China's Future Investment Watch series - China Chips: Will China Dominate the World Semiconductor Market in 5 Years?
Antitrust, reform in education, real estate and medical care policies, Common Prosperity... all of these show that China is undergoing earth-shaking change. These changes have brought large uncertainty, making many investors afraid to invest in Chinese projects and companies. However, over the next decade, China sure will become the world's largest economy. How to better understand the opportunities and risks of the Chinese market and deal with certainty and uncertainty is a crucial problem. EqualOcean launched a series of research, China's Future Investment Watch, hoping to provide clues for global investors.
Recently, a trending voice on the Chinese Internet has been suggesting that Chinese netizens should calm down and face up to the gap existing in the global chips market between China and developed countries, rather than blindly think that China will catch up with and surpass these countries in a short time.
As mentioned above, we believe that China will, finally, do all of these and take the shares of neighboring countries, Europe and the United States, to realize a domestic only industrial chain. We will prove our point of view from three aspects: the current situation of all manufacturing links in the industrial chain, China's opportunities, and China's upcoming challenges.
Nine links of the industrial chain
The key to predicting the window period is to clarify the gap between each link of the chip manufacturing industrial chain, especially the advanced nodes. The whole wafer manufacturing is divided into 9 manufacturing links, including wafer cleaning, diffusion, CVD, mask Aligner, coater & developer, etching, ion implantation, CMP, and testing. Different links have different paths and progress.
1. Wafer cleaning system gap：4 years
Japan has the most advanced wafer cleaning system capability and takes over 65% of market shares globally. Tokyo Electron (TEL) launched CELLESTA SCD single wave cleaning system in 2021, creating a pattern collapse-free drying method that uses a supercritical fluid and can well match the 5/7nm node production line. China's most advanced cleaning machine can match the 14nm production line. The exciting news is, at present, cleaning systems under development is for 5/7nm node. In May 2021, ACM Research released a wafer cleaning system for advanced node applications named ultra-C WB system, which is equipped with the SC1 and SC2 Mega sonic functions and matching 14nm node line. Referring to the development history of Japan, we believe that the gap between China’s manufacturing level and the world’s advanced node level is four years.
2. Diffusion equipment gap：5 years
In the oxidation diffusion equipment sector, Japan is still the most advanced country with a technical level also matching 5/7nm node line, accounting for about 80% of the global market share (represented by Tel and Hitachi Kokusai electric). Kokusai Electric DD-802V vertical diffusion furnace is one of the most advanced vertical diffusion equipment, which can perfectly match the 5/7nm node line. At present, China's NAURA Technology Group can independently produce a 14nm oxidation diffusion furnace to match the 14nm chip manufacturing line. We expect that it will take China five years to reach an advanced technical level and realize domestic replacement.
3. CVD equipment gap：9 years
The United States and Japan are the two main players in the CVD market, both of which can match 5/7nm node technology, accounting for about 75% of the global market share. It is also one of the links to China that depends heavily on imports with a domestic rate of about 2% in the beginning years of the 2010s. However, NAURA Technology Group has made breakthroughs in recent years. Equipment matching 28nm node line has been successfully introduced to the customers, and a 14nm node product is on a trial run. In addition to NAURA Technology Group, Piotech is also on the road, aiming for 5/7nm technology. Considering the running-in period and research progress in and out of China and the concentration of global market shares, we believe that it will take nine years for China to catch up with the main players.
4. Stepper gap: at least 12 years
In the sector of stepper, ASML in the Netherlands is a monopoly player, taking 84% of the global market share by itself, and, notably, the stepper market contributes at least 20% of the whole profit of the whole chip industrial chain. It is the only company in the world that can match 5/7nm or even 3nm nodes lines in the future. It is worth noting that the Netherlands also has no way to create stepper independently because all components are being controlled by other countries: the light source is from the United States; the lens is from Japan; the nano-CNC mechanical system is from Germany, restricting and promoting each other. At present, the most advanced company in China in the stepper sector is Shanghai Microelectronics. At present, it has successfully broken through the technology of 90nm and can produce 28nm chips through a multi-exposure method, which is costly and cannot achieve high production. The technology under research is 65nm, but it is far from the advanced technology level of ASML. The node gap is massive (at least 3 node levels, including 28nm, 14nm and 7nm) and even they use different technics (ArF vs. EUV). If China wants to be a market leader in this area and realize full-domestic-replacement, it needs to build production lines for all the components that came from the United States, Japan and Germany, instead of only creating a stepper assemble line like ASML did. Considering the huge gap and the difficulties involved in breaking through, we conservatively estimate that it will take China at least 12 years to reach the world advanced level – 15 years may in fact be a better estimation (it usually takes 3 years to build a new components line in chip areas).
