Following the sharp drop yesterday, some semiconductor stocks in A shares fell sharply today, with a drop of more than 30% in two days.
At the same time, the global semiconductor stock crash continues.
Today, TSMC set a record decline. Among the US stock
, among the US stock , Nvidia, Broadcom and other super semiconductor big bull stocks in the past, the cumulative decline this year was terrible.
Due to the double blow of shrinking downstream demand and US sanctions, is China's semiconductor still saved? How to break the deadlock? From which field is the breakthrough?
TSMC's epic plunge,
Some A-share semiconductor stocks fell more than 30% in two days
Today, the Taiwan weighted index closed down 4.35%, and the stock price of TSMC, the global semiconductor foundry leader, which Tsai Ing-wen calls the "Shengshan for Protection" TSMC, fell sharply by 8.3%, setting the largest drop in history.
In fact, before the epic drop in Taiwan's stock market, TSMC had already fallen to the forefront. In the past two trading days, TSMC's US stock market has fallen by 9.29%. Since hitting a record high on January 13 this year, TSMC's stock price has so far halved, with a drop of 53.49%.
Compared with TSMC, the trends of some A-share semiconductor listed companies are even more scary. In just two trading days, the stock prices of Haiguang Information, Shengmei Shanghai and Tuojing Technology fell by more than 30%, and the leading 100 billion Northern Huachuang even hit the limit for two consecutive trading days.
Judging from the Dragon Tiger List data released after the market, institutions fled the above shares frantically, among which Haiguang Information was net sold by five institutions for 505 million; Northern Huachuang was net sold by four institutions for 238 million; Tuojing Technology was net sold by five institutions for 282 million, and Shengmei was net sold by three institutions for 121 million. The institution fled a total of 1.146 billion from four companies. Such a decline law as
is really rare.
How miserable are global semiconductors this year? The value line compiles the changes in the market value of the top 20 semiconductor companies in the US stock market and the top 20 semiconductor companies in the A-share market this year.
It can be seen that the top 20 semiconductor companies in the US stock market value have evaporated by nearly US$1.4 trillion this year, or nearly 10 trillion yuan, almost halfway. Among them, Nvidia's market value evaporated by nearly 3.2 trillion yuan, and TSMC's market value evaporated by nearly 2 trillion yuan. The situation of A-shares is similar. Not counting Haiguang Information and Tianyue Advanced, which are listed this year, the market value of the top 20 semiconductor companies in A-share market value has evaporated by RMB 831.2 billion this year, and the latest total market value is only about RMB 1.2 trillion, less than half of a company of TSMC.
The most core reason for the collapse of global semiconductors is the shrinking downstream demand. A research report released by Strategy Analytics pointed out that global smartphone shipments will drop by 7.8% year-on-year in 2022; the latest data released by IDC shows that PC shipments in the third quarter of 2022 fell by 15% year-on-year; in addition, GPU, which was previously hot in demand due to "mining", has also sharply reduced demand with the plunge of cryptocurrency.
For domestic semiconductor companies, in addition to the depression of downstream demand, they also have to deal with the escalating sanctions from the United States, which is undoubtedly a worse situation.
Diplomatic semiconductor dilemma
Semiconductor is a huge industry, including integrated circuits (ICs), optoelectronics, separators and sensors, among which the scale of IC accounts for more than 80%.
The so-called chip is silicon wafer containing integrated circuits. It is divided into dozens of major categories and thousands of minor categories. Manufacturing a small chip involves more than 50 disciplines and thousands of processes, including three major links: design, manufacturing and packaging.
In this industrial chain, the gap between domestic enterprises and the world's advanced level is comprehensive.
First look at the design, Huawei HiSilicon and Unisex Zhanrui ranked in the top two in China. At present, the two companies are already the world's leading level in many fields, but a huge problem is that the core of their architectural authorization is controlled by outsiders.
