Global chip crisis, the United States is no exception. U.S. President Biden met with a group of members from both parties in the Oval Office of the White House on the 24th to discuss U.S. supply chain issues, especially the global chip shortage caused by production reduction, CCTV financial aid Reuters reported. An executive order was signed after the meeting to conduct a wide-ranging review of the supply chain of key industries, including the semiconductor industry.
Blocked by the United States, China's chip industry has encountered a "neck jam". In the stock market, the relevant chip stocks are soaring. Can China achieve "independence" in the semiconductor industry? What is the current level of the industry? What are the investment opportunities for the semiconductor industry.
Yang Shengjun, partner of cornerstone capital and Senior Semiconductor expert, gave us an interpretation. Yang Shengjun graduated from Fudan University and the Department of electronic engineering of Oregon State University. He once worked in radico microelectronics, and has more than ten years of industrial experience in the semiconductor chip industry. Have a deep understanding and rich investment experience in the field of science and technology.
Cornerstone capital is a well-known private equity investment institution with more than 20 years of investment experience and accumulated asset management scale of more than 50 billion yuan. In the semiconductor design, manufacturing, packaging, testing and other industrial chain, a large number of leading companies have been set up. Famous investment projects include image sensor chip design enterprise Howell Technology (Weill Co., Ltd.) and gekewei, RF filter chip manufacturer Haoda electronics, chip emerging packaging and testing company Yongsi electronics, etc.
Question 1: why does the supply of chips suddenly exceed the demand, and how do you view the current prosperity of the chip industry? Where is the future development space of semiconductor industry? How to look at the development of the third generation semiconductor industry?
Yang Shengjun: the growth opportunities of China's new generation of chip enterprises can be summarized in two aspects: one is the national system, including domestic substitution, as well as the technology innovation board and registration system, which have been mentioned many times in the past two years; the other is the industrial upgrading and emerging industrial applications, which have greatly expanded the chip market.
Secondly, the emergence of emerging industrial applications such as artificial intelligence, 5g, IOT and new energy vehicles has opened up a new chip market. New energy vehicles are the most promising applications. The new four modernizations of new energy vehicles (electrification, intellectualization, networking and sharing) are the main trend of automobile development. Its intelligent driving, power transmission, body control, safety system and entertainment equipment all need a large number of advanced chips, and the charging pile also needs chips. The richer the functions of the car, the more advanced the intelligent driving technology, and the higher the chip requirements. Compared with mobile phones, the volume of new energy vehicles may be one order of magnitude lower, but the unit price is more than one order of magnitude higher. Therefore, if its growth forecast can be fulfilled as scheduled, it will support a trillion scale for the semiconductor industry.
To sum up, China's chip production capacity is very scarce, and supplier resources have become the core demand of domestic and foreign manufacturers. There is no possibility of a U-turn in the semiconductor industry in the short term, and no cycle in at least two years. The leading enterprises or relatively leading enterprises in the industrial chain have achieved rapid growth. For example, the profit of Weill (howay Technology) invested by cornerstone capital is only 400 million in 2019, and it has risen to more than 3 billion in 2020.
Let's talk about the third generation semiconductor industry. First of all, the semiconductor industry has extreme requirements for the purity and complexity of the materials used, so materials play an important role in the semiconductor industry. The level of semiconductor materials is an important symbol to measure the level of a country's fine chemical industry.
There are two kinds of materials in semiconductor industry, one is the main material, such as silicon or compound wafer material, the other is the auxiliary material, such as photoresist. In terms of industrial accumulation in the field of domestic semiconductor materials, there is a big gap between the main materials and auxiliary materials and foreign leading enterprises such as the United States and Japan. For example, photoresist is the pain point of domestic semiconductor industry chain.
According to the development history of the main material system of the semiconductor industry, germanium and silicon belong to the first generation, gallium arsenide and indium phosphide belong to the second generation, silicon carbide and gallium nitride belong to the third generation.
