To prevent getting lost, the elevator goes directly to the safety island to report Liu Yadong A
Source: Chinese think tank
Author: Xinwang
From traditional manufacturing to modern manufacturing, from the low-end of the industrial chain to the high-end, it is the only way for China to move from a manufacturing country to a manufacturing power, and it is also the only way for China to get rid of the middle-income trap. The most typical and important thing in modern manufacturing is chips, and the process of manufacturing from low-end to high-end in the supply chain, industrial chain, value chain, and innovation chain is inseparable from chips. The whole process of intelligent manufacturing, from design, production, supply and marketing, service, and products, all need to rely on chips. Without chips, intelligent manufacturing is impossible. It can be said that the degree of application of chips is the standard and mark that defines high-end manufacturing. And China's chips are the most dependent on others, and they are also the most easily controlled by others.
- 01 -Chinese chips, "latecomer" has a long way to go
The origin of modern chips is in the United States. In 1947, the American invented the transistor, and nine years later, the American engineer Jack Kilby and another American physicist co-invented the integrated circuit, which concentrates many shrunken transistors on a single semiconductor silicon wafer. The conductivity of silicon is between conductors and insulators, so it is called semiconductors, and silicon is purified into high-performance semiconductor materials, which is the parent body of chips. Silicon Valley in the United States is named after the research and production of chips with silicon as the matrix of semiconductors. Therefore, the chip industry is called the semiconductor industry, also known as the large-scale integrated circuit industry. Kilby, who invented the chip, was awarded the Nobel Prize in Physics in 2000.
Kilby and his integrated circuit products
As the computing and processing center, the chip laid the foundation of modern industrial civilization and opened the prelude to the information revolution in the 20th century. Maybe people didn't realize this at that time, because chips were still relatively rudimentary at that time. At that time, who would have thought that today's chips are getting smaller and smaller, and the performance is getting more and more powerful. The Chinese translation of the word chip is really a stroke of genius, meaning the heart and the engine. In the future, even if we enter the cloud era and the quantum era, chips will still be indispensable.
After the chip, the United States has further related inventions, such as computers, the Internet, mobile phones, smart phones, etc., which jointly promote the information technology revolution. Therefore, in this round of information technology revolution, the United States is a pioneer and leader, and now the United States is taking the initiative everywhere and upstream everywhere, just because the United States is the source of technology.
In 1958, chips appeared. Five years later, in 1963, the Japanese imported integrated circuits from the United States. At that time, Japan was in the post-World War II recovery period, and the United States adopted a supportive policy and provided technical assistance to Japan. As a result, in the field of microchips, Japan's craftsmanship spirit has been brought into full play, and the blue is better than the blue. As a result, Japan's semiconductor industry has caught up and pushed into the field of civilian consumption on a large scale. So in the 60s and 70s, Japan was the leader of the global semiconductor industry.
In 1965, Japan and South Korea normalized diplomatic relations and Japan began to set up factories in South Korea, and since then, South Korea has mastered all the technology that Japan has imported from the United States, and in some respects it is superior to Japan. In the 60s and 70s, the United States, Japan and South Korea, like a wave, chased after me and formed an international chip industry chain.
In 1965, China realized that the development of large-scale integrated circuits was a very important thing. During the "Cold War", although China developed chips with the system of researching "two bombs and one satellite", until the early 80s of the 20th century, there was still a big gap between China and developed countries in terms of high and new technology, and there was no great progress in the chip industry.
After the reform and opening up, the State Council established the Office of the Leading Group for Electronic Computers and Large Integrated Circuits in 1982. It should be said that in the past 40 years, we have not paid attention to chips. However, domestic chips are still in a relatively backward state in the world, especially high-performance chips, with a huge gap, three-quarters of the world's chip market is in China, but about 80% of China's chips still have to rely on imports. Imports of chips cost more than twice as much foreign exchange as oil. Oil imports were $140 billion, and chip imports were more than $300 billion. We also have some low-end chips for export, about more than 100 billion US dollars, but the deficit is still more than 200 billion US dollars.