5. Coater & Developer (Track) gap: 8 years
In the global coater & developer equipment market, the market share of Japanese TEL company has reached 88.7%, reaching an absolute monopoly position. The TEL LITHIUS Series is a leading-edge industry-standard system with high reliability and productivity, which well matches the 7/5nm nodes line and is favored by worldwide chip manufacturers. However, China has made a breakthrough in this area. Shenyang Kingsemi Equipment is researching the manufacture of domestic alternative products and has reached the process level of 28nm nodes, with 14nm nodes also being a target.
6. Etching machine gap: no gap
Etching machines are the strongest link in China's current chip industry chain. It is at the forefront of the world. The representative company is China's Advanced Micro. The company is able to produce etching machines matching a 5/7nm node line. The 5/7nm level etching machine (Primo D-RIE) of China Advanced Micro has been successfully launched and has even become an equipment supplier of TSMC. Moreover, the company is still researching at the front line, and the technology of 3nm node level is under development. Notably, like the stepper, etching machine market is also playing a key role in the industrial line, taking 20% of the profit the line creates.
7. Ion implanter gap: 10 years
China faces great difficulties in the field of ion implants, mainly because American products in this field occupy an absolute monopoly position and have over 90% of the global market share. The representative companies are Applied Materials and Axcell in the United States. Applied materials' VIISTA 3000XP deliveries the angle precision required for advanced node process need, based on its unique claimed variant dual magnet single water architecture. Compared with the etching and stepper markets, the ion impactor market is relatively small: the global market sales of the whole of 2020 are only about USD 1.8 billion. At present, Beijing Zhongke Xin Electron Equipment and Shanghai Kingstone Semiconductor in China have successfully produced 28nm node ion implanter and 14nm node technology is being developed recently. Considering the market share, Sino-US technology blockade and the research process in China, we believe that the development process will take about 10 years.
8. CMP gap: 5 years
The most advanced country in the CMP market is the United States, and the technical level has reached the matching 5/7nm node line. The United States accounts for 71% of the global market share. Leading by DuPont (the U.S.), whose Visionpad portfolio is designed for advanced 7/5nm nodes processes, offering a higher removal rate. In addition, the United States and Japan together account for about 98% of the global market share. However, China's Hwatsing Technology and China Electronics Technology Group have launched the technology matching the 14nm production line and are developing the 5/7nm technology, with only one node gap. Considering the research progress of China, the United States and Japan, we believe that it will take about five years for China to reach a truly world-class level in this field, but it may take longer to occupy more shares in the market.
9. Testing equipment gap: 8 years
Testing equipment runs through every process of chip manufacturing industrial line. Early detection of bad chips and elimination can greatly improve economic benefits. Therefore, the global market of this link is massive and reached over USD 10 billion. In this field, the most advanced country is still the United States, even reaching 3nm node technology level. The representative companies are KLA Tencor, ADVANTEST, Teradyne, and Agilent Technologies, accounting for 90% of the market share globally. At present, China has launched equipment matching 28nm node (Skyverse Technology and Shenzhen Chuangniu Technology) and is developing 14nm technology. Due to the technical blockade and China's research progress, combined with the analysis of market share, we think it will be difficult for China to catch up and surpass other countries in less than 10 years.
In fact, the weakest link of the industrial chain often determines the advancement of the whole industrial chain, especially for China, which wants to realize the domestic substitution of the whole industrial chain. This is also the fundamental reason why we believe that it will take 12 to 15 years for China to achieve the world's advanced level and self-sufficiency, that is, as we all know, the stepper, as the weakest link, determines the length of this process. However, as for the etching machine, which is as important as the stepper link in the line, the etching machine technology independently developed by China has reached the world advanced level of 5/7nm in six years, successfully breaking the blockade of western developed countries (the blockade was lifted by the United States in 2015). This also makes China’s market actors full of confidence in China's chip market. In addition to the analysis of various industrial chain links, a once-in-a-lifetime era is also the key.
