This article first appeared in Nikkei Asia on October 3, 2024.

 

To maintain progress, Tokyo must solve financing issues and engineer crunch

Japan’s semiconductor industry revitalization plans, initiated in 2021, are progressing well and faring much better than similar plans elsewhere. Taiwan Semiconductor Manufacturing Co., the world’s top chipmaker, opened its first plant in Kumamoto in February and will begin mass production later this year, while its Arizona plant in the U.S. has been delayed. Japanese chipmaker Rapidus is set to open its pilot fab in Hokkaido in April.

Even just a few years ago, Japan’s chip revival was deemed unlikely. Having lost its global dominance in the 1990s, Japan has fallen about a decade behind technology leaders in Taiwan Province of China, Korea and the U.S.

The mood was somber in 2019 when Japan’s Ministry of Economy, Trade and Industry (METI) held initial discussions to formulate a new national semiconductor strategy. Many METI council members had little faith in a successful revival. However, two events triggered a turnaround — TSMC’s decision to build a fab in Kumamoto and IBM’s call to propose a partnership with Rapidus.

In the late 1980s, Japanese semiconductor manufacturers had captured more than half of the global market for dynamic random-access memory (DRAM). Threatened by Japan’s success, the U.S. imposed anti-dumping guarantees and applied 100% tariffs on Japan’s DRAM, forcing Japan to open its market to foreign producers.

Other factors contributing to Japan’s semiconductor downfall include the rapid appreciation of the yen after the 1985 Plaza Accord, the failure to invest in logic chips during the personal computer era, and the inability of Japanese companies to adapt to a new fabless-foundry business model.

While the Japanese had accepted “three lost decades” in semiconductors as a defeat, recent geopolitical developments and the rapidly evolving semiconductor landscape led to a major rethink of the country’s chip industry.

Japan’s trade surplus in electrical and electronic equipment and components has shrunk since the mid-2000s and turned into a deficit in 2022. The deficit is likely to worsen with increased digitization.

Chip shortages during the COVID pandemic severely affected Japan’s sizable automotive industry, highlighting the importance of supply chain resilience.

Moreover, industrial policy has made a comeback globally. Semiconductors have come under the spotlight due to national security concerns. The U.S.-China trade and technology tensions changed Japan’s threat assessment. In particular, Japan had to position itself as an indispensable node in the U.S.-led alliance to build a global supply chain in advanced semiconductors.

Against this backdrop, it is clear why Japan needs a new direction to revive its semiconductor industry.

METI’s semiconductor revitalization strategy comprises three steps: strengthening domestic production capacity, forming alliances with the U.S. on next-generation technology and developing game-changing future technology.

The Japanese government spent 3.9 trillion yen ($27 billion) from fiscal year 2021 to 2023 to support the semiconductor industry, outpacing the U.S. CHIPS Act in terms of its share of gross domestic product.

As part of the first step, Japan Advanced Semiconductor Manufacturing (JASM), a joint venture between TSMC, Sony and Denso, was established in Kumamoto to produce logic chips. Having received the largest portion of government subsidies, JASM will build a second fab in Kumamoto at the end of this year with the same partners, plus Toyota.

A core component of the second step in METI’s strategy is Rapidus, a government-backed startup with a consortium of eight major Japanese companies including Toyota, Sony, Denso, Kioxia, NEC, NTT, Softbank, and Mitsubishi UFJ.

Rapidus is collaborating with IBM and IMEC, Europe’s leading microelectronics R&D center, to mass-produce 2-nanometer chips by 2027. Progress appears promising. About 100 engineers from Rapidus are in New York working with IBM engineers on technology development. Meanwhile, extreme ultraviolet lithography equipment from the Netherlands’ ASML used for manufacturing advanced semiconductors is scheduled to be delivered at the end of this year.

An important but less known development is the establishment of the Leading-Edge Semiconductor Technology Center (LSTC) in 2022. LSTC spearheads R&D while Rapidus handles production. LSTC is crucial for Japan’s advanced semiconductor development, which includes chip design, physical transistor design, fast turnaround production process, materials technology, and advanced packaging.

In the third step of METI’s strategy, Japan aims to produce game-changing technology based on the convergence of photonics and electronics; potentially benefiting data centers and 6G technologies that demand ultra high speed data transmission, low latency, and energy efficiency.

Despite all the promising signs, Japan’s chip renaissance faces serious obstacles. The country’s lack of trained engineers is a major problem. Many experienced engineers have left to work overseas and are already in their 50s. Japan’s declining birthrate and decreasing number of graduates suggest that the domestic talent pool is insufficient to meet the industry demand. Relatedly, many of Japan’s small and medium-size suppliers are owner-operated and have to deal with business succession.

The government’s gambit on Rapidus-LSTC is a risky investment. Rapidus has no chipmaking experience and relies on IBM’s 2-nm technology. At present, Japan’s most advanced chip technology is at the 40-nm node. Some analysts called the attempt to progress to 2-nm within two years an “unparalleled technological feat.”

Although Rapidus is receiving government subsidies to the tune of nearly 1 trillion yen, its consortium members have put up only a combined 7.3 billion yen. The financing falls short of the expected 5 trillion yen needed to start mass production. Moreover, chip companies must invest during down cycles to be ready for the rebound, but Rapidus’ joint entity structure might delay decision-making.

Another concern is the effectiveness of LSTC in bridging R&D and commercialization. As most of its committee members are from academia, linking research from the lab to manufacturing at the fab will be challenging.

Still, Japan’s renewed commitment to support its chip industry is bold and strategic. Unlike previous attempts, the latest semiconductor industrial policy leverages strong international technology alliances and provides massive subsidies to foreign firms. In an ironic twist of events, the U.S.-China tensions are providing an opportunity for Japan to reclaim its past glory in semiconductors, which some say to be its “last chance.”