In the global chip industry chain, we are in the middle and lower reaches. The upstream high-end technology, core technology, key components, and key patents are not in our hands. Among the leading enterprises in the global chip industry chain, there is not a single Chinese company. At present, the United States accounts for half of the global market, South Korea accounts for about 24%, Japan accounts for 10%, Europe accounts for 8%, China accounts for 3%, and the market position of mainland enterprises is very marginal.
Looking back at the development of Chinese chips over the years, we missed two opportunities. The first time was in the golden stage of chip invention and development, the United States, Japan, South Korea and Europe were basically synchronized to form a global industrial chain, and we did not keep up with it at that time; the second time was after the 90s of the 20th century, the route of market for technology encountered the 1996 "Wassenaar Agreement" obstruction, delaying our pace of catching up. In the past decade, China's chip industry has grown rapidly, with more than 200 chip manufacturers that can mass-produce 14nm chips, and the industrial chain has gradually moved from the middle and lower reaches to the middle and upper reaches.
At present, China's chip technology iteration is two to three generations different from the first camp of the United States, Japan and South Korea, and the goal is to narrow the gap with the first camp to two or one and a half generations by 2025. What does that mean? From the current mass production of 14 nanometers, by 2025 it can mass produce 7 nanometers and 5 nanometers. It has been reported that TSMC will start mass production of 3 nanometers at the end of 2022. Of course, chips are not infinitely smaller, and scientists estimate that the limit of chips is 2 nanometers, which is determined by the molecular diameter of silicon.
- 02 - Why is chip manufacturing so difficult
The chip industry has its own unique internal structure and industrial characteristics. The chip industry chain is divided into five sub-chains, or the chip industry is divided into five major industries.
First, design. How hundreds of millions of lines fit together first requires design. The world's largest chip design company is the British ARM, and in the field of design software, the United States EDA has a monopoly position. Recently, the biggest news in the chip industry is that Nvidia is going to acquire ARM from the United Kingdom, and the United States will be stronger in the chip industry. Huawei's HiSilicon has a design capability of up to 7nm. I once asked Huawei's vice president why China didn't buy ARM? His answer was: China will never have the opportunity to buy such companies in Europe and the United States.
Second, manufacturing, including finished product production and semi-finished product production. The upstream semi-finished products of chips are wafers, and the manufacturing of high-purity wafers is basically monopolized by Japanese companies. In terms of silicon smelting, Japanese companies can increase the purity of smelting to 99.99%, ensuring that the finished wafers are the best. The largest company in this industry is TSMC to manufacture chips on the basis of wafers, and SMIC on the mainland is currently the fifth in the world - of course, only the fifth in the world in terms of output, and the chips it produces are of low grade and the profit margin is not high, because many patented technologies are not in their own hands and are subject to strict supervision in the United States.
Third, packaging and testing. The chip is compressed onto a board for qualification testing. Because there are too many lines and contacts on the chip, even a mistake of a few tenths of a ten-thousandth of a ten-thousandth of a thousand will eventually cause a huge error, so it must be tested one by one. Packaging and testing is basically a labor-intensive industry, and there is little gap between China and the rest of the world in this industry, and it is even in a leading position.
Fourth, equipment. Everyone knows the equipment for the production of chips, among which the most sophisticated EUV lithography machine manufacturer is ASML in the Netherlands, and the equipment manufacturers for wafer production are in Japan, mainly Mitsubishi, Sony and other enterprises. The 7nm process lithography machine is currently only available to ASML in the Netherlands, with a price of more than $100 million, and money may not be able to buy it. Shanghai Microelectronics has been able to produce equipment for the production of 28-nanometer chips.
Fifth, auxiliary materials. Including photoresists, masks, targets, packaging substrates, etc., these materials are still no breakthrough in China.
Chip manufacturing is so difficult, yet so important. It occupies a fundamental and strategic position in the entire national economy, whether it is people's livelihood, national defense, industry, equipment, aerospace, etc., a chip problem is equivalent to a problem with the human heart. Chip is also a global fully competitive industry, but the threshold for entry is high, the cycle is long, capital-intensive, technology-intensive, talent-intensive; investment is often tens of billions of dollars, the required R & D personnel are tens of thousands, the current chip industry is basically the competition between the world's limited oligarchs, is the international market competition across national borders.
