As featured in Fierce Electronics
The CHIPS and Science Act of 2022, signed into law by President Biden last week, is a momentous piece of legislation.
With $52.7 billion slated for grants to support Sections 9902 through 9906 of the 2021 National Defense Appropriations Act (NDAA), Washington faces daunting next steps to define and execute on a CHIPS Act program to move the U.S. semiconductor industry toward domestic chip supply chain resiliency and manufacturing and technology leadership over the next five years.
With the growing tensions across the Taiwan Strait, the timeline for achieving domestic supply chain resiliency seems to be increasingly urgent. This is especially true for leading-edge manufacturing as the U.S. continues to trail Asia. There are nontrivial risks that need to be addressed for Washington’s near- and long-term promises for the CHIPS Act to be realized.
Leadership is complicated
Curiously, the U.S. Department of Commerce has defined advanced manufacturing as 28 nm or denser for logic chips as of the signing of the CHIPS Act. This expands eligibility of funds to a much wider swath of U.S. and non-U.S. semiconductor companies that address the chip shortage associated with trailing-edge manufacturing, especially 28 nm, the Goldilocks node for automotive chips at the moment.
It begs the question: how will the grant awards be sized and prioritized for the purpose of “advanced” semiconductor supply chain resiliency (especially for chips at less than 5 nm) compared to the objective of developing leading-edge manufacturing know-how and capacity in the U.S.? Further, how will Washington and the U.S. semiconductor industry collaborate with chip manufacturing leaders in Taiwan and South Korea to mitigate the unique risks they face having to sole source leading-edge chips from Asia?
Little help from our leading-edge friends
In November of 2020, TSMC’s board of directors approved an initial investment of $3.5 billion to build their second chip manufacturing site in the U.S. in Phoenix, Arizona. The company plans to invest a total of $12 billion into building a 5 nm fab facility and recently held their topping-out ceremony commemorating the completion of their Fab 21 building.
Samsung is in the process of investing $17 billion into a chip manufacturing site in Taylor, Texas with the expectation that it will bring additional “advanced node” capacity to the U.S. Samsung has yet to specify which process technology will be running in their new fab but state that it will be sufficient for “application in areas such as mobile, 5G, high-performance computing (HPC) and artificial intelligence (AI).”
Neither of these industry-leading chip manufacturers will be bringing the leading-edge of manufacturing process as part of their U.S. investments. TSMC’s 3N (3 nm) process will continue to be exclusively operated out of Taiwan. When TSMC’s new Arizona site comes online with 5 nm production at the beginning of 2024, the company will be on the verge of risk production of their 2 nm process in their new fab in Taiwan, leaving their U.S. site two generations behind.
TSMC and Samsung continue to double down on their investments in their chip manufacturing technology and domestic production capacity and competencies at the leading edge. TSMC is slated to invest $100 billion over three years to fortify its leadership in semiconductor manufacturing and technology. Samsung has committed to invest KRW 450 trillion ($345 billion) over five years into advanced technologies with the goal of becoming a semiconductor powerhouse contender to TSMC.
Most of these investments will happen in Taiwan and South Korea to fortify their domestic chip industry leadership and national interests in the semiconductor and broader electronics industries.
A carrot too small
Unless the smartphone supply chain is materially shifted to the U.S., we can expect leading-edge chip manufacturing to remain geographically exclusive to South Korea and Taiwan for years to come. Eighty percent of leading-edge manufacturing capacity is reserved for chips that go into premium and high-end smartphones and tablets. Most of these smartphones are assembled in China. More than three quarters of smartphones shipped globally are from Chinese OEMs, according to Counterpoint Research.
The CHIPS Act grants under NDAA Section 9902 are limited to $3 billion per “individual project.” This is likely an insufficient sum for Samsung or TSMC to offshore all that comes with chip manufacturing leadership such as premium jobs, know-how, and semiconductor industry and economic differentiation to the U.S.
Moreover, CHIPS Act grant money comes with strings attached, such as a 10-year moratorium on any covered entity from investing in chip manufacturing in countries of concern, which undoubtedly includes China. The moratorium could be a sticking point for both companies, especially given their massive electronics industry trade with China. Both companies also have billions invested in fabs located in Nanjing, Xi’an, Suzhou, and Shanghai.
Grants aside, both Samsung and TSMC projects should be able to benefit from the 25 percent federal tax credit over the next five years for their investments in their new U.S.-based fabs with fewer strings attached.
When it comes to leadership, the U.S. is on its own
It is important for U.S. policymakers to recognize that as much as Taiwan and South Korea are geopolitical allies, they and their chip industry champions are competitors who will not entirely share Washington’s agenda of reconstituting America’s chip manufacturing leadership. After all, they are the global leaders today and the future will be determined by the outcome of a global competition that they intend to win.
Comprehensive chip supply chain resiliency which includes absolute leading-edge manufacturing will not be easy to achieve on U.S. soil unless measured and sufficient investments are made to foster home-grown semiconductor technology leadership and manufacturing excellence. One thing is certain, this will take time. It’s time to get moving. TSMC and Samsung are already moving.