The Japanese Semiconductor Renaissance
Will It Be Successful?

Japan is in the midst of a semiconductor renaissance. In February 2024 the first Japan Advanced Semiconductor Manufacturing (JASM) plant (a joint venture between TSMC, Sony, DENSO, and Toyota) was completed and began operations. The plant began mass production at the end of 2024. Following the completion of this first plant, JASM decided to build a second plant in Kumamoto, which has further contributed to making Kumamoto the center of the Japanese semiconductor industry. At the same time, Rapidus, a foundry that manufactures 2-nanometer semiconductors, was established with the cooperation of IBM and the Belgian company IMEC. It is currently building a factory in Hokkaido. Furthermore, TSMC and the National Institute of Advanced Industrial Science and Technology have collaborated to establish a research institute for next-generation layering technology, which is essential for advanced semiconductors. In addition, Samsung is building a state-of-the-art semiconductor research institute in Yokohama.

Collectively, these actions demonstrate that the global semiconductor industry is gathering momentum in Japan, rewriting the geopolitical landscape surrounding semiconductors. This essay assesses the past failure of Japan’s semiconductor industry and considers the goals and challenges of the country’s new semiconductor strategy.

Why Did Past Semiconductor Strategies Fail?

In the 1980s, Japanese semiconductors accounted for half of the global market, but today they account for less than 10%.[1] Japan is internationally competitive in the fields of semiconductor manufacturing equipment and materials, and it has a large share of the global market. Why, then, has its market share of finished semiconductors fallen so much? There are multiple reasons for the decline of Japan’s semiconductor industry. It is necessary to analyze these factors and past failures when considering current semiconductor strategies.

The first reason for the failure of Japan’s semiconductor strategy was its adherence to the integrated device manufacturer (IDM) model. Most of the Japanese semiconductor industry existed within general electrical manufacturers such as Hitachi and Toshiba and was positioned as a division of these huge companies that carried out everything from design to manufacturing and packaging. Semiconductors are a capital-intensive industry, and in order to win in the market it is necessary to make huge investments on an ongoing basis. Even for huge companies with financial resources, it is difficult to continue making large investments in a single division.

On the other hand, Taiwan’s TSMC was the company that started the foundry business, which specializes in semiconductor manufacturing, thereby reducing the burden of capital investment for IDM companies. Many semiconductor manufacturers concentrated on high-value-added businesses such as development and design. TSMC took the strategy of improving profitability by concentrating production in a single place, improving yield with advanced technology, and then using the profits to make further capital investments. In other words, even though the semiconductor industry had moved to a global, horizontal division of labor, Japanese companies persisted with the IDM approach. As a result they were unable to compete with foundries such as TSMC, which had superior financial and technological capabilities.

The second factor was the lack of willingness for industrial restructuring in Japan. The assets involved in semiconductor manufacturing in Japan were spread across multiple companies, and each company retained its equipment and sought to engage across the entire chipmaking process (from design to the final product). This made it difficult to expand semiconductor manufacturing, where economies of scale are important. The government attempted to restructure the industry by consolidating the skilled labor and manufacturing capabilities of each company as part of the “all Japan” project. The competing companies, however, avoided revealing their cards and did not provide top-level engineers and scientists, so in the end these projects were not successful.

On the other hand, TSMC, which concentrated its production capacity as a foundry, quickly grew into a global company, while Japan ended up exhausted by domestic competition in the face of global competition. Ironically, as a result of the exhaustion of semiconductor manufacturing companies and their inability to make continuous investments, the industry began to realize the necessity for restructuring. National champions were born, such as Renesas Electronics, but it was already too late to reverse the decline of Japan’s semiconductor industry.

It is often argued that the Japan-U.S. Semiconductor Agreement that was concluded in 1986 prevented Japan from entering the U.S. market, and that this led to the decline of the semiconductor industry.[2] But as mentioned above, a more reasonable view is that the semiconductor agreement was only one trigger and that broader structural problems were the primary cause.

What Is the Aim of the Current Semiconductor Strategy?

Based on the circumstances of this past failure, Japan is restarting its semiconductor strategy as part of its economic security policy. What is the aim of this semiconductor strategy? First, it is to ensure “strategic autonomy” in economic security policy. Before and after the Covid-19 pandemic, there were situations where the supply of semiconductors was disrupted due to a combination of factors such as the cold wave in Texas and the fire at the Hitachinaka plant of Renesas Electronics. As a result, the supply of semiconductors was recognized as a serious problem, with the production of not only automobiles but also hot water heaters and contactless credit cards coming to a halt. The government aimed to secure domestic semiconductor production and create a system that would ensure a stable domestic supply even if the supply of semiconductors produced overseas were to stop. This included supporting foreign companies that produce in Japan, regardless of their nationality.