Due to the important strategic value of chips, this competition is not only market competition, but also competition between countries, and the chip industry has even become a powerful weapon in the trade war and a key industry for restrictions and sanctions between competitors.
- 03 - Chips in the context of Sino-US competition
The "Iron Curtain" disappears and the "Silicon Curtain" opens. Since the Sino-US economic and trade frictions, "chips" have become a hot word and the focus. On September 15, 2020, under the pressure of the U.S. technology monopoly, TSMC officially stopped OEM for Huawei's Kirin chips. Huawei spent 6 million yuan to bring back the last batch of chips from a charter plane in Taiwan, and it is said that the 6 million yuan was the money raised by all Huawei executives. TSMC gave Huawei everything it could give. However, Huawei's stored chips are only enough to support mobile phone shipments in the first half of 2021.
It has been reported that the U.S. Department of Commerce will add SMIC to the list of entities subject to sanctions. SMIC has just gone public in China, raising more than 20 billion yuan. If SMIC has problems with upstream equipment and technology, the production of high-performance chips will be variable, and the future is not good. In 2023, the United States will completely cut off the supply of Huawei chips.
In May 2022, U.S. President Joe Biden signed the CHIPS and Science Act totaling up to $280 billion to promote the return of semiconductor companies and increase production in the United States to help the United States lead the world in the production of advanced chips. On October 7, 2022, the U.S. Department of Commerce's Bureau of Economic Security once again launched restrictions on chips in China, including equipment, parts, technology, talents and other aspects, with a ruthless and wide scope, in four key areas: supercomputing, artificial intelligence, automobiles, and mobile phones, almost one-size-fits-all. On January 28, 2023, the United States, the Netherlands, and Japan reached an agreement on restricting the export of advanced chip manufacturing equipment to China, involving equipment products from ASML, Nikon, Tokyo Electron and other companies.
In the field of chips, China is basically unable to counter US sanctions, supply and demand and technology level are extremely asymmetrical, for Intel, Qualcomm, Apple, Microsoft and other foreign companies, we are strong dependence, and it is weak dependence on China. Just as the United States wants to curb Douyin and WeChat, we can't counter it, because Google has long left China, and Facebook has never been to China at all.
Both the US National Security Strategy and the US China Strategy regard China as its number one competitor, and it has become an established policy to contain China in an all-round way. Using chips as a force point for asymmetric precision strikes is the least opportunity-cost approach for the United States, and it does the most harm to China. The economic and trade friction with tariffs as the subject matter has been transformed into technological friction and industrial friction with chips as weapons. Huawei is its victim, the successor was once detained, the chip supply was cut off, the operating procedures expired, and the U.S. Department of Commerce has more than 1,000 companies on the entity list for export control to China, and Huawei and its related companies account for more than 100. Of course, there is a reason why the United States is afraid of Huawei. The information technology revolution in history was mainly initiated by American telecommunications companies. Huawei is a company with the genes of the communications industry, not only with mobile terminal products, but also with world-leading 5G technology. In the coming digital era and intelligent era, how can we compete for dominance? Chips have become a key link in key industries.
Although it is a fool's dream for hawkish politicians in the United States to advocate a comprehensive decoupling from China, the decoupling of science and technology has already begun. The first way that comes to mind is to rob people in order to raise their wages by several times the number of times.
Ren Zhengfei went to universities in the southeast coastal region to recruit talent, because Huawei needed to add 30,000 engineers to the R&D team. Because what enterprises lack most at present is talent, and the war of chips has become a war of talents. One of the characteristics of high-end manufacturing is that it is closely related to the level of science and education, and its competition is also a competition for education and scientific strength between countries. At present, China's education system has not independently produced Nobel laureates in science and engineering disciplines, which means that we lack the ability to disrupt and innovate from 0 to 1. Where did Huawei's 5G come from? The technical principle was originally a hypothesis proposed by a Turkish scientist, and finally Huawei discovered and turned it into a product. Advanced technologies are first hypothesized by scientists, invented in the laboratory, and finally industrialized by entrepreneurs and scientists. China's industrialization is rapidly catching up with the West, and many links are still lacking. Due to poor basic research, the underlying hardware and software have to rely on others, which is a fundamental gap.
