This article is from the WeChat public account:Qin Shuo’s Circle of Friends (ID: qspyq2015), Author: Chen Gang, Original title: “What Has Japan Done to Become a Semiconductor Power?”, Cover image from: Visual China
The Japanese Ministry of Economy, Trade and Industry, industry, and media are eagerly hoping that the chip factory established by Rapidus in Chitose, Hokkaido, in 2027 will lead to a revival of the semiconductor industry.
On September 1, 2023, the groundbreaking ceremony for the Rapidus Chitose factory was officially held, with plans to mass-produce 2-nanometer chips by 2027, with an investment of 5 trillion yen (approximately 250 billion RMB).
In 1988, Japan’s semiconductor products accounted for 50.3% of the global market, but by 2006, this had dropped to 25%, and by 2019, it had fallen to 10%. The Ministry of Economy, Trade and Industry predicts that if the semiconductor industry is allowed to decline, Japan’s share of the global semiconductor market will be nearly zero by 2030.
When the United States decided to suppress China in the semiconductor sector, Japan immediately sensed an opportunity for resurgence.
Starting in September 2019, the U.S. began restricting companies from engaging with Huawei in the semiconductor sector, and in December of the same year, the Japanese government officially announced strict limitations on business dealings with Chinese companies like Huawei in the semiconductor field.
Since 2020, Japan has begun to emphasize economic security issues, and from 2021 through 2022, it established a comprehensive set of economic security laws primarily targeting China, severing ties with China in terms of advanced semiconductor materials and manufacturing equipment.
On the surface, Japan’s economic security measures hindered Japanese companies’ interactions with Chinese firms, damaging the benefits Japanese companies gained in China. However, on a deeper level, since 2021, the Japanese government has strengthened financial subsidies for the semiconductor industry, aiming to restore Japan’s status as a semiconductor powerhouse by cutting off exchanges with China in advanced products and technologies.
On August 10, 2022, Toyota, Denso, Sony Group, NTT, NEC, SoftBank, Kioxia, and Mitsubishi UFJ Bank jointly invested 7.3 billion yen (approximately 360 million RMB) to establish the chip manufacturing company Rapidus (meaning “fast” in Latin), and the Ministry of Economy, Trade and Industry immediately decided to provide 330 billion yen (approximately 16.2 billion RMB) in subsidies to the company.
On February 21, 2023, Rapidus decided to build a chip factory in Chitose, Hokkaido, leveraging 7.3 billion yen to stimulate an investment of 5 trillion yen (the ratio of registered capital to total investment is 0.146%). This appears to be a case of leveraging a small amount of capital for a large outcome, but more importantly, this company has suddenly given the Japanese semiconductor industry immense hope.
Why can a newly established company with only 7.3 billion yen in capital bring such great hope to Japan? In what ways does Rapidus differ from previous Japanese semiconductor companies in its business model? Can the Japanese semiconductor industry regain its glory because of Rapidus?
Throughout the decline of Japan’s semiconductor industry, the Ministry of Economy, Trade and Industry has never ceased its subsidies, yet almost none have succeeded. Why is there hope for success this time? From Rapidus, we can not only see the characteristics of Japan’s semiconductor policies but also foresee the future of the country’s semiconductor industry.
Leveraging Small Investments for Big Outcomes Requires Public Support
From the perspective of Japanese public opinion, Rapidus is already different from previous Japanese semiconductor companies and carries a significantly new meaning.
Over the past thirty years, whenever Japanese semiconductor companies faced technological challenges or operational difficulties, they typically resorted to corporate restructuring. However, Rapidus is different; it is a new enterprise established with investments from manufacturing companies, software firms, semiconductor experts, and banks, backed by the state.