The third generation semiconductor has the characteristics of high temperature resistance, high voltage resistance, high power and radiation resistance. It is suitable for manufacturing microwave RF, optoelectronics, power electronics and other devices. It is suitable for high voltage and high power scenarios. It is the best choice for photovoltaic, UHV power transmission and new energy vehicle chip control materials.
The third generation semiconductor materials have three advantages
One is faster, which helps to improve the performance of the chip. The third generation semiconductor uses wide band gap material, which makes the leakage current smaller when turning off, the on impedance smaller when turning on, and the parasitic capacitance much smaller than that of silicon process material, so the chip runs faster, consumes less power and has longer standby time. The third generation semiconductors can achieve part of the performance of silicon advanced nodes with larger process nodes.
Second, high energy conversion efficiency and low power loss. Taking new energy vehicles as an example, the data displayed by Ford motor show that compared with electric vehicles driven by traditional silicon chips (such as IGBT), the energy consumption of new energy vehicles driven by third generation semiconductor chips (such as SiC chips used by Tesla Model 3) is about five times lower, which greatly increases the driving range. From the perspective of energy saving, the power consumption of a large data center room in a year is equivalent to that of a medium-sized city. If the third generation semiconductor chip is used to control the power supply, it will save a lot of power compared with the traditional silicon chip.
Third, it can withstand higher power and higher voltage. Third generation semiconductors can greatly improve the power density of products to meet the needs of future electric vehicles with higher power, higher voltage and higher current.
Based on the above advantages, the opening of new application scenarios such as new energy vehicles, 5g, artificial intelligence and super large data center will bring huge development space to the third generation semiconductor, and give birth to a potential market of more than one trillion yuan. More importantly, the third generation semiconductor will play an important role in helping mankind popularize emerging energy, develop clean energy and realize the great goal of carbon neutrality in the future!
Question 2: secondly, the most puzzling question of the general public is: why can't China build a lithography machine even though China can build an atomic bomb?
Yang Shengjun: first of all, we need to correct a saying that China is not unable to build a photoresist, but it is unable to build a high-end photoresist at present. For example, Shanghai microelectronics has been able to build a photoresist with 90nm process, and it is said that it will deliver the first photoresist with 28nm process in 2021-2022, but this is compared with the 7Nm or even 5nm of ASML, the most leading company in the industry EUV lithography, there is still a big gap. Now the most high-end flagship mobile phone SOC chip is 5nm process, 28nm process can not meet the needs of consumer market competition.
Figure: Shanghai microelectronics ssa600 / 20 lithography machine and its technical parameters source: Shanghai microelectronics official website
So we need to discuss not only whether it can be made, but also whether it is good enough.
As a topic of military strategy, whether the atomic bomb can be built or not is the core. After a theoretical breakthrough, the atomic bomb can be detonated as long as the engineering difficulties such as high concentration of nuclear materials and centrifuges are overcome. Even if the atomic bomb is built, its strategic deterrent effect will be achieved. As for the equivalence of the atomic bomb, the range and the precision of the atomic bomb, these problems are not good enough, which can be solved slowly in the future, that is, its technology iteration cycle It could be longer.
But the lithography machine is not the same, just do it is not enough, must be good enough to have meaning. In addition to national military security, lithography is also related to national economic security and commercial competition in the context of globalization. The semiconductor industry is a "winner takes all" global industry. For example, in the most competitive mobile phone market, if a competitor's flagship mobile phone SOC chip adopts the most advanced 7Nm or 5nm process, and you are still using 14nm, your mobile phone performance will definitely fall behind, and you will lose market competitiveness. Just like now Intel's latest computer CPU is the 11th generation i7, computers with I3 processor can not occupy the high-end market. Therefore, the performance of the lithography machine should reach the level similar to that of the most leading competitors in the industry before it can be competitive. This means that the technology iteration speed of the lithography machine is much faster than that of the atomic bomb, which greatly increases the difficulty of the technology development of the lithography machine.
