Why is the Japanese semiconductor industry gradually falling behind?
Many people believe that Japan’s semiconductor industry surpassed the United States in the late 1980s, but this assertion is not entirely accurate. In the 1980s, Japan did indeed surpass the United States in the field of dynamic random-access memory (DRAM), but this did not represent an overall surpassing of the semiconductor industry compared to the United States. The stability of quality and overall production efficiency of Japanese industrial products at this time have caught up with those of the major developed countries in Europe and America. This phenomenon is not limited to just the semiconductor industry. Since the post-war period, Japan has consistently acquired various new technologies from the United States and Europe, while simultaneously gaining a foothold in the American market through price advantages. Before the Plaza Accord in 1985, the Japanese Yen exchange rate was relatively low, and labor costs in Japan were also lower than in Europe and America, thus providing a price advantage. The management characteristics of Japanese companies were also a reason for their ability to surpass those of Europe and America. Starting from the 1960s, Japanese companies began implementing Total Quality Control (TQC) systems. Many Japanese companies engaged in “No Defects” (ND) activities, establishing production systems aimed at zero defects. While the United States has traditionally employed a probabilistic approach, where achieving a certain level of acceptance is deemed sufficient, Japan has taken a different route. This contrast stems from the differences in national character and cultural norms between the two countries.
In the 1990s, the influence of Japanese semiconductor companies globally began to decline gradually. There were three main reasons behind this trend. Firstly, there was obstruction from the United States, symbolized by the “U.S.-Japan Semiconductor Agreement” signed by both governments in 1986. This agreement prevented the Japanese government from continuing to provide extensive support to the semiconductor industry, resulting in a slowdown in the expansion pace of Japanese companies.
The second reason was the bursting of the Japanese economic bubble in the 1990s, which placed Japanese companies under multiple pressures simultaneously: significant appreciation of the Japanese yen, saturation of the domestic market in Japan, and a severe decline in corporate profitability. Consequently, Japanese semiconductor companies faced inadequate equipment investment, leading to a decline in their capability for continuous research and development.
The third reason was the rise of neighboring countries and regions. Prior to the 1990s, the semiconductor market was primarily dominated by competition between the United States and Japan. However, after the 1990s, South Korea and Taiwan also entered the competition. The rapid development of the semiconductor industry in the 1980s led to the commodification of most semiconductor manufacturing technologies, with decades of technological accumulation condensed into semiconductor manufacturing equipment. South Korean and Taiwanese companies could achieve a competitive edge by purchasing relevant equipment, thereby establishing a solid foundation for catching up.
At the same time, Japanese semiconductor companies also failed to nurture Electronic Design Automation (EDA) firms. This is related to Japan’s longstanding focus on manufacturing and undervaluation of the software industry, as well as its lack of supportive capital markets for emerging enterprises. During the peak of Japan’s semiconductor industry in the 1980s, major Japanese companies developed their own design software internally, considering it a key component of their competitive advantage. They were reluctant to share this software with external entities, which resulted in Japan’s failure to foster EDA companies. This isolationist approach extended across various domains including industrial design software, simulation software, and statistical software. As a result, Japanese semiconductor companies lagged behind their counterparts in the United States and Europe in the field of industrial software.
Why did Nikon lose to ASML in the competition of lithography machines?
In the mid-1980s, Nikon and Canon rose to prominence in the field of lithography machines, surpassing their American counterparts. However, by 2002, Nikon’s market share was surpassed by the Dutch company ASML. At this time, there was essentially no technological gap between the two sides, and the combined market share of Nikon and Canon still exceeded that of ASML. However, after 2007, the technological gap between the two companies has widened. Today, ASML dominates the field of advanced EUV lithography machine manufacturing. The decline of the Japanese semiconductor industry, represented by Nikon, may primarily stem from several factors. Firstly, it lies in the international competitive landscape. In the 1980s, Japan’s semiconductor industry grew rapidly, with its market share surpassing that of the United States in the late 1980s. However, the United States later pressured Japan to sign the U.S.-Japan Semiconductor Agreement, which hindered Japan’s semiconductor industry. The indirect consequence of this friction between the U.S. and Japan was the rise of South Korea, China Taiwan, and the Netherlands. In the lithography machine market, companies like Samsung and TSMC, as users, naturally desired multiple options to have bargaining power. Around 2000, Intel, Samsung, and TSMC became shareholders of ASML, essentially forming a strategic cooperation alliance, while Japanese companies were excluded from this alliance.
