Japan's newly founded advanced chips manufacturing company Rapidus aims to develop 2nm chips mass production technology through cooperation with mainly the US and foray into the foundry service market in the second half of the 2020s. But the company is facing daunting challenges in materializing the ambitious goal, and it remains to be seen whether Rapidus can really become a new player beyond TSMC, Intel and Samsung Electronics in the advanced chips manufacturing arena.
Rapidus is jointly established by a consortium of eight major Japanese corporates including Toyota Motor, Sony Electronics, SoftBank, Kioxia, NEC, NTT, Denso and MUFG Bank, with their total initial investment of JPY7.3 billion (US$52 million) and the Japanese government to provide a subsidy of JPY70 billion.
Among them, Toyota has expressed its support for building next-generation semiconductor production mechanism in Japan to meet future demand for automotive chips, as sub 2nm chips will be widely used in autonomous driving and AI applications in 2030s.
At a press conference held in November to mark its formation, Rapidus announced plans to mass produce 2nm chips in Japan in 2027 for 5G communications, quantum computing, data center, self-driving vehicle and smart city applications.
Based on the Rapidus schedule, Japan is aggressively striving, within five years, to reach the point where its advanced chip manufacturing technology could lag only 1-2 years behind TSMC and Samsung, both of them are set to mass produce 2nm chips in 2025.
Rapidus to introduce first EUV machine into Japan
Apparently, Rapidus will have to catch up with the progress in a short period of time through technical cooperation with other overseas enterprises, and it will also introduce EUV lithography equipment, which will be the first such machine to be adopted in Japan.
Rapidus chairman Tetsuro Higashi, a former president at Japan's equipment maker Tokyo Electron, told the press conference that the company's goal is to become the world's number one in specific chip segments, stressing that gathering technologies worldwide is critical to achieving the goal as uniting Japanese firms alone can hardly secure a win in the semiconductor industry.
Its president and CEO Atsuyoshi Koike, an ex-president of Western Digital's Japan unit, also noted that 2nm process node is a difficult technology, but there is still a chance for Rapidus to catch up through cooperation with IBM and other advanced semiconductor R&D specialists.
Rapidus signed a partnership pact with IBM on December 13, allowing it to commercialize the US firm's 2nm manufacturing process and ramp up mass production of the chips. Just one week earlier, Rapidus also inked a memorandum of cooperation (MOC) with Belgium-based microelectronics research hub IMEC on advanced semiconductor technologies.
Rapidus, LSTC to jointly revive Japan's semiconductor industry
Both IBM and IMEC are joining a new Leading-edge Semiconductor Technology Center (LSTC) to be established by Japan's Ministry of Economy, Trade and Industry (METI) by the end of 2022 as an R&D platform for next-generation semiconductors in cooperation with the US and Europe, with Rapidus chairman Higashi to also head the LSTC's board of directors so that both parties can cooperate closely to revitalize Japan's semiconductor manufacturing capability, according to METI sources.
Other participants in the LSTC include the University of Tokyo, National Institute of Advanced Industrial Science and Technology (AIST), government-backed research body Riken, Tokyo Institute of Technology, Tohoku University, and National Institute for Materials Science (NIMS).
The establishment of Rapidus and LSTC signifies broad cooperation between the US and Japan in developing advanced semiconductor mass production technology. METI minister Yasutoshi Nishimura has stressed it is a kind of collaboration pursued by the 14-member Indo-Pacific Economic Framework, and it's also hoped that both countries can become the world's supply centers for key semiconductor materials.
It is clear that through Rapidus and LSTC Japan is attempting to acquire overseas technologies, mainly from the US and Europe, to build mass production lines for sub-2nm chips in the country. Japanese news media reports and commentaries have supported the attempt and shown high regard for the grand lineups of Rapidus and LSTC, while Korea media also reported that Samsung is closely watching Japan's latest semiconductor development efforts which it would not underestimate.
Challenges facing Rapidus
TSMC has to walk for 17 years from launching its 40nm process node in 2008 to a planned 2nm chips commercialization in 2025, and therefore industry analysts in Japan pointed to many challenges facing Rapidus in catching up to produce 2nm chips in five years (2022-2027), given Japan's best process node now is 40nm technology adopted by IDM Renesas in its automotive MCU production lines.
First, Rapidus' startup capital of JPY77.3 billion (US$570 million) lags far behind huge investments needed for wafer fab construction by major foundry houses. Its president Koike has said the company will not seek to match TSMC and Samsung in terms of production scale, but will adopt a different approach, such as producing small volumes of chips with high unit prices. Nevertheless, analysts said, the capital required will surely not be small for Rapidus to fast progress to 2nm chips mass production within only five years.
They estimated a capital injection of JPY2 trillion is needed during the R&D stage and another JPY3 trillion is required for the mass production phase. It seems that Toyota, Denso, and Softbank are still taking a wait-and-see attitude given their initial capital contribution of only JPY1.0 billion each to Rapidus. According to a Nikkei report, some corporate stakeholders in Rapidus are just investing at the request of METI and have little confidence in the success of the new venture.
However, Reuters reported that Rapidus is expected to secure additional funding from unspecified sources to be able to build its own wafer fab in Japan, suggesting many of Rapidus' backers will also be forking out some cash to help the company get off the ground.
The second challenge is the shortage of talent supply. As Japanese semiconductor firms have not invested in process technology upgrades for many years, quite a few Japanese engineers have chosen to develop overseas, resulting in a shortage of talent in Japan. Amid fierce international talent recruiting competition, it's hard to say whether Rapidus can hire sufficient talent within five years.
The third challenge refers to patent technologies. TSMC, Intel and Samsung have spent many years accumulating proprietary chip manufacturing technologies and building their own ecosystems, and thus it would be a tough task for Rapidus to reach a similar level in five years.
One more challenge is associated with the market demand side. The most advanced chips are a high-cost business, and therefore there must be sufficient clients using advanced chips in their end products to support viable manufacturing operations for the chips, rather than relying on additional investments to survive. Whether Rapidus can secure sufficient chip demand from potential clients will determine if it can gain solid ground in the advanced chips foundry market, the analysts said.
Article translated by Willis Ke