What is the difficulty of lithography technology? The difficulty is that it is the product of the most advanced theory, technology and technology across many disciplines and industries. Photoresist is known as the "pearl on the crown of semiconductor". A photoresist is composed of more than 100000 parts, which is the product of top technology in the fields of mathematics, chemistry, precision optics, fluid mechanics, precision machinery, automation control, software engineering, image recognition, electronic circuit, etc. Therefore, lithography manufacturing depends on the progress of various disciplines, as long as one of the disciplines bottleneck, it can not be made.
Take the manufacturing process as an example, the most difficult part of the lithography machine is its precision control. Under the requirements of nanometer level and extreme precision, many neglected details will become problems, such as vibration, light source frequency, reflector, etc. even the air in the dust-free room is required to be 10000 times cleaner than the outside. An American engineer once said that the adjustment time of a lithography machine is as long as several decades, and the number of size adjustments is more than a million times. Of course, there must be exaggeration in this engineer's statement, but the precision control of the lithography machine can be seen.
Even ASML, which is the only company in the world, can't manufacture lithography independently. Instead, it integrates global technology. 90% of the more than 100000 parts are provided by many advanced high-tech industrial factories. For example, Cymer (which has been acquired by ASML) from the United States is used as the light source system, Zeiss (optical system) from Germany is used as the lens (optical system), and so on.
China is solving the previous generation of 90nm and 28nm processes. It may take 5 to 10 years or more to develop to 7Nm and 5nm. In the future, once our high-end photoresist is made, it means that all the basic science related to our national photoresist has made a major breakthrough, and our whole supporting semiconductor and its related industries have also ranked among the top in the world.
Question 3: what are the key factors for the breakthrough development of the semiconductor industry?
Yang Shengjun: in the past half century, the development of semiconductor industry has been mainly affected by the following three factors: (1) the establishment of theory and technology system; (2) engineers and entrepreneurship; (3) global industrial chain.
First of all, theory and technology system. Semiconductor is the integration of modern science and engineering technology system. In the 70 years since the invention of transistor, through the efforts of several generations of outstanding scientists and engineers, the basic theory and technology system of semiconductor has been completed. However, a large number of sub basic theories and technologies still need to be solved in order to transform these theoretical designs into manufacturing and commercialization.
Semiconductor chip is the most interdisciplinary technology field. From modern natural sciences such as basic physics, mathematics and chemistry to modern engineering theories such as information theory, cybernetics and communication theory, and then to modern engineering technologies such as material engineering, fine chemical engineering, optical engineering, ultra precision machinery and basic software, we need to master complex theoretical design as well as complex industrial system No enterprise can master the complete industrial chain alone.
In contrast, photovoltaic, lithium battery and other industries have relatively simple technology system for IC chip industry. There are differences of order of magnitude in theoretical design, manufacturing process, equipment precision, material requirements, etc. these industries have been basically mastered in China.
Second, engineers and entrepreneurship. The progress of the chip industry in the past few decades is more due to the outstanding contributions of outstanding scientists and engineers.
Semiconductor is a rare industry with strong personal heroism. A few outstanding entrepreneurs have played a very important role, and they even say that "one general is one division". Because a very good general can basically promote a technology for several generations. The most typical one is Mr. Liang Mengsong of SMIC. In less than a year, he realized the 28nm process of SMIC 14 nm, and achieved a significant increase in yield of 14 nm.
The main reason is that most of the high IQ technical talents are conceited and do not respect the value of business management talents and commercial talents. Outstanding scientists and engineers often lead their companies to disaster and decline.
The third is the global industrial chain. The success and maturity of chips need a lot of verification and shipment, so the global supporting industry chain is the key driving force for the development of chip industry.
The evolution of chip technology comes from the successive iterations, and the huge R & D costs and capital expenditure of the chip industry need the long-term and continuous profit support of the downstream industry chain. The cross generation development of chips is driven by the strong driving force of downstream applications. For example, without PC and Internet, there would be no Intel, and without smart phones and mobile Internet, it would not be able to achieve TSMC; Intel, which has lost its downstream strong growth, has gradually become a "toothpaste factory" lagging behind the industrial development, and now its market value has been greatly surpassed by TSMC and NVIDIA.
