and smart TVs. In the past 20 years, the company has integrated dozens of small design teams into one combination, which today seems almost unmatched.
-2006–In the pre-smartphone era, ARM-based processors have been used in 98% of mobile phones; at that time, the company had sold (licensed) 1 billion cores;
-2010-ARM sold nearly 2 billion cores every quarter;
-2014-A total of 50 billion chips designed by Arm were produced;
-2017-the cumulative total is 100 billion cores;
-2019-130 billion cores in total;
-2020-180 billion cores.
-2021-Their website claims that 200 billion Arm-based chips have been shipped.
This means that everyone on the planet has 25 Arm-based chips, or 70% of the world’s population is using ARM-based products.
ARM is not just for the marginal market of the masses. Its technology is also the basis of the fastest supercomputer in the world. The arm is currently targeting the server market.
When the company focused on “one trillion cores”, management saw other new markets that need to be swallowed, including the “Internet of Things” (IoT). In this typical large-scale edge market, their advantages should be obvious. In fact, for the applications it leads, the basic functional advantage of Arm is that it can provide high performance with low power consumption.
Although Intel has historically been considered a leader in chip manufacturing, Arm has slowly entered a niche market where computing devices no longer need to be faster, but they need to be more efficient and portable. This is why Arm dominates the mobile processor market, and almost every major version is built on its architecture.
In short, if you want to build a data center server, where power is not the main limiting factor, Intel’s X86 architecture is obviously the best choice, and Intel still has more than 90% of the share. However, if the goal is to design a smartphone or smartwatch or sensor or other edge products that are not connected to the Internet, where power consumption is the key limiting factor, Arm will be a clear choice, and it has more than 90% of the market share.
Value creation, what is the value of ARM?
Last year, Nvidia proposed to acquire Arm from SoftBank for US$40 billion in stocks and cash, which set a specific value point for us. In fact, since the offer was made, Nvidia’s stock price has risen, increasing the value of the stock portion of the offer. As a result, today’s transaction value is approximately 54 billion U.S. dollars.
Arm has achieved approximately $2.1 billion in revenue and earned $125 million in the past four quarters.
Something with a P/E ratio of 300-400 may not make sense now. The market price/sales ratio is more stable. It is much higher than the average level of any other market segment and is comparable to the most expensive fabless IC company Nvidia itself.
In short, these two auxiliary fields-EDA and core IP, are powerful value creation engines. The market places high premiums on the sales, earnings, and cash flow of these companies. Functionally, they enable companies doing chip design to do their jobs. In short, core IP and EDA are technologies that provide recipes and ingredients for chip manufacturers’ secret sauces.
Bad news for Beijing
As mentioned above, in the semiconductor technology-geopolitical competition, the relative strength of China and the West is very obvious.
-China’s position in high-value fields such as design, manufacturing, and manufacturing equipment is negligible.
-China’s position in the low-value packaging field is not high, and it has not brought them important leverage.
-China’s position in EDA and core IP fields is also negligible, even weaker than other fields.
To quote a report from the Center for Security and Emerging Technologies (CSET) in January 2021.
-“China has made progress in some areas, but it is difficult in other areas. China is the strongest in packaging. It is making progress in design and manufacturing, albeit with the help of state support. However, China is in EDA and The core intellectual property is struggling.”
-“The American company is the exclusive supplier of EDA software, with fabless and IDM engineers with all-round capabilities required to design cutting-edge chips. Chinese chip designers rely on American EDA tools, especially all chips from Synopsys and Cadence Design, China’s EDA companies will not be able to support the chip design of cutting-edge nodes, nor will they be able to compete with the top US EDA companies.”
Let’s talk about core intellectual property (IP)
-“China is weak in core IP development. Chinese chip designers rely heavily on non-Chinese core IP. 95% of Chinese chips use IP licensed from ARM. For example, the leading chips used in smartphones are mainly based on ARM Architecture and core IP. Huawei’s subsidiary HiSilicon, China’s top fabless design company, also relies on ARM’s core IP.”
The concept of national self-sufficiency is irrelevant in the semiconductor field. China’s ambitions in this regard are just illusory.
On the other hand, in some cases, the leverage of the state may be quite real. The US’s squeeze on Huawei by refusing to obtain US semiconductor technology is quite effective, perhaps more than many people expected.
Now, although some claims in the media are very alarmist, the United States and its allies hold most of the important cards. EDA and core IP are the two winners.