Among the investors, Toyota and Denso are Japan’s most powerful manufacturing companies;
Sony, NEC, and Kioxia, which spun off from Toshiba’s most profitable flash memory division, are well-versed in semiconductor product manufacturing;
SoftBank is a major investor in ARM, the world’s most important semiconductor circuit design company, and is an expert in IT investments;
NTT is Japan’s most important data communications company, with huge demands for storage and computing;
Mitsubishi UFJ Bank stands out in Japan for its capital-raising capabilities.
In terms of manufacturing technology, familiarity with semiconductors, and funding, Rapidus can achieve the best in Japan. Although the investment amount itself is not large, the strategy of leveraging small investments for big outcomes is quite significant.
However, it should be noted that Toyota earned 1.1209 trillion yen (approximately 55 billion RMB) in operating profit in the second quarter of 2023, but it lacks future products such as electric vehicles in important markets like China, and fuel vehicles, including hybrids, can only be maintained until 2035 at most.
When companies need to transform, the 1 billion yen investment from Toyota to Rapidus seems more like a tentative step. For vehicles, including electric ones, a 48-nanometer chip is sufficient, and there is no immediate need for 2-nanometer products; investing in a chip company that is too advanced is not aligned with Toyota’s transformation goals.
Other companies, whether SoftBank or NEC, have incurred losses in recent years, preventing them from contributing more capital. The 300 million yen investment from Mitsubishi UFJ seems somewhat like a token gesture rather than a genuine commitment.
Japanese companies are skeptical about the semiconductor policies promoted by the Ministry of Economy, Trade and Industry. Politician Akira Amari, who has previously served as Minister of Economy, Trade and Industry and has been vocal about semiconductor policies, stated to Japanese reporters on November 11, 2022: “All past semiconductor strategies have failed.”
This statement is true. Reviewing the relevant materials on Japan’s national semiconductor projects, from 2001 to 2010, the Japanese government implemented the Asuka Project 1 and 2, which ended in failure by 2010. The Milai Project, which was promoted during the same period, also yielded no results. The Haruka Project, initiated from 2001 to 2004, was halted shortly after its launch. The Tsukuba Semiconductor Cooperative, started in 2006, barely lasted until its dissolution in 2021.
After 2021, the Ministry of Economy, Trade and Industry first focused on attracting investment from Taiwan to encourage TSMC to set up a factory in Kumamoto, directly offering 600 billion yen in financial subsidies. However, this merely brought 48-nanometer chip production technology to Japan, raising Japan’s semiconductor production technology from three digits to 48 nanometers, which is not revolutionary. Japan must have a domestically capitalized company and chip factory.
As for Rapidus, which only has 7.3 billion yen in registered capital but aims to undertake a 5 trillion yen project, Toyota and other companies have never stated they would increase their investment, and they still rely on the government for funding. A thorough search of information related to Rapidus’s fundraising reveals that it still hopes for government funding; the initial 330 billion yen is a tentative step, and if there are no significant objections within Japan, the government will likely provide further subsidies.
On October 12, 2023, the Asahi Shimbun reported that the Japanese government would invest an additional 1.7 trillion yen in Rapidus, bringing the total subsidies to over 2 trillion yen. However, this is still not the final amount; most Japanese companies believe that the government will ultimately provide 5 trillion yen, which is quite probable.
The Japanese government will not only increase subsidies for Rapidus; Japanese media reports indicate that when TSMC establishes its second factory in Kumamoto, the government will again subsidize 900 billion yen, and an additional 700 billion yen will be provided for existing semiconductor products like Sony’s image sensors.
It is particularly noteworthy that the level of subsidies for semiconductor localization in Europe and the U.S. is relatively limited. The U.S. subsidy rate is 5%-15%, while Germany is higher, reaching 50%. In contrast, the Ministry of Economy, Trade and Industry’s subsidies for Japanese semiconductor companies are almost all-inclusive.
