Having TSMC build production facilities and manufacture chips in the U.S. should give a shot in the arm to U.S. hopes for an improved domestic semiconductor industry. One issue that won’t help is the 30% higher production costs that TSMC will have to pay to produce the 4nm chips in the U.S. This should result in higher prices for TSMC’s U.S. clients.
Initially, under Phase 1, TSMC expects to produce a wafer output of 20,000 units a month. Customers are expected to include Apple, Nvidia, AMD, and Qualcomm. Part of the reason for the higher costs in the U.S. has to do with the lack of semiconductor supply chains in the U.S. Phase 2 is more ambitious as it calls for TSMC to manufacture 2nm chips in 2028.
Two other fabs are being built in the Phoenix area with the second one set to start manufacturing 2nm chips by 2028. While TSMC’s Taiwan facilities will start mass production of 2nm chips during the second half of next year, by 2028 the company could be manufacturing 1.6nm chips in Taiwan. The reason the process node is so important is because as that number shrinks, so does the size of the transistors used to power a chip.
Smaller transistors usually mean a higher transistor count and transistor density for a particular chip. Typically, the higher a chip’s transistor count, the more powerful and/or energy-efficient that chip is.
GIPHY App Key not set. Please check settings