The semiconductor market is a global industry that is likely to reach $1 trillion in 2031 or 2032.
The US is the leader in processors and other key products. South Korea is the leader in memory products. European companies are leaders in power semiconductors, where electric vehicles are the big market. Japan is a key vendor base for image sensors in smartphones and autonomous driving.
China is projected to consume 47.5% of semiconductors globally in 2020. However, there are major changes in the consumption pattern of semiconductors in China due to the strengthening competitiveness of Chinese equipment vendors.
A perspective on the semiconductor consumption in China is shown in the following figure.
In 2006, 11.5% of semiconductors in the China market were consumed by Chinese equipment vendors substantially for personal computers and automobiles. Chinese companies will consume 50.0% of semiconductors in the China market in 2020, with smartphones being the largest segment.
In 2030, Chinese companies will consume 68.6% of semiconductors in the China market. While smartphones will continue to be a large market, the consumption of semiconductors in data centers related to artificial intelligence (AI) will also be large.
The semiconductor industry has undergone many changes in the past ten years and will experience more changes in the next ten years, with the increasing adoption of AI capabilities being the key driver.
The industry is characterized by high levels of innovation. It is also important to have a global perspective on markets, applications, and technologies to maintain high competitiveness levels. When a country focuses inwardly and does not operate within the global ecosystem, its competitive position weakens.
The analysis of the chip war involves a perspective of the situation in 2020 and where the industry will be in 2030.
Situation in the US
US companies, including Intel, AMD, and NVIDIA, are global leaders in CPUs and GPUs for computer applications. AMD and NVIDIA use TSMC and Samsung Electronics as wafer suppliers for their latest designs.
AMD also continues to use Globalfoundries for some of its earlier designs. While Intel has its own wafer fab facilities, the company has indicated it will use TSMC or Samsung for some of its new processor designs in the future.
Qualcomm is the global leader in chipsets for smartphones and obtains wafer supply from Samsung or TSMC for advanced designs. Apple also has smartphone chipset designs, with wafer supply from TSMC.
Broadcom is the global leader for Ethernet networking designs, with wafer supply from Asia. Marvell Technology Group is also in the networking market, with wafer supply from Asia.
Micron Technology is a competitive participant in DRAM and 3D NAND, and its wafer supply is concentrated in Taiwan, Singapore, and Japan.
Texas Instruments (TI) is a leader in analog products. Its wafer supply is concentrated in Texas for mature technologies, and the company also has wafer fab facilities in Germany, China, and Japan.
Globalfoundries has a wafer fab in Malta (New York), and Samsung has a wafer fab in Austin (Texas). However, neither location operates at the most advanced technology nodes.
The US has a number of other semiconductor companies that have limited wafer fabs or are fabless. While there is some wafer capacity in the US, such as at ON Semiconductor, there is extensive utilization of foundry capacity that is primarily in Asia.
While the wafer supply of US semiconductor companies is in Asia, a number of leading vendors for manufacturing equipment are in the US, including Applied Materials, Lam Research, KLA, Veeco Instruments, and others. The large market for these manufacturing equipment companies is in Asia.
Synopsys, Cadence Design Systems, Mentor Graphics (Siemens), and Ansys are the leading software companies for designing semiconductors. While these companies are in the US, they participate in global markets.
US companies within the semiconductor industry ecosystem participate in global markets, and it is important for them to have open access to customers on a global basis. The activities that reduce total available market (TAM) for US companies will increase the TAM for companies in other countries.
US semiconductor companies are expected to be globally competitive in 2030 if there are no major disruptions in markets and supply chains. It is critical for US companies to obtain wafers from Asia without supply chain disruptions in order to be competitive at the product level.
In many cases, the wafer value represents 35% of the final product value, and US companies want to maintain 65% of the product value.
Situation in China
Chinese smartphone vendors will supply 65.1% of total smartphones worldwide in 2020 and are consequently large users of semiconductors.
While Huawei Technologies designs its own chipsets for its smartphones, other Chinese smartphone vendors use chipsets from leading semiconductor companies outside of China, including Qualcomm and MediaTek. The wafer supplier for Huawei’s advanced designs was TSMC.
Similarly, the key DRAM vendors to Chinese companies are Samsung, SK Hynix, and Micron. China is investing in ChangXin Memory Technologies (CXMT), but its output is limited to date and will not likely change in the next five years. However, SK Hynix has established large manufacturing capacity in Wuxi (China) to support the China and global markets.
There is a similar situation in NAND where Chinese companies rely on products from Samsung, Kioxia, Micron, SK Hynix, and Intel. China is investing in Yangtze Memory Technologies (YMTC), but the impact of YMTC on the China market will not be significant in the next five years. Samsung is investing an additional $15 billion in Xian (China), which will have a significant impact on the supply chain for NAND in China as well as global markets.
China is expanding its wafer manufacturing capacity and providing funding of $10 billion to SMIC for capacity expansion in Shanghai (China). The purpose for this investment is to allow China to gain internal control of wafer fab capacity even though the technology of SMIC is two to three generations behind TSMC.
Also, SMIC will not be able to produce wafers at 5nm (in Apple’s upcoming iPhone) because the US is blocking access to the extreme ultraviolet (EUV) lithography scanners from ASML.
Putting SMIC on the US entity list will slow China’s ability to build up its own supply of wafers. However, SMIC has approximately 5% of the global foundry market, and other foundry vendors will be able to replace SMIC in 18 to 24 months.
