Hou Yongqing, senior deputy general manager of TSMC, said that in response to the outbreak of demand for AI and high-performance computing (HPC), TSMC is advancing its production expansion plan at "twice the speed" compared to the past. This year, TSMC will have five 2nm wafer fabs entering the production capacity ramping stage at the same time, writing the most aggressive production expansion record in the past. Benefiting from this, the output of 2nm in the first year will increase by approximately 45% compared to the same period of 3nm.
TSMC mentioned in the earnings call that in order to meet the strong demand for AI applications, the company is increasing capital investment to expand 3nm production capacity.
Among them, the Tainan Science Park 3nm factory is expected to enter the mass production stage in the first half of next year; the second factory in Arizona, the United States, has completed construction and is expected to start mass production of 3nm wafers in the second half of next year; and the second Kumamoto factory will also include the 3nm process and is expected to enter mass production in 2028.
In addition, TSMC plans to build an A10 wafer fab in Tainan, Taiwan. The P1 to P4 fab areas will be used to develop advanced process technologies of 1nm and below. Trial production is expected to start in 2029, with an initial monthly production capacity of 5,000 wafers.
