As featured in Fierce Electronics
The CHIPS and Science Act of 2022, signed Tuesday by President Biden, includes the stated purpose of challenging China’s pursuit of technological and manufacturing leadership in the global semiconductor industry. The bill easily passed both the Senate (64-33) and the House (243-187) with bipartisan support in both houses.
While the passage of the CHIPS Act is a historic achievement that directly addresses the need for chip supply chain resilience add capacity in the U.S., it is also a timely and important commitment to a much longer and comprehensive policy effort to ensure long-term U.S. technology and market share leadership in the vital semiconductor industry.
As President Joe Biden stated in his Rose Garden speech, “We are at an inflection point.” The seemingly worsening U.S.-China trade and geopolitical relations and the Coronavirus pandemic have highlighted profound national security and economic risks that the United States is exposed to with the current state of the global semiconductor supply chain.
Persistent chip supply instability due to serial demand shocks around the globe has impacted numerous domestic industries since the advent of the Coronavirus pandemic. Most notably, the auto industry continues to suffer periodic production shutdowns due to shortages of largely low-cost mature node chips.
The hope in Washington is that the CHIPS Act and the provisions in the 2021 National Defense Authorization Act (NDAA) that it funds will help the U.S. establish a degree of domestic chip self-sufficiency and supply chain resiliency. The worsening situation across the Taiwan Strait has heightened the need to reduce the mounting risk facing U.S. industries that are increasingly dependent on semiconductors of all varieties sourced from Asia.
For this purpose, the new law budgets $39 billion in incentives to “covered entities” for the development of domestic U.S. semiconductor manufacturing with $37 billion of the sum allocated to advanced manufacturing designated by the U.S. Department of Commerce as 28nm density or higher. The remaining $2 billion is earmarked for “legacy chip production”.
The CHIPS Act includes an additional $13.2 billion in appropriations for research and development in the areas of advanced packaging and secure chips, programs to accelerate the commercialization of chip manufacturing and design innovations, and workforce development programs to help fill the 90,000 semiconductor jobs projected by 2025.
A new global equilibrium
Beyond chip supply chain resiliency, the CHIPS Act aspires to restore U.S. chip manufacturing leadership. In this regard, the newly minted law is about much more than China. It’s about the broader dominance of Asia as the epicenter of the global electronics industry. Whether its Apple, Intel, AMD, Nvidia, Applied Materials, Cadence, or Microsoft, almost every U.S. “tech sector” company is highly dependent on the complex electronics (including microelectronics) supply chains that circulate heavily across Asia and transverse China.
Shifting the global electronics supply chain back toward the West in an economically sustainable and competitive way will be no easy task or trivial investment. It is yet to be seen whether the $53 billion CHIPS Act and 25 percent tax credit for investments in chip manufacturing and R&D will be enough to nudge the status quo to a more equitable and secure regional balance.
The challenge? Washington will have to reckon with the billions in semiconductor industry trade that passes in and out of China. According to BCG and SIA, over $400 billion in semiconductor trade flowed in and out of China in 2019. And 32 percent of that trade ended up in finished electronics that landed in North America, mostly in the United States.
The U.S. will also have to contend with economics that favor Asia, especially China. BCG estimates that the TCO (Total Cost of Ownership) over a 10-year investment period for chip manufacturing in China is two-thirds of that in the United States.This will likely make chips more expensive to produce in the U.S. and will likely require sustained subsidies beyond the CHIPS Act’s 5-year timeframe for domestic chip production to be competitive in the global market.
Nevertheless, the incentives and investments outlined in the CHIPS Act are long- and near-term table stakes in leveling the playing field vis-à-vis Asian rivals, in particular China. They are also essential in realizing a more self-sufficient and globally competitive domestic semiconductor industry.
At a minimum, a level of chip supply and supply chain resiliency that garners industrial and economic stability in the face of crisis and disruption is painfully needed. The U.S. is clearly not where it needs to be, yet.
Manufacturing leadership through partnership
Though a recent TechInsights report suggests that China’s largest domestic chip manufacturer, SMIC, has ramped its 7nm N+1 process in production for more than a year, China is far from a leader in both the chip manufacturing and technology races. They would need access to the latest and greatest fab equipment and tooling currently being restricted by Washington if they were to be a contender in the long run.
Taiwan’s TSMC and South Korea’s Samsung are the established leaders in leading-edge chip manufacturing with generations of experience with ASML’s EUV (Extreme Ultraviolet) lithography technology for logic and memory. Samsung recently started risk production of their 3nm process with GAAFET (Gate-All-Around Field Effect Transistor) which is a first in the industry for this transistor architecture putting them ahead of the field for the moment.
