At the end of each year, we will make some unsupported outlooks and opinions on the development of the global and my country’s semiconductor industry in the coming year, and continue this year. The year 2019, which is about to pass, should be a difficult year for most semiconductor companies in the world. The industry as a whole is hovering at a low level. The major application markets such as AI and automobiles in the past two years have been weak and sluggish, superimposed on Sino-US trade. The friction continues to escalate, the Trump administration has begun to deeply adjust its science and technology strategy towards China, and forcibly promoting the “decoupling of science and technology” between China and the United States has become one of the important policy options. Characteristics of the prosperity of the semiconductor industry. The global semiconductor industry situation in 2018 and 2019 is a double sky, and the regional market is also showing a very differentiated market. However, judging from the current data guidance in the third and fourth quarters, the overall development direction of the semiconductor industry in 2020 is biased. Optimistic, here is also an outlook and prediction:
(1) The global economy is warming up, but it is difficult for American semiconductor companies to avoid a decline in development.
After the entire industry experienced a severe recession in the first half of 2019, it has been developing towards a steady recovery and growth since the third quarter. The price of memory has stabilized, the utilization rate of foundry, packaging and testing capacity has increased significantly, and the data of major leading companies have also continued. The rebound, and from the downstream point of view, such a state has a certain degree of continuity, so it can be seen that the signal of the global semiconductor industry’s recovery in 2020 is very obvious. However, from a regional perspective, the follow-up development of the US market may not be optimistic. Among the regional market data given by WSTS, in 2019, the market demand in the semiconductor industry in the United States declined the most, reaching 26.7%, and in China’s $158.4 billion semiconductor market, the share of American companies was as high as 47.5%, and 1245% of the rest of Asia Pacific. In the US$100 million market, the market share of US companies is also as high as 48.7%. In view of the uncertainty of the US in terms of trade frictions and restrictions on the export of high-tech products to China in the coming year, US semiconductor companies may not be able to change the slump in 2019 and usher in the industry-wide recovery.
(2) 5G and automobiles are the biggest drivers of global semiconductors and the most innovative application markets.
5G released a large amount of orders to the global semiconductor market as scheduled in 2019, which to a certain extent will help the gradual recovery in the second half of the year. Due to the lack of scale, the penetration rate of various 5G terminals is still low, and various 5G applications have not been fully explored. Therefore, in 2020, and even in the next three to five years, it can continue to benefit from the “5G” industry chain. Various types of innovation. In addition to 5G and automotive semiconductors, which have maintained the largest growth rate in various applications in recent years, this trend will continue in 2020. The development trend of safety, interconnection, intelligence and energy saving has gradually shifted the automotive value chain from mechanical power structure to Electronic information. System and value chain reconstruction have made new players in automotive semiconductors continue to emerge. At present, autonomous driving and vehicle electrification are the two mainstream applications affecting the automotive semiconductor sector, while automotive-grade sensors, automotive intelligent computing and communications, and power semiconductors will show high innovation activity in 2020.
(3) The window for large-scale semiconductor mergers and acquisitions has passed, and U.S. companies are “grouping together for warmth”, Japan and Europe are “exporting to China”.
In 2019, international regulators, including the Committee on Foreign Investment in the United States (CFIUS), have increased their scrutiny of cross-border transactions, and the semiconductor field, as a sensitive industry involving national security, has naturally received more special attention, and semiconductor acquisitions have become more and more important. It is getting more and more difficult, and it is unlikely that large-scale semiconductor mergers and acquisitions will happen again in the future, but more reflected in the vertical integration of subdivisions. In 2020, U.S. companies may still face export restrictions brought about by trade frictions. U.S. semiconductor companies that are highly dependent on the Chinese market may integrate their individual business lines by “holding together for warmth”. In the face of obstacles in international M&A transactions, Chinese mainland semiconductor companies have also turned their attention to the domestic market. In the first eight months of 2019, there were 20 international mergers and acquisitions (including agreements) in the semiconductor industry, with a total value of about US$28 billion, none of which involved mainland China. However, in 2020, with the real needs of mainland China in automotive semiconductors, industrial control electronics, equipment parts and materials, the exchanges between mainland Chinese capital and enterprises, Japan, and European industries will gradually increase, and Japan and Europe will replace in some areas. The role of the United States is to “export the core elements of the industrial system to China.”