When the proportion of manufacturing in GDP peaks and begins to decline gradually, in order to maintain the momentum of industrial development, it is necessary to increase investment in research and development, so that innovation ability becomes the first driving force of industrial manufacturing. As far as a country and region is concerned, maintaining R&D investment exceeding 3%~4% of the output value of the manufacturing industry, maintaining basic research and innovation investment from 0 to 1 accounting for more than 20% of the total R&D investment, and maintaining the leading unicorn enterprises in manufacturing innovation accounting for more than 30% of the market value of the capital market are the signs of a manufacturing power.
It should be said that we are still catching up. Putting forward a new pattern of development does not mean that the problem has been solved, but that it is necessary to make necessary ideological and work preparations in order to cope with the complex and severe international environment.
Some economic experts have always emphasized giving play to our comparative advantages, but now our comparative advantages are getting weaker and weaker. If we continue to insist on our comparative advantages, the mainland will always be in the middle and lower reaches of the industrial chain, and high-end things will always be in the hands of others. High-end manufacturing industries such as chips have a huge market in China, and it is not a long-term solution to not develop themselves but be controlled by others.
The new crown epidemic has led to the restructuring of the global industrial chain, and the original industrial chain is to follow the cost, where the cost is low, and now it has become where to deploy safely. Of course, the principles of efficiency and profit are still at play. Although the U.S. government proposes decoupling, it may also compromise in the face of the huge economic interests of corporations. The so-called "Entity List" is for U.S. companies to apply for approval, and after approval, they can still do business. Since then, a large number of U.S. companies have begun to apply to the U.S. Department of Commerce. However, repeated games and gradual decoupling will be the general trend. China's science and technology circles and business circles should prepare for the worst.
- 04 -China's chip self-improvement road
In 1982, China established the Office of the Leading Group for Electronic Computers and Large-Scale Integrated Circuits, referred to as the "Daban". The establishment of a leading organization of the industry at the national level can be imagined as a matter of policy intensity. In terms of promoting the development of the chip industry, there was the 908 project in 1990, the 909 project in 1999, and the 18th document on the development of chips issued by the State Council in 2000.
Starting from the National High-tech Research and Development Program (863 Program), many policies and resources have been tilted towards the chip field. After the establishment of the top ten special funds, the No. 01 and No. 02 projects were given to the chip industry. In 2014, China Development Bank Capital, China Tobacco and other enterprises set up a "big fund" to support the semiconductor industry, namely the National Integrated Circuit Industry Investment Fund. According to media reports, the first phase of the fund raised 130 billion yuan. In 2023, the second phase of fundraising will begin, with a capital scale of 210 billion yuan, and will be managed under a shareholding system, with more than 20 shareholders, including the Ministry of Finance and China National Tobacco Corporation. In addition, many local governments have also listed chips as a key development industry.
Industrial policy is important, but policy is not omnipotent, and there are three problems. First, the policy is top-down, casting a net in an all-round way, and does not focus on breakthroughs in subdivided fields. And who will break through the subdivision? It can only be the enterprise closest to the market. The so-called subdivision is actually defined by the enterprise - when the enterprise becomes bigger, a subdivision is established, but the person who makes the plan does not know which subdivision can be bigger. Second, the fund is likely to be used for infrastructure construction by enterprises, especially large state-owned enterprises and central enterprises. Because the term of office of enterprise leaders is only one to two terms, if they engage in basic research and development and talent investment, it is difficult to see the effect during their tenure, and they would rather buy technology to make quick money. Third, most of the formulators and review experts of all relevant policies and projects are drawn from central enterprises and universities, and the committee members and experts cannot be completely neutral, they are both referees and athletes, and no one bears the ultimate responsibility for the project.
In 1982, the "Daban" plan determined the location layout of the chip industry of "two bases and one point". "Two bases" refers to Jiangsu, Zhejiang, Shanghai, Beijing, Tianjin and Shenyang, while "one point" refers to Xi'an. However, it was hard to imagine that Shenzhen and Chengdu would emerge at the time. Although the unified national system has certain advantages, the main body of technological innovation is always the enterprise. There is a pair of contradictions that are difficult to reconcile between the national system and the main players in the market.