At present, only of the Chinese Academy of Sciences and Shenwei of the General Staff have independent architecture. The former is used for Beidou navigation , and the latter is used for Shenwei supercomputers. The civil field is basically blank.
equipment and materials are another very big shortcoming.The three major devices for making chips, lithography machine , etching machine and thin film deposition. Only the dielectric etching function of China Micro Semiconductor has kept up with the industry pace, and its 7-nanometer equipment has been shortlisted for TSMC's list.
In addition, Beifang Huachuang has performed well in oxidation furnaces and thin film deposition equipment, but it is basically at the 28nm level. Other equipment, such as ion implanter, polisher and washing machines, are similar. The biggest gap between
is the lithography machine. The lithography machine is used to expose the designed circuit diagram onto the silicon wafer, while the etching machine is responsible for microscopic engraving and engraving grooves or contact holes. At present, the most advanced EUV lithography machine of the Dutch company ASML will soon be put into the 7-nanometer process of Samsung and TSMC, while the domestic Shanghai Microelectronics lithography machine is still at the level of 90-nanometer mass production. In terms of
materials, Japan is the world leader. Among the 19 major materials for making chips, 14 of Japan rank first in the world, with a total share of more than 60%. Nearly 70% of the world's silicon wafer is produced in Japan, which is the foundation of chip manufacturing.
chip manufacturing, the most advanced in China are SMIC and Xiamen United Xin, which can currently achieve 28nm mass production. And their competitors, giants such as Samsung and TSMC, will soon mass-produce 7 nanometers this year, a difference of two or three generations.
In such a long industrial chain, global cooperation is indispensable. Taking lithography machines as an example, Netherlands ASML is ahead of its own, but its success is thanks to the strong cooperation of various countries. The lens comes from Germany Zeiss and the light source comes from the United States. This is almost the technological crystallization of Western industry in the past century.
. China is in a disadvantaged position in this industrial chain and is now facing a very unfriendly industrial environment. How does the
industry break through?
Founder Securities semiconductor researcher Chen Hang deduced the future breakthrough path of the domestic semiconductor industry in the latest research report. The situation before
was:
design company felt that the domestic Fab (OEM) technology was not good, so it preferred to choose foreign-funded wafer factory, which led to the domestic wafer fab level being on the spot.
Domestic Fab feels that domestic semiconductor equipment is not powerful, so they delay progress and KPIs, and they buy a few units to deal with national projects, and most of them buy imported goods.
is as large as an etching machine, as small as photoresist , without downstream iterative feedback, how can we achieve perfection every day?
You should know that TSMC's wafer foundry level depends on the repeated cultivation and learning of design companies such as Apple and Nvidia to achieve perfection.
You should know that ASML's lithography machine relies on the joint cooperation with TSMC, in-depth research and development, and iteratively, to become the world's number one.
Many domestic manufacturers are waiting and watching. If you have American equipment, buy American products and put them aside. Now, the strictest chip blockade in American history is coming, and the fenceists finally have no way out!
In the future, there will be only one situation:
design companies have lost foreign capital fabs, and have honestly turned to domestic wafer fabs and jointly optimized.
Domestic Fab has lost American equipment and has honestly turned to domestic semiconductor equipment and worked together.
Domestic Fab has a material crisis and is honest and comprehensive in cooperation with domestic materials that he doesn't like, and adjust and tackle key issues.
The Chinese semiconductor industry, which is forced into a corner, will form three major "study groups" of design-factories, wafer fabs, and wafer fabs, and materials.
In addition, the population of 1.4 billion in China is PK 1 billion in Western Europe + Japan and South Korea + North America, the population of STEM engineers is several times that of the latter. Combined with the advantages of late-developing countries, the upstream equipment, materials, parts, large chips, and EDA/IP of domestic wafer factories will usher in a new stage of development. Which areas of
are expected to be the first to make breakthroughs?
The basis of everything is materials and equipment.
semiconductor material , in addition to the main material of silicon, the remaining materials can be roughly divided into two categories: lithography and auxiliary lithography materials.
semiconductor equipment has become related to the water-emulsion fusion of materials, so semiconductor equipment can also be roughly divided into two categories: lithography and auxiliary lithography equipment.