(4) Moore’s Law has slowed down, and semiconductors have entered the era of “heterogeneous assembly”.
Although Moore’s Law is still advancing towards 5nm-3nm-1nm, it seems that there are fewer and fewer players who can continue to shrink following this path. The industry has also begun to pay more attention to heterogeneous assembly, including the recently extremely sought-after Chiplet (DARPA’s CHIPS project, Marvell’s MoChi, Intel’s EMIB and Foveros), silicon photonics technology, Micro LED, etc., all of which are semiconductor heterogeneous assembly. and integrated use cases. This concept of heterogeneous assembly originated from multi-chip modules and was born in the 1970s, but it is especially suitable for today’s fragmented needs, which can support rapid development and reduce chip implementation costs. At the same time, heterogeneity can also be combined with new materials to break through the physical limitations of silicon, and to apply silicon to various fields. It is expected that more heterogeneous ecosystems will emerge after 2020. New product forms, new data computing architecture designs, and new material combinations are leading the industry to explore new possibilities for chip architecture, computing efficiency, and load functions. , the future will be the era of “heterogeneity”.
(5) Domestic substitution caused by trade frictions is still the main line of the development of the domestic semiconductor industry.
The Huawei incident in 2019 accelerated the reshaping of the semiconductor supply chain system. It can be said that the entire domestic semiconductor industry chain has encountered a rare historic opportunity. Affected by the Huawei incident, many leading system-level manufacturers in various fields in China are also accelerating the introduction of domestic semiconductor products. In addition, Japan also began to sanction the field of semiconductor materials in South Korea during the year. The “efficiency priority” of the globalization of the semiconductor industry chain for 30 years has been challenged. The current global semiconductor supply chain is more “safe and controllable” as the main line. Therefore, even though the semiconductor industry in mainland China has repeatedly emphasized the need to open up and cooperate, the industry itself has become a bird of fear. In 2020, domestic substitution will continue to be the main line of the development of the domestic semiconductor industry, and the leading enterprises of domestic substitution may expand from Huawei to more domestic systems. Manufacturers, the replacement products are also upgraded from low-end products to storage, analog, radio frequency and other strategic-level general-purpose or high-end products with a wide range of products. It is the general trend to accelerate the establishment of a complete and independent domestic semiconductor industry system with independent core technologies. Domestic foundry, packaging, testing, and supporting equipment and materials will also accelerate domestic substitution in 2020.
(6) The semi-monopoly industry market with Chinese characteristics will become the “touchstone” for the high-quality upgrade of domestic semiconductors.
In mainland China, certain industries and markets have some monopolistic characteristics due to specific technical and economic characteristics or scarcity of resources, or involving national security, such as high-speed rail, smart grid, Beidou navigation, ultra-high-definition video, security, etc. As we all know, the survival and success of the semiconductor industry still depends on shipments, and these semi-monopoly industry markets with Chinese characteristics have a certain shipment basis, and have a multi-level industrial chain from systems to software to chips. Certain products have the right to speak, so domestic chips are imported to replace them, creating a touchstone market with a good independent ecology. Affected by the Huawei incident, in 2020, system manufacturers in these semi-monopoly industry markets are expected to comprehensively promote breakthroughs in systems and mechanisms and expand further cooperation with domestic chip suppliers.
(7) China will continue to increase investment and integration in advanced technology, and IDM will appear in characteristic fields.
At present, in terms of advanced process foundry, it has become a situation where Samsung and TSMC are competing for hegemony, and only SMIC is following up in the second echelon. UMC and GF have basically given up their investment in the most advanced process nodes, with the main goal of filling capacity utilization. Therefore, it is foreseeable that there will not be much changes in the overall regional pattern of global foundry in 2020, and more attention should be paid to the domestic integration in the foundry field. At present, the domestic manufacturing capacity gap of mature processes above 28nm is 350,000 pieces/month (12 inches), which is higher than the gap of 300,000 pieces/month of 28nm and below, but there are as many as 16 production lines under construction and already built at mature nodes. , and more than half of them will be put into production in 2019 and 2020, so the domestic mature process production capacity will increase in the next two years, followed by fierce competition. With the launch of the second phase of the big fund, the state’s top-level planning and guiding role in the manufacturing industry is bound to increase, and there will be integration for IDM or virtual IDM around some featured products such as analog (radio frequency) and sensors.
