Zhu Jing from Beijing Semiconductor Industry Association: The localization of integrated circuits will be affected by six factors

On June 25, the fifth Jiwei Semiconductor Summit was officially opened in Haicang, Xiamen. The theme of this summit was "The Heart and the Core are in the Same Form, and Changes are Like Kunpeng", which lasted for two days. At the first day's analyst conference, Zhu Jing, deputy secretary-general of the Beijing Semiconductor Industry Association and senior economist of Beijing International Engineering Consulting Co., Ltd., analyzed the factors and progress of localization substitution in the current integrated circuit industry chain.

During the just-concluded 13th Five-Year Plan period, the scale of my country's integrated circuit industry achieved rapid growth, with the overall scale achieving an average annual compound growth rate of nearly 20%, and the average annual compound growth rates of the three industries of design, manufacturing, and packaging and testing reaching 23%, 23%, and 13%, respectively. The independent research and development capabilities in key integrated circuit products, equipment, and basic materials have been continuously improved.

According to Gartner's forecast, China has already achieved about 10% of the global market share in cellular baseband chips and consumer electronic chips such as CIS, AP, and smart cards. In the next three years, advanced packaging will also achieve this result. NAND storage will take five years, and DRAM, MPU for computers and servers, advanced processes, semiconductor equipment/materials/equipment components, and EDA tools may take ten years.

Zhu Jing pointed out that the localization of integrated circuits in my country will be affected by six major factors, including geographical factors, technological innovation and upgrading, super-large-scale market, semi-monopoly industry, spillover of top talents and industrial chain coordination.

Geopolitical factors. The Atlantic Council, a US think tank, released this year a report titled "The Geopolitical Impact of New Technologies and Data," which states that geo-technology refers to a technology that is important enough to affect the functioning of society, and will be considered as "geo-technology." In the next decade, the world will usher in the "GeoTech Decade," which will involve six major areas: digital economy enabling technology, intelligent system technology, global health and wellness technology, technology that expands the boundaries of human, corporate and government activities, human enhancement technology, and future basic technology.

The current integrated circuit industry is regarded as "geo-technology" by the United States because of its importance as a basic technology for the future. The Sino-US dispute has provided external conditions for geopolitical factors to drive the localization process of integrated circuits. The localization process in upstream fields such as semiconductor equipment and EDA, where American companies have strong monopoly, high technical barriers and scarce talent, is more obviously affected.

Zhu Jing said that my country's foundation in the above-mentioned fields, especially in photolithography machines, coating and developing machines, testing equipment, advanced process ion implantation machines, as well as sub-sectors such as photoresists and silicon wafers, is poor, the task is heavy, and R&D is difficult. In addition, the investment cycle is long and there are few talents. It may take more than ten years for the domestic substitution cycle to occur. The national system needs to work tirelessly with the spirit of nailing down the nails.

Technological innovation window factor. As Moore's Law is getting closer to the physical limit and process technology is facing huge challenges in lithography, new materials and processes, process fluctuations, new structures, process integration and yield improvement, the new device technology of the post-Moore era marked by FinFET technology, which can continue Moore's Law and solve the problems of short channel effects, high leakage circuits and 60mV/dec subthreshold swing limitations faced by nanoscale devices, has arrived. Zhu Jing generally divides the post-Moore era into three stages: the continuation of Moore's stage with transistor structure, process and material innovation as the main focus; the extension of Moore's stage driven by application-driven heterogeneous integration; and the Beyond Moore era with new devices, new principles and new materials application innovation.

In recent years, my country's research and development level in advanced fields has gradually improved. For example, last year, a total of 23 domestic papers were selected for ISSCC, ranking third in the world. More than a quarter of these papers were completed by academics and industry, a proportion that exceeded that of South Korea and the United States. This means that my country is expected to catch up with the window of opportunity for technological innovation and replacement in new devices, new principles, and new materials in the future.