Golden age’s coming
China's once-in-a-lifetime opportunity period is mainly caused by the failure (or slowdown) of Moore's Law. Moore's Law is the observation that the number of transistors in a chip doubles about every 1.5 to two years; it was created by Intel's co-founder Gordon Earle Moore and the semiconductor industry regards it as the golden rule. However, Moore's law is approaching the limits of physics. In the 2000s, the performance of global chips increased by about 50% every year, the rate slowed to 23% in 2010 and to 12% in 201. The average rate in the recent five years is as low as 3% per year. As early as 2014, Moore's law had an omen of failure: in terms of unit cost, when the chip process evolved from 28nm node to 20nm node, the unit price increased by 10%.
At present, the development of the chip industry mainly depends on increasing the transistor density. Jim Keller, the lead architect of the AMD K8 micro architecture, once said that along the process evolution path of FinFET (three times the density), GAA (two times), stacked nanowire (two times), wafer-to-wafer (two times), and die-to-wafer (the 3D stacking bonding technology, two times,), there is still 50 times the density of silicon-based semiconductors (based on Intel's 10nm node technology). However, after 3nm, the process with stepper as the key link currently forms 20 nm graphics with 193nm wavelength, which faces great physical challenges (the density change diagram is shown in the figure). Therefore, the subsequent density increase will actually come from the stacking of various levels, that is, the transistor stacking in various dimensions, which will increase the transistor density in a disguised manner.
In addition to increasing density, changing the semiconductor material itself is another strategy. As the silicon-based chip manufacturing process gradually approaches the physical limit, major chip-related companies have no clear answer to the trend of chips below the 3nm node. A major potential substitute is the chemical compound chip. At present, there are three strong competitors: GaAs, Gan and SiC. However, none of them can have the characteristics of high-temperature resistance, good radiation resistance, convenient fabrication, and good stability of silicon-based semiconductors. It is worth noting that graphene chip (carbon-based chip) is also a potential alternative mainly discussed. Compared with the current mainstream silicon-based materials, the thickness of single-layer graphene is only 0.27nm, the electron mobility of crystal structure is high, and the electron moves 1,000 times faster than that of silicon chips, up to 1/300 light speed. Graphene can also make flexible films, so that electronic products may be able to bend in the future. In 2020, Shanghai Institute of Microsystem and Information Technology unveiled the world's first carbon-based wafer capable of mass production: 8-inch graphene single wafer. This brings the possibility of large-scale mass production of carbon-based chips in China. However, according to the research of SMIC, it is not possible to create another innovative semiconductor road with the help of graphene wafers at this stage. Peng Lianmao, a leader in the Chinese graphene chip area, also expects that graphene chips will not become the mainstream in the next 15 years (silicon-carbon fusion technology may become the mainstream in the next 10-15 years).
The iteration speed of the same process has greatly slowed down, and the research on new materials will make it difficult to breakthrough in the short term to provide a golden age for the development of China's chips industry. Combined with the four decisive factors for China's success mentioned in our previous article, we believe that if we maintain active and high-intensity research investment and effective policy support, China will reach the first tier in the global chip market and realize all-domestic coverage in 12-15 years.
China + advanced-tech = historical opportunities
As Lu Qi, a Sino-U.S. software executive and founder of Miracleplus, said, the development of modern human science has entered the fourth Paradigm era (1+1 > 4), a data-driven era, exploring the cutting-edge of science and technology through computing power. In 2020, there was an undeniable trend that the global center was shifting to Asia, not only the economic center but also the science and innovation center. The rise of the United States and its financial founding is a historical accident. China's economic development is more like other parts of the world, from Agrarian society to Industrial society, and then to the current Information society. Therefore, it is easier for Chinese innovation to spill over to other countries except for North America. After the commercialization of Chinese innovation results, it will be easier to go to countries and regions in Southeast Asia, Europe, Africa, and Latin America.
From the perspective of market, economy, and development, we think an unheard-of opportunity is coming. To win in such a historical period, China has no choice but to seize the keynote of the era, that is, technology-driven advancement. China realized the importance of independent technology very early (even earlier than the United States), especially in the chip sector. Therefore, we are strongly optimistic about China's chip industry. But we still firmly believe that blind self-confidence cannot solve practical problems. If China plans to lead the development of advanced nodes in the global chip sector and occupy advancing market shares, we think 12 to 15 years in the future is a realistic timeline.
EqualOcean operates offices in Beijing, New York, and Shanghai. We welcome investors interested in the Chinese market to contact us via (contact@EqualOcean.com) or visit our offices. We believe the exchange of views will make you have a clearer prediction of the future.