(8) my country’s memory has ushered in the most critical year, and there will still be integration and variables in the DRAM field.
In 2019, the price of memory continued to be sluggish, and several manufacturers have reduced production and delayed equipment investment. In 2020, driven by downstream demand, both DRAM and NAND will gradually return to normal development, and supply and demand will maintain a relatively stable relationship. What is more eye-catching is the domestic memory part. Whether Yangtze Memory 3D NAND or Hefei Changxin’s DRAM, 2020 will usher in the stage of increasing production capacity and increasing the volume of large-scale production. Although the planned production capacity does not account for more than 3% of the world’s total, Coupled with the optimistic price increase expectations next year, domestic memory will not have an impact on the global memory chip pattern in a short period of time. However, after large-scale mass production, it means that it will face market inspection, price competition and subsequent patent disputes, as well as the ability to continuously invest capital. Therefore, my country’s memory industry will really usher in a key test in 2020. It is expected that the second phase of the National Fund will continue to increase the amount of memory, and the national layout of DRAM will be further clarified in 2020.
(9) The Science and Technology Innovation Board promotes fierce competition in various fields of integrated circuits, and leading companies take the lead in promoting the reshuffle of sub-tracks.
The introduction of the Science and Technology Innovation Board has made the semiconductor industry the most favored industry sector in the domestic capital market in 2019. It has also driven the market value of semiconductor companies on the main board to record highs. The primary market has benefited from the profit-making effect of the Science and Technology Innovation Board. Relying on the rapid advancement of the listing on the Science and Technology Innovation Board, a very considerable book investment floating profit has been achieved. On the other hand, the introduction of the Science and Technology Innovation Board will also help to accelerate the formation of leaders in the semiconductor segment. With the boost of entrepreneurship projects and capital, there has been a phenomenon of overcrowding of entrepreneurial projects and capital, but in the next two or three years, the Science and Technology Innovation Board will accelerate the survival of the fittest in these crowded tracks, and the leading force formed by the Science and Technology Innovation Board and the main board listed companies It will be the first to push these subdivision tracks to shuffle.
(10) The second phase of the major fund will be compared with the first phase of shareholder expansion, and will cooperate with local governments to help leading projects.
The second phase of the Big Fund will officially start the “investment investment” model in 2020. From the perspective of proportion and capital scale, perhaps nearly half of the funds will still be invested in manufacturing such as foundry and memory. In terms of design, we will pay more attention to the areas of high-end general-purpose chips that are really stuck, such as EDA, FPGA, processors, etc. At the same time, we will pay more attention to the ecology established by the downstream application side, hoping to drive the development of the design industry through the downstream application ecology. The second phase of the big fund will also strengthen investment in upstream fields such as semiconductor equipment and materials, as well as some IDM projects in the integrated circuit field. In addition, it can be noted that the number of shareholders in the second phase of the big fund has expanded compared with the first phase, from 16 shareholders to 27. Compared with the first phase, which is mainly funded by central enterprises, the second phase is mainly based on local government investment platforms. China Regions where the development of the integrated circuit industry in mainland China is relatively concentrated and mature have all participated. Therefore, it is foreseeable that the second phase of the large fund will play more role in industry guidance and orderly integration, and cooperate with local governments to invest in local benchmarking leading projects Therefore, the competition for large projects will continue to compete in the next year.
The global semiconductor industry in 2020 will generally be better than 2019. Although there are many “uncertainties” in the evolution of China-US technology decoupling towards normalization, the characteristics of semiconductor globalization will not be completely changed even if it is impacted by this. As far as the domestic semiconductor industry is concerned, this situation is both a challenge and an opportunity. “The sword is sharpened, and the plum fragrance is bitterly cold.” Whether it is replacement or exporting value to the global market is a protracted battle. Only by constantly sharpening oneself can we seize the opportunity. Don’t fall behind this best and most bumpy era.
About the Author
Zhu Jing: Beijing International Engineering Consulting Co., Ltd., senior economist, and deputy secretary-general of Beijing Semiconductor Industry Association.