Seemingly Hasty, but Prudent TSMC's New Fab Plan in the U.S.
Exhibit 1: The Comparison of TSMC Fab 18 and New Fab in Arizona
The share of 5nm in TSMC’s production is expected to reach 10% and 38.5% in 2020 and 2024, respectively. 3nm is expected to start mass production by 2022 and will take at least 2 to 3 years to ramp up. The chart below is Counterpoint Research's forecast of TSMC's process distribution over the next 5 years based on several assumptions and detailed analysis. Please contact Counterpoint to get more insights.
Exhibit 2: Forecast of TSMC Revenue by Technology, 2018-2025
The advantages of building new fab in U.S.
- As a company that plays a key role in the global electronics supply chain, TSMC has always had a need to diversify its manufacturing sites. TSMC has been Taiwan's largest listed company by market value, and its main production base is also in Taiwan. As a result, the source of talent will gradually limit its growth. TSMC has a 12-inch wafer fab in Nanjing, China, with a 12/16nm process, however, TSMC has been unable to expand its China facility amid sensitive geopolitical issues. In addition, under the constraint of Wassenaar Arrangement, TSMC has difficulty in migrating its process node in its China fab to 7nm (EUV) and below.
- Another reason why TSMC needs to diversify its production base is the concern of the natural environment in Taiwan. Taiwan is in a seismically active zone, on the Pacific Ring of Fire. In the past, the "921 earthquakes" of 1999, with a magnitude as high as 7.3 on the Richter scale, caused a certain degree of damage to TSMC's production. After than, TSMC has accumulated sufficient experience in preventing hazards caused by earthquakes. However, with TSMC now accounting for more than 50% market share and playing an even more critical role in the global electronics supply chain than in 1999, it is still prudent to diversify its production base.
- Over the next five years, the demand drivers for semiconductors will be automotive and consumer electronics. In order to achieve electrification, wireless connectivity, and autonomous driving, automotive electronics will require a large number and variants of semiconductors. The chips used in autonomous vehicles will require high-speed, but they don't have to be as small as the chips in smartphones or laptops. On the other hand, reliability is very critical. Therefore, the manufacturing process node required for automotive electronics often doesn't have to be the most advanced but has to be mature and reliable.
- The majority of TSMC’s manufacturing equipment is from American companies. According to our latest data, the share of the top three US semiconductor equipment companies (including AMAT, Lam Research, KLA) in total was more than 40% in 2019. Therefore, TSMC is able to get immediate support and is able to work efficiently with equipment suppliers to retrofit the equipment. In addition, the United States is also TSMC's largest market. In 2019, the United States accounted for 45% of TSMC's business. It can likely also get tax breaks and other incentives from Arizona.
- The initial monthly capacity of TSMC's Arizona plant is only 20,000 wafers. Therefore, other suppliers in the ecosystem are unlikely to follow TSMC to open a branch in the United States. However, if the operation of the new fab is running well and the demand is growing, it's expected to see them follow TSMC to build a complete ecosystem there.
- Unlike memory manufacturing, the operation of foundry fab must be flexible in their production line to meet the different requirements for a variety of products while also maintain a high utilization rate. However, the new process development and initial mass production normally requires engineers to work overtime. In addition to the lower salaries than U.S. engineers, Taiwanese engineers have more flexible working hours and can complete time-consuming tasks such as new process optimization and troubleshooting. However, 5nm will be a relatively mature manufacturing process in 2024 and this new fab in the US is supposed to be highly automated, so there will likely be even less extra work for engineers in the new fab.
Receive our insightful weekly newsletter and stay ahead of the competition.
Author
Brady Wang
Hi, I’m Brady Wang, a seasoned professional with over 20 years of experience in the high-tech industry, spanning semiconductor manufacturing, market intelligence, and strategic advisory roles. Currently, I serve as an analyst at Counterpoint Research, where I specialize in semiconductors with a focus on advanced applications such as automotive, server platforms, and cutting-edge process nodes. My core research centers on AI servers and their key components, including GPUs, custom accelerators, high-bandwidth memory (HBM), CPUs, and advanced packaging technologies. I also track the evolution of AI server architectures, interconnect technologies, and data center deployment trends. By combining deep technical knowledge with market insight, I help clients navigate the fast-changing AI infrastructure landscape and make strategic, data-driven decisions.