Ultra-large-scale market factors. China has the world's largest demand for branded complete machines, and has the right to define scenarios or product pricing in the fields of ultra-high-definition video, consumer wearables, portable energy storage, and pan-security. It has fast technology iterations, many micro-innovations, and strong integration and solution capabilities. The localization of chips for the Chinese market will benefit from the acceleration of the internal circulation, and domestic substitution will be quickly achieved within 1 to 3 years, or domestic chips will be dominant from the beginning.

Industry monopoly advantage. The new infrastructure industry (innovation) will significantly increase the localization rate. Thanks to the trend of system + chip collaborative development, supply chain security (actively cooperating with small and medium-sized enterprises), ecological awareness and other factors, the localization rate in power semiconductors, home appliance-related chips, smart grid-related chips, 5G-related chips and other fields will be greatly increased.

Outflow of top talents. In recent years, there has been a significant outflow of talents in some fields that have long been monopolized by overseas giants, have high technical barriers, and are extremely short of domestic talents, such as GPU, EDA, simulation, RF (filters), equipment (front-end measurement, lithography components), etc. This is due to the shrinking domestic layout of overseas companies due to domestic substitution, the impact of the epidemic, the wealth effect brought by the Science and Technology Innovation Board, changes in overseas companies' mergers and acquisitions in China, and the integration of domestic companies.

One problem that the promotion of localization based on the spillover of top talents will face is the dispersion of technical forces. Taking the GPU field as an example, in recent years, the talents spilled over from Silicon Valley Nvidia, Qualcomm, HiSilicon and domestic AMD branches, Zhaoxin (S3) and HiSilicon have spawned a group of GPU entrepreneurial teams such as Denglin, Tianshu Zhixin, Muxi, Moore Thread, Haifei Technology, and Biren. These companies can help China achieve a breakthrough from 0 to 1 in GPU to a certain extent, but the difficulty of replacement in these fields is often not solved by talent alone.

Industrial chain synergy. The localization of some industrial links requires the localization of the downstream of the industrial chain to be improved first, which greatly increases the difficulty of domestic substitution. For example, lithography machines and photoresists need to be used together, so there is a certain synergy between the two in new product development and product sales. Therefore, the prerequisite for photoresists to be well done is that domestic lithography machines must also reach a certain level. Other similar fields include key IP, equipment parts, etc.

Zhu Jing believes that driven by the above six factors, the process of localization substitution of integrated circuits in various fields will have different degrees of rhythm. Under the influence of ultra-large-scale markets and semi-monopolistic industries, consumer electronic chips with low thresholds, industrial and rail transit chips, automotive semiconductors, 5G base station chips and other fields will be the first to achieve localization substitution in the next five years; under the influence of technological innovation and replacement factors, intelligent computing, new storage, AI+EDA, advanced packaging, and wide bandgap semiconductors with high thresholds will achieve localization substitution in the next 5 to 7 years; in the fields of GPU, EDA and IP, RF front-end modules, photoresists, semiconductor components and enterprise-level SSD master control, which are influenced by technical geography, the need for industrial chain collaboration, and talents, semiconductor equipment such as advanced processes, front-end measurement/ion implantation/lithography machines, and semiconductor materials such as silicon wafers, chemicals, and special gases will require 7 years or even more than 10 years of localization substitution.

Faced with the fact that the localization process in different links of the industrial chain is affected by different factors, Zhu Jing believes that industrial technological innovation under restrictive conditions is sometimes more difficult than original innovation. She puts forward several suggestions in this regard.

First, during the 14th Five-Year Plan period, we will increase attention to technology and engineering innovation in the integrated circuit industry, and promote integrated circuit companies to truly become the main body of technological innovation and become the carrier of "proposing demands, increasing investment, organizing research, and leading application transformation."

Second, establish a relatively independent disruptive technology innovation plan outside the existing science and technology plan management system to support various types of innovation entities to carry out more challenging and high-risk innovation activities to discover technologies that can bring fundamental changes to the integrated circuit industry.