1.4nm争霸战，打响

Core Insights - The global semiconductor industry is engaged in a strategic competition centered around the establishment of 2nm wafer fabs, seen as a critical threshold for AI computing sovereignty in the AI era [1][18][25] - TSMC plans to expand its 2nm production capacity in Taiwan from seven to ten fabs, with significant investments in overseas projects in the US, Japan, and Germany [2][3][19] - Intel is positioning itself with its 18A process technology, backed by national capital, to compete directly with TSMC's 2nm offerings [4][8][19] - Samsung is ramping up its 2nm production capabilities, achieving a yield of 55-60% and securing significant contracts, including a long-term deal with Tesla [9][10][19] - Japan's Rapidus is also entering the 2nm race, supported by government initiatives, aiming to establish local production capabilities [12][16][19] TSMC's Strategy - TSMC's plan includes ten 2nm fabs in Taiwan, with an estimated cost of approximately NT$300 billion (US$80-100 billion) per fab, totaling around NT$900 billion for the additional three fabs [2][19] - The company emphasizes that advanced processes will primarily serve top-tier clients, particularly in AI and high-end computing sectors [2][3] - TSMC's strategy is a response to both the surging demand for AI chips and competitive pressures from rivals like Intel and Samsung [3][19] Intel's Positioning - Intel's 18A process is designed to compete with TSMC's 2nm technology, with recent reports indicating improvements in yield and production stability [4][6][19] - The company aims to establish itself as a significant foundry player, focusing on collaboration with design partners to enhance its manufacturing capabilities [8][19] Samsung's Developments - Samsung's 2nm process yield has improved significantly, and the company plans to increase its production capacity by 163% by the end of next year [9][10] - The company has secured a major contract with Tesla for AI chip production, which is expected to enhance its standing in the US market [9][10][19] Rapidus and Japan's Strategy - Rapidus is working on establishing 2nm production capabilities with government support, aiming to start mass production by the second half of the 2027 fiscal year [12][16] - The company is also planning to develop 1.4nm products, indicating a long-term commitment to advanced semiconductor manufacturing in Japan [16][19] Market Dynamics - The competition for 2nm fabs is driven by the need for higher transistor density and lower power consumption to support AI applications [18][19] - The establishment of these fabs is not only a corporate decision but also a reflection of national industrial policies, with significant government backing for companies like Intel and Rapidus [19][25] - The high capital expenditure required for 2nm fabs necessitates partnerships with key clients and government subsidies to mitigate financial risks [18][19] Conclusion - The race to build 2nm fabs represents a pivotal moment in the semiconductor industry, with implications for technological leadership, market dynamics, and geopolitical considerations [25] - Companies that successfully navigate this high-capital, high-risk environment will be well-positioned to influence the future of semiconductor manufacturing and AI capabilities [25]