In November 2022, the Japanese government and eight Japanese multinationals invested in a firm to make cutting edge semiconductors. The company, named Rapidus, aims to mass produce two nanometre logic chips by 2027. This is a technological leap for Japan, as it is currently only able to produce 40 nanometre logic chips. But its success is not guaranteed.
The company faces headwinds. One is a shortage of engineers with the skills to make semiconductors. Workers need to understand materials such as gallium nitride and be able to master packaging and artificial intelligence technologies. Since Japan lost its lead in semiconductor manufacturing in the 1980s, it no longer has enough workers with the requisite knowledge to manufacture advanced chips. Though Rapidus has sent several hundred workers to IBM for training, the Japanese Electronics and Information Technology Industries Association forecasts that Japan will need 40,000 more semiconductor engineers.
Another challenge that Rapidus faces is that customers may be reluctant to purchase from an unproven company. There are already established manufacturers, such as Taiwan Semiconductor Manufacturing Company (TSMC), that make reliable chips and have earned customers’ trust.
On the other hand, Rapidus has several factors working in its favour. One is that, like TSMC, it is a pure play manufacturer. This means that it will only manufacture for its customers and not simultaneously compete with them. Chip designers trust that TSMC will not steal their intellectual property because TSMC does not design or manufacture semiconductors in its own name.
Another factor favouring Rapidus is that there is enormous pent-up demand for advanced chips that can meet the needs of artificial intelligence and data centres. TSMC may eventually be unable to satisfy this demand. This provides an opportunity for Rapidus if it can produce quality chips.
In addition, the weak yen means that the US dollar cost of producing chips in Japan has fallen. From 2007 to 2012, a strong yen decimated the Japanese semiconductor industry. Electronic components are often priced in US dollars, and when the yen is strong the costs rise relative to the US dollar revenues. A weak yen has the opposite effect and should help Rapidus to be profitable, which is important as it must invest copiously in new plants, equipment and research and development.
Rapidus will also be able to communicate well with the many Japanese firms involved in the semiconductor supply chain. Manufacturing chips requires not only factories but also upstream firms that provide chemicals, cleaning machines, valves, tubes, pumps, plastic containers, gases, metals and many other products. Only a few firms can meet these exacting standards for each input, and many of those firms are Japanese. Japanese firms interact easily with other Japanese firms, solve problems well together and exhibit high degrees of trust. This should help Rapidus to succeed.
The Japanese government has provided 920 billion yen (US$5.7 billion) to support Rapidus. There is concern that industrial policies could produce rent-seeking waste. But in the case of Taiwan when it was developing its semiconductor industry, these losses seem to have been minimal. This could be because there was a sense that developing the semiconductor industry was essential for Taiwan’s survival.
When scholars and government officials began considering developing the industry, Taiwan faced multiple crises. It left the United Nations in 1971 and severed relations with Japan in 1972, the latter of which damaged local industries as Japanese companies were a crucial source of goods and technology. Taiwan also suffered a 47 per cent increase in consumer prices between 1972 and 1974 due to the first oil shock. In this environment government officials, scholars, businesspeople and others pulled together to help Taiwan succeed.
Taiwan invested 25,000 times its annual per capita income to develop the semiconductor sector. These expenditures led to the creation of two successful semiconductor companies, TSMC and United Microelectronics Corporation. It also contributed to a vibrant supply chain that supported the semiconductor industry.
Japan could learn from Taiwan’s experience by emphasising that developing advanced chips is vital for Japan’s security. Japan risks losing access to imported chips due to a blockade, war or natural disaster. As Japan learned during the COVID-19 pandemic, semiconductor shortages can stop the production of many goods.
Japan should start educating middle school and high school students and the general public on the importance of developing an indigenous chip industry, which could help attract workers to the sector. If producing semiconductors domestically is framed as a national security imperative, Japan is more likely to avoid rent-seeking waste and see the country pull together to achieve this goal.
Following the 2011 earthquake, Japanese firms quickly restarted factories. Japanese companies also rebounded rapidly when the 2011 Thai floods disrupted their supply chains. But Japanese firms suffered from a drop in demand during the 2008–2009 Global Financial Crisis. They were also decimated by the strong yen afterwards, proving that Japanese firms manage supply-side challenges better than demand-side challenges.
Since the difficulties that Rapidus faces are largely supply-side issues such as training workers and improving yields rather than demand-side shocks such as weak demand or a strong yen, its prospects for success in manufacturing cutting-edge chips are strong.
August 30, 2024 East Asia Forum
