Photo credits: William W. Potter/Adobe Stock
You can listen to this episode on the CSIS website, or wherever you get your podcasts.
In this episode of GeoTech Wars, Kirti and Andrew discuss the state of the U.S. semiconductor industry and the CHIPS Act with Dan Kim, Chief Economist and Director of Strategic Planning and Industry Analysis at the CHIPS Program Office in the Department of Commerce. Prior to joining the CHIPS Program Office, Dan was Vice President and Chief Economist at South Korean chipmaker SK Hynix, and before that he was Director of Economic Strategy at Qualcomm.
As past guests have highlighted, technology leadership is essential for the United States’ economic and national security in the modern age. Technologies such as semiconductors, quantum computing, and artificial intelligence are considered “foundational technologies” as they are applicable across a wide variety of domains, including defense. Semiconductors, for example, not only power the smartphones, computers, and automobiles we use today, but also the cutting-edge technology employed by modern militaries. Ultimately, technology leadership is at the heart of the United States’ ability to compete internationally, both in economics and national security.
In the semiconductor industry specifically, which was born in California, the United States has retained leadership in certain segments, but not in others. For instance, U.S. companies have done fantastically well in chip design, but have lost leadership in chip manufacturing. Even when U.S. companies do manufacture chips, they often do so in fabrication plants located in Asia.
Accordingly, the share of semiconductors produced in the United States has fallen from 37 percent in 1990 to 12 percent in 2021.
As recognized in the CHIPS Act, semiconductor manufacturing is key to cultivating technology leadership. The CHIPS Act commits $39 billion to manufacturing incentives and over $13 billion to workforce development, along with an up to 25 percent tax credit for investments in semiconductor manufacturing facilities. Within the semiconductor industry this is a significant but not massive amount—TSMC, the world’s largest semiconductor manufacturer, spent over $225 billion on operating costs in 2022—meaning the ultimate success of the CHIPS Act will come down to its implementation.
As Dan explains, to effectively develop semiconductor manufacturing leadership, there are three key factors the United States must consider when deploying CHIPS Act funds.
First, leadership in the semiconductor industry is highly dependent on economies of scale. There is simply a lot of learning by doing, whether that be designing chips, manufacturing chips, or packaging chips. Therefore, it is incredibly difficult for newcomers to catch up to the leaders as the leaders typically operate at higher volumes and can therefore identify and implement process improvements more quickly. As in most things, practice makes perfect, and the biggest companies can just practice more.
Second, semiconductor leadership depends on understanding the needs of the customer. At the end of the day, this is about private enterprises selling products to customers. And what customers want has shifted over time. At first, chip development was vertically integrated, with companies designing and manufacturing chips for their own end products. But, over time, specialization naturally increased. Today, there are design-only companies, manufacturing-only companies, and, critically, companies who do not produce their own chips but instead buy chips to insert in their end products. In other terms, the semiconductor industry is incredibly dynamic in both technology innovation and business innovation. The United States must develop not just great chips, but a reliable and vigorous business ecosystem which can adapt to customers’ needs.
And finally, long-term technology leadership depends on developing the technical know-how to produce semiconductors in a cost-effective manner. Nothing can function without the right workforce. While it is straightforward to measure progress by the number of new fabrication plants coming online, it is imperative to also consider how human capital is being developed so that the United States can innovate in a competitive landscape for decades to come.
Given the global interconnectedness of the semiconductor supply chain, however, the United States must be realistic in its goals. Complete self-sufficiency is not only impractical but may be detrimental. Research and development (R&D) costs for new semiconductors are incredibly high, meaning leading semiconductor firms must sell their products globally to earn enough revenue to finance the next generation of semiconductors. Of TSMC’s $225 billion in operating costs in 2022, over $160 billion was spent on R&D.
Ultimately, the United States’ goal is resiliency. Producing critical semiconductors at a sufficient volume to protect its core interests, while also boosting the competitiveness of its semiconductor industry in a global environment of heavy government subsidies. That does not mean insuring against all black swan events which would be prohibitively expensive, but cultivating the capital and know-how to innovate independently. The CHIPS Act is a seminal moment in U.S. industrial policy in a rapidly changing geopolitical landscape. If the CHIPS Act achieves its goals, it may make the case for not only another CHIPS Act in the future, but similar initiatives in other critical technology industries.
This piece summarizes the discussion in GeoTech Wars, "Achieving Technology Leadership with CHIPS Office Chief Economist Dan Kim.” It does not represent the opinions of the hosts.