The government’s support for the establishment of Rapidus, which will produce 2-nanometer semiconductors, is aimed at ensuring “strategic indispensability.” It is expected that the demand for such semiconductors will increase over the long term as the technology of generative artificial intelligence (AI), represented by ChatGPT, improves dramatically and becomes more widely used. Only a limited number of companies, such as TSMC, are able to produce semiconductors suitable for AI, and supply may not always be able to meet demand. Japan has not yet mastered the know-how for manufacturing 2-nanometer semiconductors, as its most advanced fabs currently only produce 40-nanometer semiconductors. Therefore, if Rapidus succeeds in manufacturing 2-nanometer semiconductors, it could become an indispensable presence in the global market.

Japan’s Ministry of Economy, Trade and Industry (METI) has created the Leading-edge Semiconductor Technology Center, which will collaborate with research institutions and companies from supportive countries and regions. These include the U.S. National Semiconductor Technology Center, Belgium’s IMEC, France’s CEA-Leti, and Germany’s Fraunhofer. Additionally, the June 2023 update of METI’s Semiconductor and Digital Industry Strategy outlines the promotion of collaboration not only with the United States and the European Union but with various other countries and regions, including Belgium, the Netherlands, the United Kingdom, South Korea, and Taiwan. It also mentions utilizing multilateral forums such as the G-7 summit and the Government/Authorities Meeting on Semiconductors (GAMS) to advance these collaborations. GAMS will be utilized for public-private collaboration to promote cooperation with like-minded countries and regions.

In this way, Japanese companies have abandoned their focus on IDMs by specializing in semiconductor foundries. This decision is based on their recognition that they should not be fixated on domestic know-how but instead play a part in the international division of labor through cooperation.

Challenges for the New Semiconductor Strategy

Japan is moving forward with a new semiconductor strategy, learning from past mistakes. Yet, in order for this strategy to succeed, it must overcome various challenges.

One of the issues is how to raise funds. The government has provided subsidies of 1.2 trillion yen to JASM’s first and second factories in Kumamoto and 920 billion yen to Rapidus.[3] In the case of JASM, TSMC, Sony, and DENSO have invested significant funding in the company. By contrast, in the case of Rapidus, even though eight companies are investing in it, the total amount of investment is only 7.3 billion yen, a fairly small amount compared with government subsidies. In September 2024, that three megabanks and the Development Bank of Japan plan to invest 25 billion yen in Rapidus, but even this may not be enough to enable the company to continue its business.[4]

The semiconductor manufacturing industry is an equipment industry, and the lesson from past failures is that it is impossible to survive in international competition without maintaining cutting-edge production capacity through continuous capital investment. Based on this lesson, it is necessary to create a financial cycle that enables continuous investment. Once mass production begins and profits start to be made, it will be possible to invest those profits. The primary challenge is to successfully raise the necessary funds to reach that point. Under the Kishida administration, the government planned to draft a bill to provide guarantees for investment in Rapidus, but this work was suspended before the general election. Passage of this legislation remains critical for attracting investment from the private sector.

The second issue is the shortage of skilled labor. Many of the personnel who will be responsible for the core operations have already received training at places like IBM and have achieved necessary levels of knowledge and skills, but the semiconductor industry requires a wide range of human capital, including various suppliers. Rapidus has already signed agreements with Hokkaido University and other institutions and is hoping to develop and supply human resources. The big question is whether it can recruit enough qualified people before mass production actually begins. Taiwan and the United States are also grappling with this challenge, which demonstrates a global shortage of the skilled labor necessary for the semiconductor industry. One of the keys to surviving international competition is to acquire the requisite labor force.

If Japan can overcome these challenges and become a center for semiconductor manufacturing, gaining strategic autonomy and strategic indispensability, then its strategy will be successful. However, continued government support is essential to secure key resources such as funding and skilled labor. The Ishiba cabinet will need to commit with the same level of passion as the previous administration.

Kazuto Suzuki is Professor of Science and Technology Policy at the Graduate School of Public Policy at the University of Tokyo and Director of the Institute of Geoeconomics.

Endnotes

[1] Mireya Solís and Mathieu Duchâtel, “The Renaissance of the Japanese Semiconductor Industry,” Brookings Institution, June 3, 2024, https://www.brookings.edu/articles/the-renaissance-of-the-japanese-semiconductor-industry.

[2] See, for example, Bryan Johnson, “The U.S.-Japan Semiconductor Agreement: Keeping Up the Managed Trade Agenda,” Heritage Foundation, January 24, 1991, https://www.heritage.org/asia/report/the-us-japan-semiconductor-agreement-keeping-the-managedtrade-agenda.

[3] Ministry of Economy, Trade and Industry (Japan), “Outline of Semiconductor Revitalization Strategy in Japan,” July 2024, https://www.meti.go.jp/english/policy/0704_001.pdf; and “Japan Approves $3.9 billion in Subsidies for Chipmaker Rapidus,” Reuters, April 2, 2024, https://www.reuters.com/technology/japan-approves-39-billion-subsidies-chipmaker-rapidus-2024-04-02.

[4] “Japan Megabanks, DBJ to Invest Up to $175m in Rapidus,” Nikkei Asia, September 26, 2024, https://asia.nikkei.com/Business/Tech/Semiconductors/Japan-megabanks-DBJ-to-invest-up-to-175m-in-Rapidus.