Question 4: what is the current level of China's semiconductor industry? From 2018 to now, in addition to the high expectations of the capital market, what has happened to China's chip industry at the physical level? Is there any progress in China's chip industry? From a qualitative point of view, can and when can Chinese chips catch up with the United States?
Yang Shengjun: according to the data of SIA in the United States, in the past 30 years, the United States has accounted for about 50% of the global chip industry chain, while China now has about 5%, which is 10 times the gap.
Source: WSTS, SIA, IHS Global, PwC
After looking at the status quo, let's look at the speed of catching up. According to the official statistics of us listed companies, the R & D investment and capital expenditure of us chip listed companies in 2019 totaled 71.7 billion US dollars. From 1999 to 2019, the total investment of us chip listed companies was nearly 900 billion US dollars. The sum of the first and second phases of China's national fund is 300 billion yuan, an order of magnitude less.
Another data is even more shocking. In the past two decades, the average annual R & D investment and sales of listed chip companies in the United States accounted for 16.4%, while that of listed chip companies in China was 8.3%, about half of that of the United States.
This shows that China has a long way to go before it can surpass the United States. As pursuers, we have to increase investment in order to be late comers. On the other hand, in 1894, the GDP of the United States was the first in the world, but it was not until after the second world war that it became the first in science and technology. China may also need such a process.
In recent years, the progress of China's chip industry is obvious to all. At present, we have solved the problems from the middle and low end. Domestic CPUs have obtained application opportunities in China. In the future, on this basis, we will continue to iterate and upgrade from there to better;
In terms of manufacturing, SMIC has also made great progress. The process has been upgraded from 28nm to 14nm. Although TSMC is faster, our progress is huge. If we are not stuck by the lithography machine, we will definitely achieve a breakthrough in 7Nm and 5nm. At present, SMIC ranks the third in the world;
In terms of semiconductor equipment, China micro semiconductor and North Huachuang also achieved a certain proportion of domestic substitution;
In semiconductor materials, 8-inch and 12 inch silicon wafers also have a certain proportion of localization;
Question 5: in this case, how can China achieve breakthrough development in the semiconductor industry? How to look at the development path of China's chip industry and the development space of domestic substitution?
Yang Shengjun: there are four basic factors for the development of China's chip industry: first, capital; second, talents; third, time; fourth, industrial chain support.
With the continuous increase of national policy support, especially after the launch of the science and technology innovation board, the situation of funds and talents has improved significantly. The huge wealth effect has attracted a large number of overseas talents to return. Many of my classmates who stayed in the United States in the past two years have returned home to start a business. The listing conditions of science and technology innovation board are more relaxed than before. As long as a certain scale is achieved, it can be listed on science and technology innovation board. The temptation of financial freedom brought by this rapid wealth creation is great.
Time is the most difficult problem to solve. The core difficulty of the chip is that there is no shortcut and we can only rely on continuous iteration. The R & D cycle of a chip is very long. Generally, it can only be iterated twice a year, which greatly restricts our catching up speed.
There is also a very critical but often overlooked factor, that is the market. The upstream and downstream industry chain of chips must be opened up, which is the most critical issue.
The restrictions and embargoes imposed by western countries on China's semiconductor industry have been clearly stipulated in the Wassenaar agreement as early as 2017. It is not after ZTE was sanctioned by the United States that ZTE suddenly appeared. The more important reason why China's chip industry can not develop rapidly is that the downstream enterprises in the domestic industry chain, especially the leading enterprises in the electronic information industry, have been ignoring or even rejecting domestic chips.
The chip industry is a very mature global competitive industry. If we compete with the world in a completely open system, as long as our chips are not as good as NVIDIA, no customers will buy them, and naturally no manufacturers will design and produce them. This is not just a matter of money, but if there is no feedback from downstream users, chip manufacturers will not be able to carry out product iteration, and they will never be able to produce good enough products. The reason why China's Internet industry is developing so fast is that we have a lot of user data, which helps us iterate at an unimaginable speed. That's why most of the big Internet companies are in China and the United States.