In fact, the Chinese government has invested the largest amount in the chip industry in the world, surpassing the United States, Japan, South Korea and European countries, however, the effect is difficult to describe. According to relevant statistics, as of August 2020, nearly 10,000 enterprises in China have turned to the chip industry.
In the capital market, venture capital and private equity companies are looking for chip projects, and chip investment consulting institutions have sprung up. Experts, entrepreneurs, investors and intermediaries encourage and mislead each other. These investment companies invest money, tell stories, cash out, and exit, so that small and medium-sized investors see chips like leeks and sickles, and they are cut to the ground. However, the companies that raise money in the market are all new companies that only have beautiful presentations but no real products. Some companies even claim to be starting mass production with just a sample. Most companies lack upstream and downstream industrial chains, have low sales, no profits, and even no teams that are really engaged in chip research.
A few years ago, the dean of the School of Microelectronics of a university bought chips from the United States, and then asked temporary workers to remove the trademark and replace it with the logo of Hanxin, so as to defraud 1.1 billion yuan of research and development funds, and worse, this behavior delayed the development of many national projects and destroyed the industrial ecology. In 2020, the Wuhan Hongxin project with an investment of 15.3 billion yuan was suspended, and then declared bankruptcy, and the only 7nm lithography machine in China owned by it was mortgaged by the bank. Local governments and enterprises are eager to pursue achievements, ignoring the research foundation and team strength of chip companies, which has led to painful failures, which is embarrassing.
- 05 -Break through the stuck point of the inner circulation
There is no doubt that China must develop its own chip industry. In the future, there may be two chip supply chains in the world, one is China's independent chain, and the other is the chain of Western countries such as the United States. As for whether China's chip industry can succeed and to what extent, it is too early to say.
As early as 2010, China became the world's largest manufacturing country, but 10 years later, small chips are still the unbearable burden of the largest manufacturing country. If the world's manufacturing industry is compared to a giant, the United States is the brain, Japan and Germany are the heart, and China is the limbs, playing the role of undertaking basic work and labor-intensive work.
The economic and trade friction between China and the United States and the new crown epidemic have made us realize that we are in the middle and low end of the global value chain. China's manufacturing industry has the advantage of variety, but lacks the advantage of quality, has the advantage of cost, but lacks the advantage of technology, has the advantage of speed, but lacks the advantage of quality, has the advantage of product, but lacks the advantage of brand. Before the start of the "14th Five-Year Plan", the country began to investigate the technical bottlenecks of various industries, and the manufacturing industry was the field with the most technical bottlenecks.
In China's chip manufacturing industry, the operating system, the new material system, and the precision equipment system are all controlled by others. Similar situations in which key technologies and core components are controlled by others also exist in other fields, such as aero engines, sensors, ion diaphragms, high-pressure piston pumps, epoxy resins, etc.
Therefore, if the external circulation is really to be disconnected, China's internal circulation may also be an empty word. The internal circulation is the main body, and the external circulation is still essential. In addition, even if the internal circulation is the main body, it is necessary to prevent the formation of a new gap due to the poor external circulation. Especially once a global alliance is formed under the leadership of the United States, China's high-tech manufacturing, advanced manufacturing, and high-end manufacturing will face more and more challenges and constraints.
It is also necessary to guard against the fact that if the economy declines for a long period of time and the fiscal situation becomes more and more difficult, the policy effect of the national system will gradually decrease, and it will also be questioned by the international market. In particular, beware of competitive pitfalls such as chip competition. What is the competition trap? When we use the power of the whole country to catch up with the West in terms of industrial level at any cost, and the West suddenly opens up its market, then all the investment at any cost is tantamount to a waste. This is the case with chips, and the same is true for all the industries listed above that are "stuck", if all industries encounter competition traps, it will be a huge drain on the national strength of the whole of China. Competition traps must be guarded against.
The chip industry must not be separated from the global innovation chain. External circulation remains crucial. On September 11, 2020, General Secretary Xi Jinping spoke at a symposium of scientists on six suggestions for accelerating scientific and technological innovation, one of which is to "strengthen international scientific and technological cooperation". At present, what is still needed is "strengthening", not "weakening". It is still important to join, rather than depart from, the global system of division of labor and cooperation.