Home-grown chip manufacturing leadership rests heavily on the shoulders of Intel which is expected to receive a significant share of appropriated CHIPS Act funding and tax credits. However, the company has struggled for years with its 10nm process and continues to suffer execution challenges and delays getting Sapphire Rapids (10nm Enhanced SuperFin) and its Intel 4 (7nm) processes (Intel’s first foray with EUV technology) into production and on schedule.
Intel is targeting 2024 to “usher in the Angstrom Era” with Intel 20A which will implement the company’s RibbonFet (Intel’s GAAFET equivalent) and PowerVia technologies. The company’s drive to restore its manufacturing leadership will be predicated on exceptional execution that the organization has yet to demonstrate under the auspices of CEO Pat Gelsinger. Much is riding on the company getting its IFS (Intel Foundry Services) business humming and product roadmaps on track and on schedule. Once they do, it will be a great boon for U.S. technological leadership in chip manufacturing.
In order to meet the manufacturing leadership and security intents of the CHIPS Act in the near term, Samsung and TSMC will need to throw their hats in the ring to build leading-edge fabs in the U.S. Securing the support and partnership of these two leaders will be important for realizing domestic production of leading-edge chips that are largely produced in Taiwan today on the order of 90 percent.
Investment in chip technology leadership
Building chip manufacturing capacity and capabilities in the U.S. is a strategically important objective of the NDAA and CHIPS Act. However, it is equally important for the U.S. semiconductor industry to continue leading in strategic areas where it currently leads such as logic IP, chip design, EDA (Electronic Design Automation) tools, fabrication equipment, and foundational research.
Why? These strengths not only foster sustainable technological leadership at a global level, but they also provide formidable supply chain and trade leverage against China which the Department of Commerce has exercised in its campaign to limit Huawei’s access to advanced semiconductors.
While chip manufacturing in the U.S. has dwindled from a global share of 37 percent in 1990 to 12 percent in 2020, fabless chipmakers such as AMD, Nvidia, and Qualcomm have led the world in logic design for advanced applications such as the next generation of advanced vehicular systems, exascale computing, new and increasingly powerful Edge AI applications, and 5G and 6G mobile communications systems.
However, Huawei’s pre-sanctioned HiSilicon and Taiwan’s Mediatek continue to demonstrate formidable chops designing competitive Arm-based chips for a wide range of consumer devices including smartphones. In the case of Huawei, the company continues to design chips for industry-leading network equipment including those used in 5G mobile networks.
Thankfully, President Biden acknowledged the important role of chip designers like Qualcomm in U.S. semiconductor industry leadership. Yet, there is no direct appropriation of funds in the CHIPS Act for sustaining strategic U.S. leadership in chip design. That being said, Section 9903 of the 2021 NDAA specifies avenues for fabless chip design firms to engage in $2 billion in projects yet to be determined under the CHIPS Act for securing microelectronics and chip manufacturing.
Additionally, Section 9906 establishes a $2 billion investment fund that supports private sector collaborations between startups, academia, and established companies to speed the commercialization of innovations that meet the objectives of the National Semiconductor Technology Center (NSTC) yet to be established by the Secretary of Commerce in coordination with the Secretary of Defense.
One of the proposed initiatives by the President’s Council of Advisors on Science and Technology (PCAST) is the development of a zettascale supercomputer which could be the CHIPS Act’s Project Apollo, the NASA program that landed the first man on the Moon.
A needed and welcome response
There is no doubt that the U.S. government had to do something to improve the security and resiliency of the nation’s semiconductor supply chain against the backdrop of increasingly acrimonious relations between the U.S. and China, and destabilizing pandemic waves and regional conflicts.
The CHIPS Act is an important and welcome legislative achievement that meets the growing challenge of a highly competitive Asia region that continues to invest heavily in advancing its semiconductor industry, technology, and regional ecosystem.
The question remains, is the CHIPS Act enough? If the goal is improving manufacturing supply chain resiliency and security, maybe. Time and quality of investment, administration, and execution will tell.
If the ultimate goal of the new law includes sustained technological and market leadership, it is likely that the CHIPS Act is but the first act in what will be a prolonged public-private partnership, one that will need to shift focus on areas such as R&D, workforce, and other investments key to maintaining the U.S.’s technological edge in chips.
Such continued investments and well-directed government support will determine if the U.S. pioneers the future of the semiconductor industry in the decades to come.