Despite the presence of companies like Toyota, which can earn over 1 trillion yen in a quarter, the government will still heavily subsidize semiconductor companies, and these subsidies can span multiple years. The Japanese public and relevant ministries will not conduct strict audits of the results of these subsidies, which “will lead to significant waste,” as reported by the Asahi Shimbun on October 12.
Thus, leveraging small investments for big outcomes is not so much about Rapidus’s capabilities, but more about persuading public opinion to agree to allocate more resources to domestic and foreign investment companies, even though all past semiconductor policies have failed, and the failures in the semiconductor business will not lead to accountability for individuals, companies, or the government.
The Innovation of Rapidus
From the publicly available information about Rapidus, the general development path before and after its establishment is as follows:
In August 2022, it was decided to invest 7.3 billion yen to establish the company, and in November, it received 70 billion yen in subsidies from the Ministry of Economy, Trade and Industry. In December, it established a strategic partnership with IBM and began sending technical personnel to IBM’s research and development bases starting in 2023.
After receiving approval from the Ministry of Economy, Trade and Industry for these actions, in April, the government decided to add another 260 billion yen in subsidies. In September, it was decided to build a factory in Chitose, Hokkaido. The installation of manufacturing equipment is expected to begin in December 2024, with the trial production line completed by April 2025, and mass production officially starting in 2027.
Building the trial production line in a year and a half and then entering mass production after more than a year of debugging indicates that Rapidus’s efficiency should be quite rapid.
From a technical perspective, the Chitose factory will produce 2-nanometer chips, while TSMC’s Kumamoto factory will produce 48-nanometer chips, placing the two at different levels of production technology in Japan. In terms of factory establishment experience, TSMC not only has its own experience in Taiwan but has also built factories in Arizona, making Kumamoto at least its third factory, and its product offerings are relatively outdated, with construction speed still not comparable to Rapidus.
It is important to note that Japan is proposing to build a 2-nanometer factory with virtually no experience in factories above 48 nanometers, and the speed of this proposal indicates considerable confidence.
This confidence likely stems from the fact that the main equipment will be provided by domestic manufacturers, and after ASML of the Netherlands decided to cooperate with the U.S. in suppressing China’s semiconductor industry, it has little important business and can assist Japan in rapidly establishing Rapidus.
More importantly, Rapidus aims to find a fast production model that differs from the traditional vertically integrated production model of Japan, outside of the level of division of labor seen in TSMC and others.
We can view the previous Japanese semiconductor production and sales model as a “department store” approach, which is vertically integrated. Companies produce what they sell in the market, and users have little opportunity to participate. Semiconductor companies can always present the latest products before users develop new products, and users use these new semiconductor products to design their new goods.
In other words, users can only follow semiconductor manufacturers to do what they want. The department store has ample product supply; the only concern is affordability, not availability.
However, companies like TSMC have changed the production method of semiconductors, transforming manufacturers into large custom factories, also known as a horizontal division of labor production model. After the advancement to a horizontal division of labor production method, planning and design are handled by fabless companies, while production is managed by foundry companies.
Rapidus will combine design-focused fabless companies with foundry companies, directly engaging in design, front-end, and back-end processes after initial communication with U.S. GAFAM (Google, Amazon, Facebook, Apple, Microsoft) to obtain relevant planning. This approach shortens the manufacturing cycle and achieves the goal of “fast” (the original meaning of Rapidus) production.
Three Major Challenges to Overcome
From publicly available information, the Ministry of Economy, Trade and Industry has allocated a budget of 10 trillion yen to catch up with other countries and regions in the semiconductor industry, and there will be substantial budgets allocated to Rapidus in the future. Rapidus has also not actively sought external funding, showing considerable confidence in obtaining subsidies from the government.
According to reports from Japanese media on October 12, the Ministry of Economy, Trade and Industry has established three funds for semiconductor production and research and development: “Fund for Strengthening Research and Development of the 5G Information and Communication System,” “Specific Semiconductor Fund,” and “Stable Supply Support Fund.”