- 06 -It is inseparable from the support of basic research
Semiconductors are the "main battlefield" of the current Sino-US science and technology war. The world's semiconductor products with an annual output value of 600 billion US dollars cover thousands of chips and nearly 100,000 discrete devices, supporting various electronic products with an annual output value of several trillion US dollars in the downstream, as well as the digital economy with an annual output value of tens of trillion US dollars. It is estimated that every $1 of semiconductor products can drive $100 of global GDP.
On August 11, 2022, the United States announced an embargo on EDA software for next-generation GAA transistors to China, and the world's fundamental research results in semiconductor physics and microelectronics are integrated into the process design kit (PDK) of EDA tools. In the past, Chinese chip companies shared the results of global semiconductor basic research by purchasing EDA companies' PDK packages, Chinese policymakers, government personnel, and even the industry believe that the semiconductor industry can be developed without semiconductor basic research. Now, in the words of Academician Li Shushen, the United States has extinguished the "lighthouse" of academic exchanges in semiconductor physics, and we have entered the "dark forest".
The upstream of the chip is semiconductor physics, and semiconductor physics is the source of all semiconductor technology. The first quantum revolution gave birth to devices such as lasers and transistors, and produced semiconductor information technology including integrated circuits, optoelectronic devices, sensors, and discrete devices, and 11 achievements in the field of semiconductors won 9 Nobel Prizes in physics. At present, transistors are approaching the physical limit, and "Moore's Law" is about to fail, and it is urgent to break through the bottleneck of new materials, new structures, new theories, new devices and new circuits. A significant portion of the core patents come from the basic research results of semiconductor physics, and these achievements do not rely on the most advanced semiconductor manufacturing equipment such as EUV lithography machines. By vigorously strengthening the basic research of semiconductors and laying out a large number of patents in Europe and the United States around the materials, devices, and processes of next-generation transistors, we can set up "checkpoints" in chip manufacturing, the "throat" of the global semiconductor industry chain, and form countermeasures, which is expected to solve the problem of "stuck neck" in key core technologies of semiconductors. The semiconductor industry chain is long and extensive. Any material, a device, or even an accessory in the upstream of the semiconductor industry chain may become a part that restricts competitors. Even the United States, the birthplace of semiconductors, cannot independently build the entire semiconductor industry chain. To this end, the United States is eager to enlist Japan, South Korea and Taiwan to form the Semiconductor Quadripartite Alliance (Chip4) to improve the security of its semiconductor supply chain and curb the development of high-end chip industries in other parts of China. In the context of the Sino-US technology war and the "decoupling" of the industrial chain, it is difficult for us to enter the international supply chain even if we design or manufacture advanced chips. Through a large amount of investment in domestic substitution, we can only achieve internal circulation or narrow the gap with the United States, and still cannot change the "stuck neck" dilemma of "I have you and you have no me".
For a long time, the U.S. has invested more than twice as much in semiconductor research and development each year as the rest of the world combined. In 2018, the U.S. federal government invested $6 billion in semiconductor R&D, while semiconductor companies invested as much as $40 billion, which is close to the total R&D expenditure of 373.8 billion yuan from China's central government. Under the slogan "America's future depends on semiconductors", the United States passed the "CHIPS and Science Act" in 2022 with an investment of $280 billion, including about $52.7 billion in financial support for the chip industry from 2022 to 2026, $24 billion worth of investment tax credits for enterprises, encouraging companies to develop and manufacture chips in the United States, and providing about $200 billion in scientific research funding support from 2023 to 2027.
China's absolute investment in science and technology is huge, with 2.78 trillion yuan in 2021, ranking second in the world, accounting for 2.44% of GDP, which has great potential. However, basic R&D accounts for only 6.2% of total investment, and this proportion will reach 8% by 2025. In developed countries, it is 15%~20%. Moreover, under various statistical standards, the highest conversion rate of Chinese results is 30%, which is less than half of that of developed countries.
This article is excerpted from Chapter 6 of "The Key Leap from Big to Strong" and Section 3 "High-end: Starting from Chips" in "Critical Moment: Where to Go for China's Manufacturing Industry"