China does not have competitive wafer manufacturing technology. Although Advanced Micro-Fabrication Equipment (AMEC) has high market valuation, the company will not likely become competitive with the equipment manufacturers in the US, Europe, or Japan over the next five or potentially ten years.
There is a similar situation in design software for semiconductor products. The gap between China and US vendors, such as Cadence and Synopsys, will not close in the next five or potentially ten years.
China is building a strong electronics equipment industry, including 5G smartphones and 5G infrastructure, personal computers, televisions, autonomous vehicles, electric vehicles, robots, high-speed trains, and data centers. For China to succeed with its growth strategies for electronics and AI, it needs the semiconductors that are produced outside of China.
A perspective on semiconductor supply in China is shown in the following figure.
The analysis indicates that non-Chinese semiconductor companies (Samsung’s NAND facility in Xian is non-Chinese) will supply 81.0% of its semiconductors in the China market in 2020. Even in 2030, China will buy 60.2% of its semiconductors from non-Chinese semiconductor vendors.
As a result, it is important for Chinese companies to have access to semiconductors from multiple sources. It is also important for non-Chinese semiconductor vendors to have access to the China market.
Resolution of trade issues
China is expected to increase funding for semiconductor products. While there has been mixed success over the past ten years, its funding will accelerate due to the trade issue with the US.
The expectation is that China will spend $500 billion over the next ten years ($50 billion per year) to strengthen the competitiveness of its semiconductor industry, which includes establishing competitive supply chains as well as supporting companies that develop new products.
The Science and Technology Innovation Board (STAR market) of the Shanghai Stock Exchange is also a good source for funding new companies that develop new high-technology products.
Additionally, a significant amount of funding will likely encourage non-Chinese companies to establish manufacturing capacity in China, continuing manufacturing patterns of the past such as with Samsung, SK Hynix, and Intel.
The China market is important for the U.S. semiconductor industry. Access to the China market is expected to be conditional upon the US government loosening restrictions on supplying products to Chinese companies.
The China market is also important to companies in South Korea, including Samsung, SK Hynix, Hyundai Motor Group, LG Electronics, and others. South Korea will not likely allow the market for its major corporations to be impacted by the activities of the US and will protect its companies and consumers from the activities of China.
There is a similar situation with European companies. Access to the China market for STMicroelectronics, Infineon Technologies, NXP Semiconductors, Volkswagen Group, Siemens, and other European companies is impacted by the activities of the U.S. government.
US companies and consumers also make large purchases of goods from China, which includes smartphones, personal computers, televisions, and many other products. It will be a major challenge to set up supply chains for these precision products in other countries due to the highly automated factories in China and the high quality of many products as well as low cost.
The ongoing attempts by the US to block the activities of China have had a negative impact on both US and Chinese companies. To date, the US government continues attempting to contain China’s activities because the Chinese government has not made any direct resistance to the US activities against Huawei, TikTok, and others.
It is, however, not realistic to expect this situation to continue. China will take both direct and indirect actions against the US in the future, which will significantly increase the trade war’s intensity. The key issues are the timing of China’s activities and the approaches that China will take against the activities of the US
One possible scenario is that the world will be segmented into countries that ally with the US and those that align with China. This fragmentation may occur in 5G where there may be two global standards, with the main difference being in software.
While the US is ahead of China in some areas such as AI in 2020, China is making much greater investments than the US in new technologies.
The technology for designing and manufacturing high-speed trains in China is significantly ahead of the US, large battery manufacturing capacity is being built in China for electric vehicles, robotaxis are increasingly used in China.
The deployment of 5G networks in China is far ahead of the US. As a result, China will be significantly ahead of the US in many areas, including AI, in 2030.
While the trade war between the US and China is accelerating China’s investments in semiconductors and applications for semiconductors, it is not resulting in the same sense of urgency in the US. The US government is not supporting the acceleration of new product development or strengthening the manufacturing supply chain.
The US regards China as a competitor but is not making the necessary investments in technology and manufacturing capabilities in the US to be superior to China. In 2020, it is not too late for the US to drive technology leadership in strategic areas, with semiconductors and AI being key areas.
Based on present US strategies, however, China will be significantly ahead of the US in many areas by 2030. Then, it will be too late for the US to close the technology and manufacturing gap with China.
The attempts by the US government to slow down China may be effective in the short term. China, however, will counter the US activities by imposing limitations on the products from US companies in the China market.
A significant escalation in the US-China trade war will be detrimental to semiconductor companies in the US, Europe, South Korea, Japan and China.
It is important to resolve the trade imbalance between China and the US, with the understanding that US businesses and consumers choose to buy Chinese-manufactured products. The US needs to build an industrial base where more products can be sold to China as well as supporting US consumers.
It is also important that the data in the commercial and consumer ecosystems is allowed to move globally without government restrictions. This is one key area where the US needs to negotiate with China, but it must be done with a win-win perspective rather than using approaches that may be considered arbitrary.
The trade war within the semiconductor field is the wrong strategy for the US because China will use the chip dispute to accelerate the development of its internal supply and to build an ecosystem of countries that want to be part of the China market. Even though the US has military and financial power, China has an extensive market and large financial resources.
Conflict wastes resources. Competition stimulates creativity. There is high value for the US to reassess its strategies and focus on areas where it can have global leadership in key semiconductor and electronics industries, especially as AI becomes increasingly important.
The US needs to regard China as a strong competitor that it can surpass with the drive of US entrepreneurs.
Handel H Jones is chairman and CEO of International Business Strategies (IBS) Inc.