The success of modern AI techniques relies on computation on a scale unimaginable even a few years ago. What exactly are the AI chips powering the development and deployment of AI at scale and why are they essential? Saif M. Khan and Alexander Mann explain how these chips work, why they have proliferated, and why they matter.

The United States and its allies enjoy a competitive advantage in the production of artificial intelligence chips necessary for leading AI research and implementation. This memo identifies chokepoints for limiting China’s access to key chipmaking equipment.

Congress has advanced legislation to appropriate $52 billion in funding for the CHIPS for America Act, which aims to increase semiconductor manufacturing and supply chain resilience in the United States. But more can be done to improve the resiliency of U.S. access to microelectronics beyond manufacturing incentives. This report outlines infrastructure investments and regulatory reforms that could make the United States a more attractive place to build new chipmaking capacity and ensure continued U.S. access to key inputs for semiconductor manufacturing.

The U.S. semiconductor supply chain’s resilience will meaningfully increase only if current efforts to re-shore fabrication (that is, to situate more facilities that make its key parts in the United States) are met with commensurate efforts to re-shore upstream material production along with downstream assembly, test, and packaging (ATP) of finished microelectronics.

ETO's Supply Chain Explorer is designed to quickly orient non-experts to the essential inputs, players, and relationships involved in producing advanced computer chips. Use the Explorer to learn how these chips are made, who makes them, and the tools, materials, and processes involved in the supply chain.

China seeks to develop an indigenous semiconductor industry. It is in the strategic interest of the United States and democratic friends for China to remain reliant on them for state-of-the-art computer chips, especially as Beijing invests heavily in advanced chips.

Between 2012 and 2018, the amount of computing power used by record-breaking artificial intelligence models doubled every 3.4 months. Even with money pouring into the AI field, this trendline is unsustainable. Because of cost, hardware availability and engineering difficulties, the next decade of AI can't rely exclusively on applying more and more computing power to drive further progress.

To reduce its dependence on the United States and its allies for semiconductors, China is building domestic semiconductor manufacturing facilities by importing U.S., Japanese, and Dutch semiconductor manufacturing equipment. In the longer term, it also hopes to indigenize this equipment to replace imports. U.S. and allied policy responses to China’s efforts will significantly affect its prospects for success in this challenging task.

Protecting international security and human rights by using multilateral controls on semiconductor manufacturing equipment and advanced chips

Offshoring the production of semiconductor manufacturing equipment would remove an important source of leverage over China and make the United States more dependent on other countries for some of the most important inputs to semiconductor manufacturing. This brief explores the factors driving U.S. SME firms to offshore production and what can be done to slow or reverse offshoring.

The semiconductor industry and the U.S. government are engaged in ambitious plans to expand domestic semiconductor manufacturing capacity. Previous CSET research has found that the CHIPS for America Act incentives, if carefully targeted and augmented by adequate regulatory and workforce support, could reverse the observable decline in U.S. semiconductor manufacturing capacity since 1990. This paper argues that targeted investment incentives to increase U.S.-based advanced packaging capacity are also important for innovation, supply chain security, and ongoing semiconductor industry leadership.

CHIPS for America Act funding will result in the construction of new semiconductor fabrication facilities (“fabs”) in the United States, employing tens of thousands of workers. This policy brief assesses the occupations and backgrounds that will be most in-demand among new fabs, as well as options for ensuring availability of the necessary talent. Findings suggest the need for new immigration pathways for experienced foreign fab workers, and investments in workforce development.

The countries with the greatest capacity to develop, produce and acquire state-of-the-art semiconductor chips hold key advantages in the development of emerging technologies. At present, the United States and its allies possess significant leverage over core segments of the supply chain used to produce these chips. This policy brief outlines actions the United States and its allies can take to secure that advantage in the long term and use it to promote the beneficial use of emerging technologies, such as artificial intelligence.

As artificial intelligence is applied to new and more complex tasks, the computational power necessary to develop and deploy it will become increasingly expensive. This policy brief offers a concise overview of the full report, “AI Chips: What They Are and Why They Matter.”

As an integral player in advanced semiconductor supply chains, the United States enjoys advantages over China in producing and accessing chips for artificial intelligence and other leading-edge computing technologies. However, a lack of domestic production capacity threatens U.S. semiconductor access. The United States can strengthen its advantages by working with allies and partners to prevent China from producing leading-edge chips and by reshoring its own domestic chipmaking capacity.

Since 1990, the U.S. share of global semiconductor manufacturing capacity has declined while the shares of China, South Korea, and Taiwan have increased. If carefully targeted, CHIPS for America Act incentives could reverse this trend for the types of chips that matter most to U.S. national security. In this policy brief, the author assesses how CHIPS Act incentives should be distributed across different types of chips.