A Japanese "Mega Project"? Japan's Strategic Bet, "Rapidus"

This is Tokyo. Today, I would like to talk about the Japanese semiconductor industry. Following the announcement of South Korea's "Mega Project," I visited Chitose, Hokkaido, last week to find out what Japan is doing and how they are doing it. In fact, when thinking of Japanese semiconductors, the TSMC plant in Kumamoto comes to mind first. The mere fact that TSMC, a Taiwanese company and the world's number one foundry, built a factory in Japan has already drawn significant attention.

However, there are several reasons why I became interested in Chitose, where the Rapidus factory is located. First, it shares similarities with South Korea's Mega Project. Instead of "external blood transfusions" by attracting foreign companies like TSMC or Micron, it is a "self-cultivation" hub to foster Japanese enterprises. It is also similar in that there was significant government intervention, and the hub was established in a regional area rather than the metropolitan area. It is also worth noting in terms of the "speed battle" emphasized by President Lee Jae-myung regarding South Korea's Mega Project. Above all, I believe Japan's true intention and real strategic bet lies in Chitose's Rapidus. In order to regain its past status as a leader in various advanced technologies, including semiconductors, a technological gamble is underway here to leap directly from 40-nanometer (nm)—the last node Japan mass-produced—to 2-nm. "Rapidus" means "rapid" in Latin, and true to its name, it has shown record-breaking speed so far. Today, let's look into what goals Japan hopes to achieve through Rapidus, its likelihood of success, and what lessons we can draw from it.

To understand the goal, we must first understand why it started. Rapidus was established in 2022. The incorporation of the TSMC plant in Kumamoto took place a year before that. Yes, it coincided with the COVID-19 pandemic. At the time, Japan's semiconductor supply chain was disrupted to the point where the entire core national industry was shaken. In particular, the automotive industry, which can be called the backbone of the Japanese economy, took a major hit. At one point, Japan's eight major automakers, including Toyota and Nissan, cut production by 1.7 million vehicles annually, suffering losses of over 10 trillion won. It wasn't just cars. Air conditioners and water heaters, which are essential in daily life, faced shortages, and even game consoles like the PlayStation, which represents Japan, could not be found. Although it happened a bit later, the issuance of new Suica and Pasmo transit cards was suspended for a long period due to a lack of semiconductors. Experiencing these crises, the Japanese government and industry woke up. They realized that relying solely on other countries for semiconductor factories could paralyze both national security and public livelihoods at the same time.

With this as a turning point, the Japanese government made semiconductors a national priority and pursued a dual strategy of "external blood transfusions" and "self-cultivation" simultaneously. In this light, it can be seen as having started for somewhat defensive reasons compared to South Korea's Mega Project. While the Mega Project began with the aggressive goal of expanding global market share, Rapidus's primary goal was to protect domestic industries. This is even clearer when looking at the actual shareholding structure. There are currently 32 private investors, including many major Japanese corporations that need semiconductors, such as Toyota, NTT, and Sony. They have essentially established a domestic ecosystem that can immediately purchase the chips once mass production succeeds. This is also why Rapidus has been able to speed ahead without hesitation.

Kiyoshi Hashimoto / Head of Public Relations, Rapidus
"There are semiconductor companies, automotive companies, and companies like NTT that want to pioneer new technologies in the telecommunications field. The investing companies are likely thinking about various business synergies in the future."

However, if asked whether Rapidus will succeed in the future, it is true that relying solely on the domestic market will be difficult. Due to the nature of semiconductor manufacturing, which requires massive initial investments, securing large-scale customers like Nvidia is essential to turn a profit. Even the Kumamoto TSMC plant, which secured major corporations like Sony and Denso as customers from its inception, remained in the red until last year. In fact, excluding TSMC, which holds a dominant position with over 60% market share, there are virtually no foundry companies making stable profits. Furthermore, succeeding with a 2-nm semiconductor prototype is a completely different matter from succeeding in mass production. Although the Japanese government has provided over 20 trillion won in research funding and directly purchased more than 2 trillion won in shares, this is far less compared to South Korea's 2,000 trillion won Mega Project. In that sense, it is indeed a risky gamble, and it is true that there are skeptical views within Japan as well.

However, there are also reasons why this company should not be underestimated. Based on interviews with Rapidus, let's break it down into three aspects: funding, technology, and customers. First, the government's investment in Rapidus should not be measured by the amount alone. Last year, the Japanese government amended the Act on Promotion of Information Processing (commonly referred to as the "Rapidus Support Act") to enable long-term government investment in private enterprises. This year, it invested 250 billion yen to become the largest shareholder. Becoming the largest shareholder is significant. It means the government can mobilize administrative power to provide tangible and intangible support. Regarding power supply alone, the government decided to invest 1.8 trillion yen to construct an undersea transmission line connecting Hokkaido and the main island, and Prime Minister Takaichi publicly pledged full support during her European tour.

Sanae Takaichi / Prime Minister of Japan (Speaking last month during a meeting with Rapidus executives)
"I will make sure to promote this actively to other countries, including the UK and Italy, focusing on our allies."

This means the government itself will pitch semiconductors, much like it does for nuclear power plants or the defense industry. Since Japan is mobilizing its full economic and diplomatic power, there is an expectation that it will not let the venture fail.

Second, technology is ultimately a matter of talent. Regarding concerns about whether it would be difficult to secure professional talent since Japan has been out of the advanced semiconductor field for 20 years, they responded as follows:

Kiyoshi Hashimoto / Head of Public Relations, Rapidus
"We are proceeding with technology sharing from IBM. We sent 160 engineers to the IBM Research Center in New York to conduct research together. More than 100 of them have already returned."

Efforts to foster talent at the regional level have also begun.

Shuichi Mori / Director of the Next-Generation Semiconductor Hub Promotion Office, Chitose City
"We are cooperating with local educational institutions with the goal of training around 600 people annually, which is the estimated number of semiconductor-related employees and personnel working for related companies needed each year."

I visited the Chitose Institute of Science and Technology, which is located near the Rapidus factory, and the students there were also filled with anticipation.

Keita Shima / Freshman, Chitose Institute of Science and Technology
"Rapidus is a fairly big career option at our school. Right now, there aren't that many semiconductor classes, but I heard a semiconductor-related department will be established starting next year."

Whether they can secure enough customers in a market already dominated by TSMC and Samsung is also a crucial point. Rapidus suggested that TSMC would find it difficult to handle all of the exponentially growing demand, indicating that they plan to target that niche.

Kiyoshi Hashimoto / Head of Public Relations, Rapidus
"For the next three years or so, looking at the supply-demand balance, demand will be very strong. There may be cases where customers want to order advanced semiconductors from TSMC, but orders cannot be placed immediately due to business reasons."

This means their goal is not to immediately become number one in the world. The important thing is for Japan to acquire advanced semiconductor technology, and Rapidus calculates that customers will naturally follow once that happens. Whether Rapidus will ultimately succeed remains to be seen until the mass production stage next year. What is more noteworthy is the process of how they got here. The fact that the Japanese government, starting from former Prime Minister Abe and Prime Minister Kishida, drew a big picture including the semiconductor industry and has consistently supported it even as prime ministers changed, is something South Korea, which has started its own Mega Project, could learn from.

Park Sang-joon / Economics Professor, Waseda University
"Since it takes at least 10 or 20 years to see the results, rather than the government trying to quickly produce outcomes or make announcements just to show off, improving step-by-step is also an achievement in itself."

※