SMIC issued its latest announcement and signed a 1.2 billion US dollar agreement for the supply of lithography machines. In the face of major changes that have not occurred in 100 years, SMIC's overall strategy will have to be adjusted in order to adapt to the new living environment.

Based on the objective rules of the semiconductor industry and the conditions and constraints facing China, we believe that SMIC will undergo three major transformations:

1. From simply pursuing advanced processes -> returning to mature processes (“new 90/55nm” is far greater than “old 7nm”).

After SMIC entered the list of entities, the practical significance of 7nm, which is based entirely on American equipment, is far less than the mature process based on domestic equipment. Foundry is not the lowest level of semiconductor technology; it is only an integrator of chip equipment, materials, and processes. The main conflict of Chinese semiconductors has moved from a lack of advanced process training to a lack of domestically produced semiconductor equipment and materials.

China lacks advanced 14/7/5nm technology, but China also lacks mature 13um/90/65/55nm processes. Weierhauer's CIS chip (55/45nm), Megayi's innovative NOR (55/45nm), Hubing's fingerprint recognition (55nm), and Zhuo Shengwei/Srip/Shengbang analog chips (around 90nm) all require mature process production capacity. Looking at the current industry, China can fully replace photovoltaic, LED, and LCD panels with domestic production, and mature process chips can also be produced by domestic equipment, materials, and processes.

Taking advantage of their technological advantages, the US will continue to attack other countries and control high-end processes and advanced chip manufacturing, and this situation will not be reversed in the short term. We expect that the low-latitude domestic substitution process will continue, and that China will control pan-semiconductor technology from the bottom up. Before significant progress is made in the fields of root technology, such as equipment and materials, mature processes must be remanufactured, then advanced process barriers are gradually overcome upward, and spiral development.

Returning to mature process reengineering based on domestic equipment is the most realistic task for Chinese semiconductors right now. The practical significance of 7nm, which is based entirely on American equipment, is far less than that of 55nm fabs based on domestic technology. The practical significance of fabless, which serves these mature domestic processes, is also greater than many FinFET processes with a small customer base.

2. From simple external circulation -> internal and external double cycle (de-A, not localization).

The global technology landscape will be reshuffled, showing a return to the ancestors of anti-globalization. Even as strong as the US, it only participated in a small part of the semiconductor industry. China, Europe, Japan, the US, South Korea, and Taiwan each occupied an integral part of the industrial chain.

Semiconductors are an industry with a fully global division of labor. No country can achieve full internal circulation alone, so there is no so-called full-link localization of the semiconductor industry, and the basis for de-beautification and de-A in some key fields is to combine equipment and materials from Europe and Japan, and manufacturing in South Korea and Taiwan Province of China.

The US, on the other hand, is based on high-end manufacturing to make up for its shortcomings downwards. The third quadrant refers to Japan (materials), South Korea (storage), Europe (equipment), and Taiwan Province of China (OEM), which rely on their leading edge in segmented industries to become intermediaries for the external circulation of the global hard technology market, independently of the internal and external circulation of China and the United States.

Global hard technology is divided into three quadrants based on resource endowments and factor distribution for one's own development:

The first quadrant: dominated by the United States;

Second quadrant: dominated by mainland China;

Third quadrant: dominated by Korea, Japan, Taiwan Province of China, and Europe (intermediate medium).

Under pressure from the external environment, China's local Fabless and Fab are all facing an upstream supply chain crisis, but the path of China's independent development will not change due to external pressure. With the proposed internal circulation policy, in the future, China will use mature Fab as the foundation and carry out external circulation with the third quadrant.

Based on the objective rules of the global industry, we believe that the external cycle of technology between China and the US will continue in the following areas:

1. Equipment: External circulation of equipment is carried out with “intermediate media” such as Europe and Japan, but internal circulation of domestic equipment is required in the fields of American ETCH, PVD, CVD, CLEAN, CMP, ANNEAL, etc. (Beifang Huachuang, Yitang, Shengmei, Huahai, Wanye, Zhongwei, Zichun, Precision Measurement, etc.);

2. Materials: External circulation of materials with “intermediate media” such as Japan and Europe (large silicon wafers, photoresists, etc.), and internal circulation of domestic materials is carried out in various fields such as large silicon wafers, electronic gases, electronic chemicals, sputtering targets, etc. (Zhonghuan, Shanghai Silicon, Liong, Jacques, Jingrui, Jiang Feng Electronics, etc.);

3. IP/EDA: External circulation of the software ecosystem is carried out with the US (ARM, Synopsys, Cadence, etc.), but domestic internal circulation is also carried out in various new scenarios and new applications in the EDA and IP fields (Huada Jiutian, Core Vision, Guangli Wei, Xinhe Technology, and VeriSilicon Co., Ltd.);

4. Fab/IDM: External circulation is carried out with South Korea's storage chips and wafer foundry in Taiwan Province of China, and internal circulation continues in domestic fabs and IDMs that rely on mature processes (SMIC, Huahong, Changcun, Changxin, China Runwei, Silanwei, Jichuang).

Therefore, in the future, China will maintain minimal internal circulation and carry out independent innovation in underlying technologies (equipment, materials, EDA/IP) in mature processes. Looking back at China's pan-semiconductor development path, in the next few years, we believe that China's semiconductors will rely on external circulation on the basis of internal circulation as much as possible to achieve a dual cycle system.

3. Technological breakthroughs take the lead -> production capacity expansion takes the lead.

SMIC undertakes three tasks:

1. R&D and breakthroughs in advanced processes, such as FinFET 14/7nm;

2. Operation and production of stock production capacity, mainly at SMIC's various plants in Beijing, Shanghai, Tianjin, and Shenzhen;

3. Expansion of new production capacity, based on the expansion of production capacity based on existing mass production process technology, mainly mature processes.

For various reasons, the expansion of mature new production capacity, which should have been given first priority, has not been fully implemented. After several setbacks, SMIC's Beijing joint venture line finally received official funding at the end of 2020, but at this point, a global production capacity crisis dominated by mature 8-inch and 12-inch processes had already broken out.

China's huge Fabless industry is facing tight production capacity in global fabs, and there is no strong local capacity reserve. According to SMIC's prospectus and IC insight data, SMIC's monthly production capacity of 8-inch chips is about 400,000 pieces, far lower than TSMC's monthly production capacity of about 2.7 million pieces. According to Huahong's official website data, Huahong Semiconductor, which is ranked second in China, has a monthly production capacity of only 220,000 pieces.

Among the three tasks of SMIC in the future, meeting domestic fabless demand and providing safe and controllable OEM services by expanding production is the first priority, and maintaining the normal operation of existing production capacity is the second priority. As for breakthroughs in advanced processes, this is not within the reach of individual companies, but a systematic project that requires semiconductor equipment, materials, and IP/EDA manufacturers throughout the United Nations to cooperate and make a breakthrough together.

Therefore, Jiang Shangyi's return and SMIC's drastic upgrading of the Capex process are in line with this major trend in the industry, and are the right and inevitable choices for SMIC to face new international forms and conditionalities (sanctions on US equipment).

Risk warning: Technology research and development falls short of expectations; industry competition intensifies; trade friction between China and the US intensifies.