If there is no external force, allowing chips to become a completely free competitive market and no downstream enterprises to use domestic products, it will form a vicious circle and never develop.
Here we can discuss what kind of role the government can play, which is also the core issue of the industrial development of the late developing countries. The government should introduce relevant industrial policies, take certain protective measures for our self-produced products, support the development of domestic chips, and give low-end products the opportunity of application and iteration.
In this regard, we can refer to the successful experience of our country in developing domestic communication industry from the late 1990s to 2010. Professor Huang Weiwei described this in detail in his book asymmetric competition. In the early 2G era, China's mobile communication equipment basically depended on imports, and the development of local enterprises was far behind the rapid expansion of the market. In this case, the Ministry of information industry at that time put forward a timely request to give priority to the selection of domestic communication equipment and systems under similar conditions. In the subsequent actual procurement, the Ministry coordinated all parties to sign a domestic program-controlled exchange with volume procurement agreement, which laid a solid foundation for the construction of national communication industry. Later, in the process of developing the 3G / 4G industrial chain, the state also followed this practice, that is, to support the domestic communication equipment industry by setting the proportion of domestic equipment and terminal procurement. Huawei, ZTE and other enterprises' communication equipment developed in this way in the early stage.
Photo: asymmetric competition by Huang Weiwei
The government can support domestic downstream enterprises to purchase and use domestic chips and core components through "asymmetric competition" strategy, and encourage the development of domestic industrial chain.
Although there may be a performance gap between domestic chips and top products, in many scenarios, it is enough to meet the requirements of domestic manufacturers. We can establish appropriate quality standards for chips, so that domestic manufacturers can safely use domestic chips that meet the standards. As long as there are customers, the enterprise's technology will continue to iterate and progress.
Over the past two years, trade frictions have also boosted the rate of chip localization. After the ban on Huawei issued by the United States, the public's awareness of domestic chips has been improved, and domestic enterprises have begun to be willing to use domestic chips. The localization rate of chips is advancing very fast. There are two reasons. First, some enterprises will worry about sanctions when they see Huawei's precedent. Second, as an industry benchmark, Huawei's procurement has played an exemplary role. Even if Huawei is forced to use it, it also shows that the quality of this product is up to standard. Huawei is equivalent to making a brand endorsement.
China imports about US $300 billion of chips a year, so it is impossible for all of these chips to be domestically produced in a short time. However, the government also has expectations to solve the problem step by step from easy to difficult from the perspective of industrial chain. Originally, there was only 5% localization rate. In 2020, there will be 17% localization rate. In two years, it will be 30%. In five years, it will be 60%. Maybe by 2035, it will be 80%. That's a great success.
Question 6: the development of China's semiconductor industry is inseparable from design software EDA tools (Electronic Design Automation) and chip design IP. Without EDA software, all chip design companies will be shut down, and the pyramid of the semiconductor industry will collapse. How do you view the lack of EDA tools and chip design IP in China's semiconductor industry?
Yang Shengjun: chip design software EDA tools and chip design IP are necessary options for the development of China's semiconductor industry, and they are unavoidable practical problems. Based on strategic security considerations, EDA must do it by itself, otherwise, if the United States does not give license, our chip design will be suspended. Because it needs a long time of extensive and profound industry accumulation, the domestic talents in this field are very scarce. At the present stage, we also need to solve the problems first, and then move forward step by step.
Although the market scale of EDA industry is small, only about 10 billion US dollars, but the threshold is very high, especially professional, which needs long-term focus and long-term investment. Compared with lithography and other hardware devices, EDA as software is relatively easy, as long as the architecture and algorithm breakthrough, continuous iteration, almost.
At present, more than 60% of the market share of EDA industry is occupied by Synopsys, cadence and mentor graphics. In China, there are Huada Jiutian, Gelun electronics and other enterprises catching up. Although compared with the top three EDA software enterprises in the world, domestic EDA enterprises are just starting and have simple functions, we hope to make great progress in five years.
