The growing emphasis on supply chain security, particularly in sensitive industries like the military, is driving Western countries and their allies in Asia, such as Japan and South Korea, to implement various policies to ensure the resilience of their local supply chains in the face of worsening geopolitical tensions. One critical component of the supply chain is semiconductor chips, which are considered to be one of the most vulnerable links. To address this vulnerability, countries like the United States, European Union, and Japan have introduced subsidy policies to ensure that at least one local semiconductor foundry can meet their respective demands. With the current trend, companies like Intel are benefiting from the surge in demand for foundry services.
The feasibility of achieving 2 nm semiconductor technology has been a topic of discussion. Looking at the investment figures provided by TSMC, a leading semiconductor foundry, we can gain some insights. TSMC announced plans to invest 40 billion in the construction of two fabs in the United States, with one aimed at producing 4 nm chips by 2024 and the other targeting 3 nm chips by 2026. Assuming the costs of building a 2 nm fab in Japan are similar to those of a 3 nm fab in the United States, the construction expenses alone would amount to a minimum of 40 billion in the construction of two fabs in the United States, with one aimed at producing 4 nm chips by 2024 and the other targeting 3 nm chips by 2026. Assuming the costs of building a 2 nm fab in Japan are similar to those of a 3 nm fab in the United States, the construction expenses alone would amount to a minimum of 40 billion in the construction of two fabs in the United States, with one aimed at producing 4 nm chips by 2024 and the other targeting 3 nm chips by 2026. Assuming the costs of building a 2 nm fab in Japan are similar to those of a 3 nm fab in the United States, the construction expenses alone would amount to a minimum of 20-30 billion. Additionally, the research and development costs incurred by TSMC in 2022 amounted to around 5.5 billion. Therefore, the total R&D investment for the next four years would be at least 20 billion. Considering these figures, the total investment for TSMC to achieve 2 nm technology would reach around 40−50 billion, exceeding the 40-50 billion, exceeding the 40−50 billion, exceeding the 35 billion investment stated by Rapidus.
It is important to note that IBM, despite divesting its semiconductor manufacturing business to Global Foundries in 2015, has continued its research in advanced semiconductor technologies. For instance, IBM has actively participated in prestigious semiconductor conferences, including the Symposium on VLSI Technology and IEDM, where they have presented significant papers outlining advancements in semiconductor technology. The 2 nm semiconductor technology licensed by Rapidus from IBM is based on IBM's previous developments in gate-all-around (GAA) technology. Rapidus established a partnership with IBM to commercialize this technology, and engineers from Rapidus have undergone technical handover training at IBM's facilities in the United States. By leveraging IBM's technology and expertise, Rapidus aims to establish its first fab in Hokkaido, Japan, with construction set to be completed by 2024 and trial production to commence by 2025.
Despite the potential benefits of Rapidus partnering with IBM, it is important to consider the significant funding requirements for achieving 2 nm technology. The estimated investment of 35 billion for Rapidus to achieve mass production of 2 nm chips by 2027 appears to be insufficient to cover the costs, especially when considering the figures needed by a well−established and technologically advanced company like TSMC. Given that TSMC′s investment is estimated to be around 35 billion for Rapidus to achieve mass production of 2 nm chips by 2027 appears to be insufficient to cover the costs, especially when considering the figures needed by a well-established and technologically advanced company like TSMC. Given that TSMC's investment is estimated to be around 35 billion for Rapidus to achieve mass production of 2 nm chips by 2027 appears to be insufficient to cover the costs, especially when considering the figures needed by a well−established and technologically advanced company like TSMC. Given that TSMC′s investment is estimated to be around 40-50 billion, Rapidus, a comparatively new company established just last year, will face challenges in meeting their ambitious targets with a budget that is potentially smaller than TSMC's. However, it is worth mentioning that Rapidus has expressed its intention to focus on serving a few major customers, particularly targeting orders for high-performance computing ASICs from leading U.S. internet companies like Google, Amazon, and Microsoft, leveraging the growing trend of artificial intelligence.
In conclusion, the drive for supply chain security and the vulnerability of semiconductor chips have led countries like the United States, European Union, and Japan to adopt measures to ensure local production capabilities. Rapidus, in collaboration with IBM, aims to establish itself as a significant player in semiconductor manufacturing, but the actual feasibility of achieving 2 nm technology remains uncertain given the substantial investment requirements and the company's relatively recent establishment. Nevertheless, Rapidus has a strategic focus on securing orders from major internet companies, indicating a potential niche market in the chip trading business. The global semiconductor industry continues to evolve, and companies like Rapidus face both opportunities and challenges in this dynamic landscape.