The second factor is the changes within Nikon itself and the domestic environment in Japan. Originally named Nippon Kogaku K.K., Nikon was established in 1917 as a company under the Mitsubishi conglomerate to produce military optical products such as rangefinders and telescopes. In 1988, it was renamed Nikon Corporation. After World War II, Nikon’s main source of revenue became cameras. After 1990, the revenue from lithography machines surpassed that from cameras. Nikon entered the field of lithography machines because it had an industrial foundation in optical lenses and precision machinery (including optical interferometers, etc.). From 1975 to 1980, the Ministry of International Trade and Industry (MITI) of Japan organized the “Ultra Large Scale Integration (ULSI) Joint Research Institute” to conduct joint research on ultra-precision processing technology for semiconductor products. Around the year 2000, with the bursting of the dot-com bubble in the United States, semiconductor investments declined, putting immense pressure on Nikon’s operations. If Nikon had accepted investment from Intel at that time, the outcome might have been entirely different. However, Nikon was a part of the Mitsubishi conglomerate, making it impractical for them to accept equity investment from Intel. Moreover, during that time, Japanese companies were generally resistant to foreign investments. After the bubble economy, Japanese conglomerates producing semiconductors faced significant challenges, and Japan’s share in the global semiconductor market continued to decline. In such a situation, it was difficult for the Japanese domestic market to sustain two lithography machine companies. However, Nikon and Canon failed to integrate their lithography machine businesses in a timely manner and instead continued to compete with each other.
The third reason is ASML’s consistent pursuit of globalization, whereas Nikon focused primarily on the domestic market in Japan. The domestic market in the Netherlands is relatively small, and there were no large semiconductor companies in Europe during the 1990s. From the outset, ASML had to adopt a global approach. Continuous technological advancement in a company relies on market support and continuous feedback from customers. Simply sourcing high-quality components from around the world isn’t enough to build perfect equipment. Precision machinery like lithography machines relies heavily on software for control programs, which is as crucial as hardware components like light sources and lenses. ASML developed its own software as well. Nikon initially accumulated a considerable amount of technological expertise, and with a relatively large domestic market in Japan, it pursued a strategy focused on domestic collaboration. However, after 2000, the number of end-users in Japan continued to decrease, leading to a decline in Nikon’s revenue. As a result, their investment in research and development couldn’t keep up.
The implications for China’s semiconductor industry
For China, which has been catching up later, the most important thing in the short term is still stabilizing relations with the United States and other major developed countries. Since the 1990s, the United States alone cannot construct a complete semiconductor industry chain. Now, the reason the United States is trying to win over the Netherlands and Japan lies here. In the medium term, it is crucial to accelerate the importation of overseas talent and the cultivation of domestic semiconductor-related talent. Currently, much of the work done domestically is about indigenous substitution, with many niche areas involving technologies and products that developed countries had achieved 10 or 20 years ago.
Currently, we are in the midst of the global economic digital transformation, with the most crucial aspect being the integration of artificial intelligence and industry. For instance, in this industrial revolution, the automotive industry is also undergoing a once-in-a-century upheaval. Automobiles are evolving into intelligent products. New energy vehicles represent the largest application market for semiconductor products, and China’s new energy vehicles have already taken a leading position in the world. However, in the long term, if China aims to dominate the semiconductor industry, it will need to make breakthroughs in basic theory. Otherwise, it will be difficult to become a semiconductor powerhouse.
Major Teng is a research fellow at Asia Pacific Regional Development Institute of Hong Kong and he focuses on East Asian studies and China-Japan comparative politics research.