Chip design IP is also a field with a big gap at home and abroad, especially the most widely used chip IP such as CPU and GPU. There is still a big gap in China, which may be larger than EDA design software. At present, the global market of chip design IP is mainly occupied by arm, imaging technologies, CEVA and several EDA companies. Whether it is EDA design software or mainstream chip IP, at present domestic talents in this field are very scarce, because it needs a long time of extensive and profound industry accumulation.
Question 7: how to look at the problem of chip manufacturing in China? At present, even if China has designed advanced chips, domestic basic industries can't make them. If qualitative, can Chinese chips be made in the future?
Yang Shengjun: why is chip manufacturing complex? The core is that the manufacturing process of the chip is very long. Chip production is a combination of equipment, materials and technology as a whole. It's not that if you buy lithography equipment, you will be able to do the chip well. You also need to understand the material, the technology and the process.
A complete set of advanced industrial chip production process has hundreds of processes. The material system and equipment used in each process are different. The process design is also different, such as how to design the temperature, humidity and pressure. The material ratio is also different. The sequence of processes is also different. Every detail is a know-how, is an engineering problem, must be solved, in order to produce, at the same time, every link will affect the quality of the final product. As a result, chip manufacturing and design need long-term accumulation. Only by continuous investment and customer use can they accumulate.
This has also led to the formation of giant monopoly in almost every direction of the semiconductor industry. In order to improve the market competitiveness, all manufacturers want to purchase the best products, unless they are targeting at the low-end products or have protective policies. In all branches of chips, the same story is unfolding. The chips made of high-end materials and low-end materials, high-end equipment and low-end equipment are very different. Customers will not choose low-end products, so low-end products will never have the opportunity to develop. This kind of pressure is transmitted layer by layer, and finally it becomes that the best materials and equipment, the best design and the best production process are needed to create a competitive product in the world. And those who can't do the best are basically eliminated.
In the past few years, with the gradual opening of downstream applications, the localization rate has been increasing. More and more enterprises have made more advanced chips. Originally, only foreign manufacturers can supply chips, but now there is a large supply in China. Our industrial chain has gradually formed, and the proportion of localization is higher and higher.
There are excellent enterprises like Huawei and Shangtang in our country, which constantly put forward higher requirements. The chip industry can certainly catch up with them. At the same time, we should also pay attention to the development of supporting industries, otherwise the United States will be in trouble with sanctions.
Question 8: how to view the dispute between Liang Mengsong and Jiang Shangyi of SMIC?
Yang Shengjun: Liang and Jiang's line is not contradictory. Mr. Liang Mengsong does advanced technology while Mr. Jiang Shangyi does advanced packaging. Advanced technology and advanced packaging complement each other. Advanced packaging is to understand and partially realize more advanced system performance and lower cost from the perspective of the overall application of the system, while advanced technology is only realized from the perspective of chip. The research and development of advanced technology is the cornerstone. According to the development law of Moore's law, the long-term sustainable development of advanced technology is beyond doubt. At the critical moment when Moore's law is slowing down and the post Moore era is approaching, it is particularly necessary to advance the development strategy of layout, advanced technology and advanced packaging.
The progress of process, design and packaging technology is the three important factors to promote the development of Moore's law.
Among them, the advanced technology corresponds to the machining accuracy, which is what we usually call the number of nano process. The so-called 7Nm and 5nm refer to the minimum gate length of the transistors inside the chip. The shorter the gate length, the more transistors can be integrated on the same size silicon chip, or the smaller the wafer area to be used under the same chip complexity. Therefore, in general, the smaller the process, the higher the performance of the chip.
In the future, we will definitely achieve 7Nm, 5nm and 3nm step by step, either waiting for ASML's EUV photoresist or waiting for us to develop advanced photoresist. This is the ultimate problem.
Packaging, generally speaking, is how to combine a single or multiple small chips with other materials into a large chip. Advanced packaging technology can improve the synergy effect, save space, reduce power consumption and cost, so as to improve the chip performance and reduce the chip cost without changing the process.