At the national level, prioritizing semiconductors gives Rapidus ample reason to utilize the relevant government funds, almost eliminating the need to raise capital as hard as companies in the U.S. and Europe. This includes external companies establishing semiconductor factories in Japan, which can enjoy government subsidies to pursue their own business goals, making this an optimal time.
However, Rapidus faces three significant challenges.
First, Japan essentially lacks a market for 2-nanometer chips, and Rapidus faces strong competitors like TSMC and Samsung.
The primary reason for the decline of Japan’s semiconductor industry has been the sharp drop in domestic demand for semiconductor products since 2000, with the world’s largest semiconductor market shifting to China. After Japan lost domestic demand for semiconductor products, companies producing semiconductors could only sit idle, and despite the Ministry of Economy, Trade and Industry’s policy support and substantial economic subsidies, the semiconductor industry continued to decline until TSMC’s establishment of a factory in Japan began to reverse the trend. However, the market issue for products remains unresolved.
Since 2020, the focus of global technological innovation has shifted to batteries, electric vehicles, 5G, industrial internet, and advanced medical equipment. Japanese companies hold most of the technology and patents in these areas, but they have not invested significantly. Without the emergence of electric vehicles and 5G trends in the domestic market, Japan lacks companies like U.S. GAFAM or China’s BATH (Baidu, Alibaba, Tencent, Huawei) that have high demand for storage and AI, resulting in a low domestic market for semiconductors.
At the same time, economic security measures prevent Japan’s advanced technologies and related products from being exported to China. Previously, Japan could export semiconductor products to China to sustain its semiconductor industry, but now 2-nanometer chips will have to rely entirely on the U.S. market, creating new risks for Japan’s semiconductor industry.
In the future, Rapidus’s 2-nanometer products will need to compete with chips from TSMC’s Taiwan factory and the chip factory being built in Arizona, U.S. TSMC is essentially using depreciated production lines, and the Arizona factory will likely produce products before Rapidus if all goes smoothly.
With a limited market, products from TSMC’s Taiwan factory and the Arizona factory will also be supplied to GAFAM before Rapidus, and whether Rapidus’s advancements in production methods can offset TSMC’s advantages remains uncertain from a technical standpoint.
Second, Japan is severely lacking in semiconductor technical talent.
The construction of the Kumamoto factory has essentially exhausted the existing semiconductor talent, and in the face of talent shortages, Kumamoto Prefecture is actively building related training institutions to cultivate talent. Hokkaido, on the other hand, is even more lacking in semiconductor talent. Rapidus has considered sharing talent with Kumamoto. However, whether this talent-sharing approach will work smoothly, especially in cooperation with TSMC, remains difficult to assess due to the lack of previous successful examples.
Rapidus is also seeking cooperation with India. India has a considerable number of IT talents, but it lacks personnel with ready experience in semiconductor production. Collaborating with India sounds promising, but it is equally uncertain.
Third, Hokkaido has limited water and power resources.
TSMC favors Kumamoto largely due to its geographical advantages, while Chitose has insufficient water resources, and its power supply is only adequate for current usage. Before establishing the factory, Rapidus will need Hokkaido to construct new water channels and activate new nuclear power plants. Especially regarding the construction of nuclear power plants, Japan has faced long-term stagnation due to the Fukushima nuclear power plant accident, and there is considerable administrative resistance to activating a nuclear power plant for a chip factory.
Even if the challenges related to talent and geography can be overcome, Rapidus will ultimately face the need to navigate a long and arduous path due to the lack of a domestic market and intense competition abroad.
This article is from the WeChat public account:Qin Shuo’s Circle of Friends (ID: qspyq2015), Author: Chen Gang
The content represents the author’s independent views and does not reflect the position of Huxiu. Unauthorized reproduction is prohibited; for authorization matters, please contact [email protected]. If you have any objections or complaints regarding this article, please contact [email protected].
End
