Japan is preparing to provide another ¥150 billion ($944 million) to Rapidus, the country’s effort to establish a domestic source of cutting-edge logic chips and reduce its reliance on Taiwan Semiconductor Manufacturing Co. (TSMC), according to a Bloomberg report.
Rapidus, established in 2022, is developing an advanced semiconductor manufacturing operation in Chitose, Hokkaido. The company is targeting mass production of 2nm logic chips in 2027, putting Japan in competition with the U.S., Taiwan, and South Korea in leading-edge chip manufacturing.
The Japanese government is investing heavily in this project. It made a ¥150 billion government capital injection in June, which increased Rapidus’ capital and capital surplus to ¥424.95 billion (US$2.67 billion). Rapidus has also gotten money from private corporations, including Canon, Fujitsu, NTT, SoftBank, Sony, Toyota and Kioxia.
The investment makes clear the strategic importance the Japanese government places on rebuilding a semiconductor manufacturing capability that largely disappeared after Japan’s dominance of the chip industry faded in the 1990s.
Rapidus says the financing will support its transition from research and development into commercial production. The company is also developing advanced packaging and shorter manufacturing cycle times as part of its strategy to differentiate itself from established foundries.
Japan’s semiconductor strategy puts it squarely in competition with TSMC, yet at the same time, one major corporation is working with TSMC. TSMC and Sony are reportedly discussing a combined ¥1 trillion ($6.4 billion) investment in a planned Japanese factory that would produce next-generation image sensors for automotive and robotics applications.
So one way or another, Japan is determined to become a player in the semiconductor manufacturing business, a position it once held decades ago.
Rapidus has no established high-volume manufacturing operation at the leading edge, or for that matter at any edge. The company is only four years old, essentially a startup, and no one wants to enter the market at the 2nm generation, the same as veteran companies with decades of experience.
However, it does have one experienced veteran on its side. The company is working with IBM on 2nm process technology and has said it intends to use extreme ultraviolet lithography and advanced packaging technologies in its production strategy.
So can Rapidus compete commercially? Government funding can give Rapidus the capital needed to reach production, but establishing a competitive foundry requires more than building a fab.
For starters, it’s going to need talent. People working in fabrication facilities don’t just push a button. Those people are going to be rare and TSMC and the Koreans will have them locked up tight.
The company must demonstrate that its 2nm process can achieve competitive yields, performance and power efficiency comparable to TSMC and company while also attracting customers willing to commit production volumes. Its first customers will undoubtedly be Japanese companies, and they are going to want the same level of quality that they get from TSMC, Samsung, etc.