Not long ago, TSMC also announced that it will focus on high-end packaging. At the International Conference on solid state circuits (ISSCC) 2021, which just ended last week, TSMC system expounded the idea of developing advanced packaging. TSMC believes that the performance improvement of advanced processes such as 3nm technology may be only 10% compared with 5nm technology. Today, the most complex chip has integrated 50 billion transistors. In the future, through the innovation of advanced packaging technology, 300 billion transistors can be integrated in an advanced packaging chip, and the performance of the chip can be further improved by 10 times or more.
Mr. Jiang Shangyi also hopes to further explore the potential of the chip through advanced packaging, realize the effect of more advanced technology, and reduce the cost of chip application. When the current semiconductor technology goes to 7Nm and 5nm, Moore's law is still continuing in technology, but the equipment investment and cost of each generation of products are greatly increased. Everyone is talking about the coming of the post Moore era. System level packaging will be the focus of the post Moore era. System level packaging is to package many chips or electronic components into one chip to improve the system integration. It's similar to packaging a small PCB into a chip.
The process evolution is relatively fast, but it becomes more and more difficult. In the past 30 or 40 years, the progress of semiconductor packaging technology is not fast, and the semiconductor packaging and substrate technology has increasingly become the technical bottleneck of the overall system performance. Therefore, in the post Moore era, advanced packaging is one of the important solutions to continue Moore's law, and it is also an industrial direction, and the future of miniaturization is inevitable Chips also need to work from these aspects. Therefore, it is a reasonable choice for SMIC to develop advanced technology and advanced packaging.
Q9: how to view the investment opportunities of semiconductor industry?
Yang Shengjun: we should rationally understand the investment opportunities of semiconductor industry.
In this context, our investment strategy can be summarized as the following four points: (1) comply with the law of industrial development and deal with the characteristics of global market-oriented chip with the principle of market; (2) respect the inherent law of the development of hard technology enterprises and adhere to the long-term principle; (3) respect the professionalism and be careful to be led by the tuyere; (4) all participants should respect and cooperate with each other and not be independent Food.
Taking cornerstone capital as an example, we have invested in more than 10 semiconductor companies in the whole industry chain, basically covering all the subdivisions of materials, equipment, design, packaging and testing.
Specifically, our first focus is on the imagination of segmentation. For example, the camera chip industry is growing very fast. In three or four years, the market size has increased from US $8 billion to US $16 billion, and it will probably reach US $20 billion in 2021. Another example is the filter, which has low localization rate, high ceiling, and fast industry growth. All these are in line with the judgment standard of a good track. In addition, the demand for security, mobile phones and artificial intelligence is also growing.
Image: Howe Technology launched the first full phase detection coverage image sensor source capable of achieving excellent auto focus function in flagship and high-end mobile phone: Howe technology WeChat official account.
In every good segmentation field, we will choose those teams with deep historical accumulation to have long-term research in a field. For example, howay technology is the pioneer of camera chip industry. Ningbo Yongsi electronics, which we invested in, is doing the sealing test and has a very good accumulation. Now it is facing the shortage of production capacity and develops faster. It was established in November 2017 and achieved nearly 1 billion revenue in 2020.
The commercialization ability of an enterprise is also very important. It should be able to complete the transformation from technical advantage to commercialization advantage. This is a very critical step, because the products will be sold in the market eventually. If we can't make a commercial product, we will lose our technical advantage in the end. Many founders of enterprises are technical background, only pay attention to technology, to prevent themselves from falling into the one-sided pursuit of technology.
The semiconductor chip industry has its own development law. After more than half a century of rapid development, the basic theory and technology system of the industry has been established. We are still in the process of rapid iterative development of the industry. This process requires time and continuous investment. The Chinese nation has always had the virtue of diligence and wisdom. These characteristics are necessary for the development of chips The Chinese are suitable for making chips.
We believe that as long as we stick to the current investment, under the current development trend, China's semiconductor industry will basically solve the situation of being stuck by foreign countries in another decade or so, and will certainly have a place in the global industrial chain.