Academician Wu Hanming: China's chip investment is far from overheating

Recently, Academician Wu Hanming pointed out that since China's chip technology is still a certain distance behind that of the United States, China's chip investment is far from overheated, and the real chip manufacturers are still in short supply. If development is not accelerated, the gap between China's chip production capacity and demand will widen to at least the equivalent of eight SMIC production capacities in the future.

Over the past 50 years, the integrated circuit industry has followed "Moore's Law": the number of transistors integrated on a chip will double every 18 to 24 months while the price remains unchanged, and computing costs are decreasing exponentially. Moore's Law is still supporting the development of new technologies such as 5G and artificial intelligence, but there is a limit to the development of things. As the process moves from micron to nanometer, the number of atoms in transistors is decreasing, and various physical limits restrict the further development of Moore's Law.

"With the evolution of process nodes, Moore's Law is becoming increasingly difficult to sustain." On April 24, at the "Chinese Academy of Engineering Information and Electronic Engineering Frontier Forum" co-organized by the Chinese Academy of Engineering journals FITEE and Engineering and hosted by Zhejiang University and Tsinghua University, Academician Wu Hanming, chief innovator in the field of Zhejiang University Hangzhou International Science and Technology Innovation Center and dean of the School of Micro-Nano Electronics, told the media including China Business News that the "post-Moore era" has officially arrived. In his view, "industrial technology is not the application development after the transformation of scientific research institutions, but one of the original driving forces guiding scientific research." In key core technology fields such as integrated circuits, it is necessary to supplement the construction of a scientific and technological culture oriented towards industrial technology, so that technology research and development and market applications can "mutually leverage each other."

Many industry insiders at the meeting also said that the open development and win-win cooperation of the integrated circuit industry is an inevitable choice. In particular, the development of key equipment involves many scientific, technological and engineering fields. We cannot rely on single-point breakthroughs on the "beach", but should integrate the efforts of the entire industry chain.

In his speech, Wu Hanming pointed out that China's integrated circuit industry is facing two major barriers: policy barriers and industrial barriers. The former include the Batumi and Wassenaar Agreement, and the latter is reflected in the fact that the world's leading semiconductor companies have accumulated rich intellectual property rights thanks to their early layout, forming a patent moat for latecomers such as China's semiconductors. These have posed huge challenges to China's semiconductors. The integrated circuit industry chain has many links, and the three key "neck-stuck" manufacturing links include process, equipment/materials, and design IP core/EDA. Wu Hanming mentioned that there are many shortcomings in the development of domestic industries. For example, in the field of testing equipment, Chinese companies rarely get involved, so the development of domestic industries in this field is almost blank; in terms of semiconductor materials, my country's photoresists, masks, and large silicon wafers are almost all dependent on imports; in the field of equipment, Chinese equipment is not seen on the world stage.

Take EUV lithography as an example. It involves more than 100,000 parts and requires the support of more than 5,000 suppliers, of which 32% are in the Netherlands and the United Kingdom, 27% are in the United States, 14% are in Germany, and 27% are in Japan. This actually reflects the result of global technology collaboration. Wu Hanming mentioned that autonomy and control are important, but at the same time this industry is also a global industry. Among them, "Which links can we control? It is the core point of our country's R&D and industrial development." Wu Hanming's topic about lithography also triggered the emotion of Zhao Haijun, CEO of SMIC at the scene.

Zhao Haijun mentioned that Martin van Der Brink, chief technology officer of Dutch optical technology company ASML, had visited Shanghai before and told him that "the gap is about 20 years". But Zhao Haijun believes that "a 20-year gap does not require 20 years to catch up. It is difficult to surpass it all at once, but it can be advanced little by little each time." Zhao Haijun mentioned that another reason why ASML is so great is that others have stopped doing this field, but it still insists on doing it. Academician Wu Hanming took the current chip manufacturing process as an example and talked about the three major challenges currently faced. Among them, the basic challenge is precision graphics; the core challenge is new materials. He mentioned that each material requires thousands of process experiments, and the support of new materials requires performance improvement; the ultimate challenge is to improve the yield. In addition to technology, wafer fabs, R&D and design costs are also another major challenge facing the chip industry. "Although the difficulty and cost of chips have been increasing, the slowing Moore's Law brings opportunities to the pursuers," Wu Hanming said. Under these challenges, the combination of advanced system structure, special processes and advanced packaging in chip manufacturing can make my country have great potential in the field of chip manufacturing. He cited data saying that advanced production capacity below the 10-nanometer node accounts for 17%, and 83% of the market is at nodes above 10 nanometers, leaving huge room for innovation.

In the case of not having an advantage in advanced process research and development, my country can use mature processes to improve the performance of chips. It is for this reason that he put forward a point of view: local controllable 55nm chip manufacturing is more meaningful than the completely imported 7nm. Wu Hanming also made suggestions for the development of my country's semiconductor industry. For example, it is necessary to establish a scientific and technological culture oriented towards industrial technology, and the technological achievements are entirely dependent on market appraisal, and to accelerate the construction of public technology research and development platforms under the national system, give play to the advantages of "concentrating forces to accomplish major tasks", and further build a global innovation path that is conducive to my country's sustainable development in this field. It has become a key task.

He also mentioned to reporters that "scam-style" chip investment may be overheated, but the actual chip manufacturers are still in short supply. China's chip investment is far from overheated. In particular, we must pay attention to the localization of the industrial chain and increase manufacturing capacity, at least at a higher growth rate than the global level.

Zhao Haijun also agrees that the current chip overheating is not the essence. According to statistics, the top five wafer fabs in terms of production capacity include Samsung, TSMC, Micron, SK Hynix and Kioxia, and no company from mainland China is among them. With the development of technology, the demand for chip production capacity will become higher and higher, and the gap between domestic chip market demand and manufacturing capacity is still large.

Wu Hanming previously said: "If development is not accelerated, the gap between China's chip production capacity and demand will widen to at least the equivalent of eight SMIC production capacities in the future."