日本半導體國家隊Rapidus有危機？專家：這點輸台積電、三星 | 遠見雜誌
日本政府通過內閣會議決定修法，允許政府投入巨資，幫助半導體新創Rapidus。朝日新聞報導，觀察半導體產業超過30年的日本機械產業振興協會特任研究員井上弘基擔心，認為追求最先進的半導體微縮製程本身已是「過時的想法」，並指出戰術已經發生變化，但日本仍固守「大艦巨砲主義」的思維。
Machine translation:
The Japanese government has approved a legislative amendment allowing the state to inject massive funds into the semiconductor startup Rapidus. According to Asahi Shimbun, Hiroki Inoue, a senior research fellow at the Japan Machinery Industry Promotion Association who has observed the semiconductor industry for over 30 years, is concerned that pursuing cutting-edge semiconductor miniaturization is an "outdated concept." He points out that industry strategies have evolved, yet Japan remains stuck in the "big battleship doctrine" mindset.
The report highlights that the biggest uncertainty for Rapidus is the lack of sufficient customer demand. Once production begins, shipments are expected to reach hundreds of millions of chips, but there are currently not enough customers to support this volume. Advanced semiconductors are mainly used in smartphones and data center processors, yet Rapidus has not secured clients in these segments. Inoue questions, "Who will buy the mass-produced chips from Rapidus, and how much will they actually purchase?"
When TSMC and Samsung develop advanced semiconductors, they collaborate with major clients such as Apple, Qualcomm, and NVIDIA in advance to secure orders and align their technology roadmaps with customer needs. This approach is completely different from Rapidus. Objectively speaking, even if Rapidus successfully builds a cutting-edge manufacturing facility, the absence of buyers could leave production capacity idle, potentially leading to financial crises.
Experts Say Rapidus' 2nm Focus Is "Outdated"
Inoue also argues that Rapidus’ decision to focus on 2nm process technology is already "outdated." As Moore's Law slows down, the benefits of further miniaturization are diminishing. He suggests that Japan should shift its focus to advanced packaging, which is closely related to semiconductor manufacturing equipment and materials—areas where Japanese companies hold a competitive advantage. Investing in this sector would yield significant results with lower capital requirements. He advises Rapidus to pause its mass production plans and concentrate its resources on R&D instead of rushing into the highly risky mass production phase.
Beyond Inoue’s skepticism toward Rapidus’ business model, past Japanese government officials and economic commentator Shigeaki Koga have also warned that the likelihood of Rapidus failing is increasing.
Lack of Private Investment and Financial Challenges
The estimated funding required for Rapidus is 5 trillion yen, yet private investment contributions have only reached 7.3 billion yen, with no further increases. This indicates that no private investors are willing to take on this project. Koga noted that while many corporate executives attended the groundbreaking ceremony for Rapidus' factory in Chitose, Hokkaido, most were representatives from companies selling semiconductor manufacturing equipment—purchased with large amounts of taxpayer money.
Rapidus is currently constructing a factory in Chitose, Hokkaido, intended to serve as an R&D and production base for next-generation logic semiconductors.
Koga advises the government to avoid being misled by bureaucratic inefficiencies that hinder the path to sound policies. Instead of creating a new company from scratch, the government should support existing Japanese companies or promising startups. This approach would require far less investment than Rapidus while unlocking new opportunities.