Materials are a paradise of chemistry and equipment is a world of physics. From mastering basic theories to the industrialization of technology, it also takes a long time to precipitate. Therefore, if you want to enter the field in equipment and materials, you can only make breakthroughs through long enough research and sufficient funds to invest, and materials and equipment cannot be developed in isolation. Especially when developing new semiconductor processes, equipment is materials based on materials, and materials are materials of equipment. Equipment manufacturers and material manufacturers should give each other tolerant and provide trial and error opportunities. Therefore, if you want to make breakthroughs in semiconductor materials and semiconductor equipment, you must not rely on simple market behavior, and you must not expect immediate results. In terms of semiconductor software, converting mathematical models into simulation results and converting executors into streaming graphics is not very difficult as long as the rules are set. The ease of use of software and user habits are often the key to whether a software can succeed. After years of survival of the fittest, Synopsys, Cadence, Silvaco and other software are left behind. These software are based on Linux. There are many Chinese designers. In theory, breakthroughs can be made in a short period of time, but they cannot be replaced. This is embarrassing. Imagine that if the Chinese develop a set of semiconductor design software for the Chinese to use, is this considered entering the market or out? Therefore, in terms of semiconductor software, just like we have our own office software such as Yongzhong , WPS, etc., but we mainly use Office to write documents. We need to have it, but it is also difficult to rely on it to enter the game. In addition, similarly, the development of semiconductor EDA software must not rely on pure market behavior. In terms of semiconductor production capacity, Taiwan, China has become one of the world factories of semiconductor chip because of its special nature. If semiconductors no longer have strategic significance, it is natural for mainland China to become a new world factory, but if semiconductors remain strategic, mainland China will still face the pressure of introducing new process equipment and orders from external customers. Therefore, in terms of semiconductor production capacity, if there has been no endorsement of independent core technologies, even the packaging and testing that have the smallest gap between China and the world will always be separated from a position not far or near the world's top factories. In the unstable upstream of China's semiconductor industry, production capacity must be cautious and cautious. Taiwan, China, will become a good buffer zone for a long time, and it will be difficult for the mainland to rely on a single-point breakthrough in production capacity to enter the market. In terms of semiconductor production capacity, it is best to make it a simple market behavior.
Semiconductor design belongs to the category of superstructure. Once a company successfully starts, it can produce blood and roll forward. There have been some companies in China that have spent the darkness before dawn in the subdivision. According to the rules of the game of semiconductor design, some modules can be purchased from EDA or Fab manufacturers; some modules need to be implemented by themselves; and some modules must be paid for patent fees to others even if they are implemented by themselves and rely on downstream applications. When we look down on China's entire semiconductor design industry from a higher perspective, China's semiconductor design industry is still weak overall, especially in analog design, not entirely because it started late, but also because of the lack of trial and error and iteration opportunities. Compared with European and American design companies, there is a gap in price and performance.
In short, to break the deadlock, we need to make continuous policy investment in semiconductor materials and equipment, so that certain key process links become part of the division of labor in the global semiconductor industry, and build an ecology of smooth production and sales and self-circulation in semiconductor design, so that it becomes part of the socialist market economy with Chinese characteristics.
The material and equipment level of China's semiconductors determines the productivity level of China's semiconductors. The design level of China's semiconductors determines the production relationship of China's semiconductors. The former is the premise and basis of the latter and determines the development and changes of the latter. The latter plays a role in the former. Whether the latter can exist in an active form will either liberate, or bind, or promote, or hinder the iteration and evolution of the former.
In the future, semiconductor materials, equipment and design cannot be separated from two elements: the will of the country and talents. The will of the country can bring funds, and talents can bring technology.
For investors, it is worth noting that China's semiconductor industry has huge room for a breakthrough, and a number of super big bull stocks will inevitably appear; at the same time, the rise of China's semiconductor industry is a very long and arduous process, and a number of companies will inevitably be eliminated. This process, if you bet too early or too late, or if you bet the wrong company, there will also be huge risks.
