TSMC, a global leader in semiconductor manufacturing, recently announced a plan to shake the industry: it is expected to achieve mass production of chips in the 1.6nm class (A16 process) by 2026, and this technological leap is very likely to be based on the existing first-generation EUV lithography machine. Behind this decision, the high cost of 2nm EUV lithography machine has become an unavoidable practical factor, which also reflects the exquisite layout of TSMC under the dual consideration of technology and economy.
Although ASML's latest 2nm EUV lithography machine was almost booked out by Intel, and it was predicted that Samsung and TSMC would compete to chase the next generation of EUV equipment, TSMC showed unusual calmness and did not rush to purchase. Far from being conservative, this move demonstrates TSMC's deep understanding of cost control and the pace of technology iteration.
Looking back on history, TSMC's choice of 7nm process is also compelling. When it was generally believed that EUV was indispensable, TSMC successfully opened up the first 7nm process production line with its mature DUV lithography technology, which effectively avoided the immaturity and high cost of EUV technology at that time, and also verified its accurate judgment in the selection of complex technology paths. Although Samsung's early attempts at EUV technology were brave, they suffered setbacks due to yield issues, which contrasted the robustness of TSMC's strategy.
Of course, the improvement of chip performance by EUV lithography machine should not be underestimated, and TSMC subsequently integrated EUV in the second-generation 7nm process, achieving a performance jump of more than 20%, fully demonstrating the huge potential of EUV technology. However, in the face of the triple jump in the price of 2nm EUV lithography machines, jumping from $120 million to $380 million, TSMC once again chose to seek the best balance between cost and technology, and insisted on using existing EUV resources on the A16 process.
TSMC's series of decisions is tantamount to a powerful revelation for China's semiconductor industry. In the context of the difficulty in obtaining EUV lithography machines, TSMC's successful case reminds us that even without the most cutting-edge equipment, by deeply tapping the potential of existing DUV lithography machines, China's chips can also launch an impact on 7nm or even 5nm processes. At present, there are signs in China that with the help of DUV lithography machine, some manufacturers have approached the 7nm process, and breakthroughs in innovative methods such as quadruple exposure technology have further enhanced the feasibility of using DUV to achieve more advanced processes. Lim, the pioneer of immersion DUV lithography machines, also expressed optimism about the potential of DUV, although he also pointed out that cost will be an unavoidable challenge.
TSMC has repeatedly proven that advanced technology does not rely entirely on the latest models of equipment, but is deeply rooted in the extreme use of existing resources and technological innovation. The close ties between TSMC and SMIC, especially the joining of former technology leader Liang Mengsong, have added confidence and expectations to the high-end process research and development of China's semiconductor industry. TSMC's strategic choice is not only a technological revolution, but also a far-reaching incentive for the independent innovation and development of China's chip industry.