This article is derived from the value line
The three major devices for making chips, lithography machine , etching machine and thin film deposition. Only the dielectric etching function of China Micro Semiconductor has kept up with the industry pace, and its 7-nanometer equipment has been shortlisted for TSMC's list.In addition, Beifang Huachuang has performed well in oxidation furnaces and thin film deposition equipment, but it is basically at the 28nm level. Other equipment, such as ion implanter, polisher and washing machines, are similar. The biggest gap between
is the lithography machine. The lithography machine is used to expose the designed circuit diagram onto the silicon wafer, while the etching machine is responsible for microscopic engraving and engraving grooves or contact holes. At present, the most advanced EUV lithography machine of the Dutch company ASML will soon be put into the 7-nanometer process of Samsung and TSMC, while the domestic Shanghai Microelectronics lithography machine is still at the level of 90-nanometer mass production. In terms of
materials, Japan is the world leader. Among the 19 major materials for making chips, 14 of Japan rank first in the world, with a total share of more than 60%. Nearly 70% of the world's silicon wafer is produced in Japan, which is the foundation of chip manufacturing.
chip manufacturing, the most advanced in China are SMIC and Xiamen United Xin, which can currently achieve 28nm mass production. And their competitors, giants such as Samsung and TSMC, will soon mass-produce 7 nanometers this year, a difference of two or three generations.
In such a long industrial chain, global cooperation is indispensable. Taking lithography machines as an example, Netherlands ASML is ahead of its own, but its success is thanks to the strong cooperation of various countries. The lens comes from Germany Zeiss and the light source comes from the United States. This is almost the technological crystallization of Western industry in the past century.
. China is in a disadvantaged position in this industrial chain and is now facing a very unfriendly industrial environment. How does the
industry break through?
Founder Securities semiconductor researcher Chen Hang deduced the future breakthrough path of the domestic semiconductor industry in the latest research report. The situation before
was:
design company felt that the domestic Fab (OEM) technology was not good, so it preferred to choose foreign-funded wafer factory, which led to the domestic wafer fab level being on the spot.
Domestic Fab feels that domestic semiconductor equipment is not powerful, so they delay progress and KPIs, and they buy a few units to deal with national projects, and most of them buy imported goods.
is as large as an etching machine, as small as photoresist , without downstream iterative feedback, how can we achieve perfection every day?
You should know that TSMC's wafer foundry level depends on the repeated cultivation and learning of design companies such as Apple and Nvidia to achieve perfection.
You should know that ASML's lithography machine relies on the joint cooperation with TSMC, in-depth research and development, and iteratively, to become the world's number one.
Many domestic manufacturers are waiting and watching. If you have American equipment, buy American products and put them aside. Now, the strictest chip blockade in American history is coming, and the fenceists finally have no way out!
In the future, there will be only one situation:
design companies have lost foreign capital fabs, and have honestly turned to domestic wafer fabs and jointly optimized.
Domestic Fab has lost American equipment and has honestly turned to domestic semiconductor equipment and worked together.
Domestic Fab has a material crisis and is honest and comprehensive in cooperation with domestic materials that he doesn't like, and adjust and tackle key issues.
The Chinese semiconductor industry, which is forced into a corner, will form three major "study groups" of design-factories, wafer fabs, and wafer fabs, and materials.
In addition, the population of 1.4 billion in China is PK 1 billion in Western Europe + Japan and South Korea + North America, the population of STEM engineers is several times that of the latter. Combined with the advantages of late-developing countries, the upstream equipment, materials, parts, large chips, and EDA/IP of domestic wafer factories will usher in a new stage of development. Which areas of
are expected to be the first to make breakthroughs?
The basis of everything is materials and equipment.
semiconductor material , in addition to the main material of silicon, the remaining materials can be roughly divided into two categories: lithography and auxiliary lithography materials.
semiconductor equipment has become related to the water-emulsion fusion of materials, so semiconductor equipment can also be roughly divided into two categories: lithography and auxiliary lithography equipment.
Materials are a paradise of chemistry and equipment is a world of physics. From mastering basic theories to the industrialization of technology, it also takes a long time to precipitate. Therefore, if you want to enter the field in equipment and materials, you can only make breakthroughs through long enough research and sufficient funds to invest, and materials and equipment cannot be developed in isolation. Especially when developing new semiconductor processes, equipment is materials based on materials, and materials are materials of equipment. Equipment manufacturers and material manufacturers should give each other tolerant and provide trial and error opportunities. Therefore, if you want to make breakthroughs in semiconductor materials and semiconductor equipment, you must not rely on simple market behavior, and you must not expect immediate results. In terms of semiconductor software, converting mathematical models into simulation results and converting executors into streaming graphics is not very difficult as long as the rules are set. The ease of use of software and user habits are often the key to whether a software can succeed. After years of survival of the fittest, Synopsys, Cadence, Silvaco and other software are left behind. These software are based on Linux. There are many Chinese designers. In theory, breakthroughs can be made in a short period of time, but they cannot be replaced. This is embarrassing. Imagine that if the Chinese develop a set of semiconductor design software for the Chinese to use, is this considered entering the market or out? Therefore, in terms of semiconductor software, just like we have our own office software such as Yongzhong , WPS, etc., but we mainly use Office to write documents. We need to have it, but it is also difficult to rely on it to enter the game. In addition, similarly, the development of semiconductor EDA software must not rely on pure market behavior. In terms of semiconductor production capacity, Taiwan, China has become one of the world factories of semiconductor chip because of its special nature. If semiconductors no longer have strategic significance, it is natural for mainland China to become a new world factory, but if semiconductors remain strategic, mainland China will still face the pressure of introducing new process equipment and orders from external customers. Therefore, in terms of semiconductor production capacity, if there has been no endorsement of independent core technologies, even the packaging and testing that have the smallest gap between China and the world will always be separated from a position not far or near the world's top factories. In the unstable upstream of China's semiconductor industry, production capacity must be cautious and cautious. Taiwan, China, will become a good buffer zone for a long time, and it will be difficult for the mainland to rely on a single-point breakthrough in production capacity to enter the market. In terms of semiconductor production capacity, it is best to make it a simple market behavior.
Semiconductor design belongs to the category of superstructure. Once a company successfully starts, it can produce blood and roll forward. There have been some companies in China that have spent the darkness before dawn in the subdivision. According to the rules of the game of semiconductor design, some modules can be purchased from EDA or Fab manufacturers; some modules need to be implemented by themselves; and some modules must be paid for patent fees to others even if they are implemented by themselves and rely on downstream applications. When we look down on China's entire semiconductor design industry from a higher perspective, China's semiconductor design industry is still weak overall, especially in analog design, not entirely because it started late, but also because of the lack of trial and error and iteration opportunities. Compared with European and American design companies, there is a gap in price and performance.
In short, to break the deadlock, we need to make continuous policy investment in semiconductor materials and equipment, so that certain key process links become part of the division of labor in the global semiconductor industry, and build an ecology of smooth production and sales and self-circulation in semiconductor design, so that it becomes part of the socialist market economy with Chinese characteristics.
The material and equipment level of China's semiconductors determines the productivity level of China's semiconductors. The design level of China's semiconductors determines the production relationship of China's semiconductors. The former is the premise and basis of the latter and determines the development and changes of the latter. The latter plays a role in the former. Whether the latter can exist in an active form will either liberate, or bind, or promote, or hinder the iteration and evolution of the former.
In the future, semiconductor materials, equipment and design cannot be separated from two elements: the will of the country and talents. The will of the country can bring funds, and talents can bring technology.
For investors, it is worth noting that China's semiconductor industry has huge room for a breakthrough, and a number of super big bull stocks will inevitably appear; at the same time, the rise of China's semiconductor industry is a very long and arduous process, and a number of companies will inevitably be eliminated. This process, if you bet too early or too late, or if you bet the wrong company, there will also be huge risks.
This article is derived from the value line
