Chips and Chokepoints: Chris Miller on the Geopolitics of the AI Supply Chain
By Jess Regan
As the world rapidly adapts to the expanding use of artificial intelligence, existing vulnerabilities in global supply chains have become chokepoints stunting advancement. Discussing these pressing developments, Chris Miller—a professor at the Fletcher School at Tufts University and a nonresident senior fellow at the American Enterprise Institute—recently served as the featured guest for the Carnegie Mellon Institute for Strategy & Technology’s Scientists & Strategists Speakers Series at 好色先生TV. Miller argues that a factor of critical importance today is how major powers leverage these supply chains to control the world’s digital infrastructure.
In his February 26 lecture, entitled “Geopolitics of AI Supply Chains,” Miller focused on the importance of the technology, why supply chain bottlenecks exist, how they emerged, and the underlying relationship dynamics at play. He argued that no issue is more pressing today than the relationship between the world’s two largest economies—the United States and China—and their militaries in the context of the AI supply chain.
Miller opened his talk by sharing a prompt that he had fed to ChatGPT, asking how AI technology actually works. While the response detailed models, data, and architecture, it notably omitted what Miller considers the primary driver of AI growth: computing power. Although the cost of computing has decreased dramatically over the decades, the foremost challenge for technology executives today remains the acquisition of the chips and electrons necessary to sustain the expanding industry. Miller emphasized that the largest US tech companies “are struggling to acquire larger and larger volumes of computing power to build vast data center campuses full of the most advanced processor and memory chips” to scale their AI operations. Designed by industry leaders like NVIDIA and produced through global supply chains, these high-end chips and GPUs have become the world’s most critical commodity. They are the essential drivers of technological progress.
Guiding the audience through a short history of the technology, Miller noted the essential role of chips in electronic warfare, communications, radar, space systems, and intelligence and military operations. He stressed that despite decades of change, the world's most powerful militaries continue to view access to advanced microelectronics as vital to modern strategic competition. While the United States has become increasingly overt about its use of AI for defense, China’s procurement documentation reveals large-scale purchases of GPU chips, and publicly available information literature suggests that Russia is actively trying to keep pace by leveraging open-source models for intelligence.
What makes this technology unique, Miller argued, is a reliance on supply chains that—for the first time in history—cannot be managed by a single nation. Instead, these goods are produced across multiple countries and continents. “Today there's not a single country in the world that can produce, not even an advanced chip, but a moderately advanced chip on its own,” he said. Instead, these goods are produced across multiple countries and continents. Miller underscored this point by sharing data regarding the various stages of the semiconductor supply chain and market share held by key geographies, including specialized segments such as electronic design automation (EDA), the production of pure silicon wafers, and the manufacturing of extreme ultraviolet (EUV) lithography machines. The data illustrated that even the US—which holds the largest overall share of the global chip industry—is fundamentally reliant on Europe, Japan, Korea, Taiwan, and China for critical links in the chain. Notably, Miller pointed out that China currently holds only a small portion of the supply chain. This is a surprising outlier given its vast manufacturing capabilities. Despite significant efforts, China has yet to successfully produce the most advanced equipment and chips domestically.
Miller also identified several chokepoints where a single region possesses capabilities that virtually no one else can replicate, noting for example that the global supply chain remains highly vulnerable to a devastating earthquake in Taiwan. Furthermore, he highlighted the widespread nature of this dependency, noting that these chips are now essential to the broader economy—powering everything from smartphones and computers to the automotive and industrial machinery sectors. As Miller put it, “As costs of AI decline, we're just going to put more of these capabilities in an ever-larger number of our devices.”
Beyond technical bottlenecks, Miller contended that the semiconductor landscape is being reshaped by the demands of global political leaders seeking to align supply chains with their own national interests. Miller previously traced this power struggle in his 2022 New York Times bestseller, Chip War, which underscores how control over computing power has historically dictated military and economic advantage. This trend has only intensified in 2026, as evidenced by the rising use of tariffs and tightening of export controls and trade regulations as nations race to secure their AI futures.
Closing his lecture by addressing the escalating national security concerns surrounding these technologies, Miller indicated that global powers are increasingly wary of foreign-made technologies capable of autonomous processing or independent communication. He noted that this skepticism raises critical questions about who has access to these autonomous systems, how they operate, who writes the software, and who provides the over-the-air updates—questions for which there are currently no clear answers. Ultimately Miller observed that while the primary concern regarding semiconductors today is access, the defining question of tomorrow will be trust.
After the lecture, University Professor of Economics and Public Policy and Provost Emeritus Mark Kamlet joined Miller to expand on the topic in a dialogue moderated by CMIST Director Audrey Kurth Cronin. When Kamlet asked what a decisive “win” in the technological race would look like for the US, Miller described a multifaceted victory, including the rapid deployment of AI, the harnessing of productivity growth, and ensuring that US tech firms remain at the forefront of the industry to maintain regulatory influence. He also emphasized the need to deploy AI for defense and intelligence faster than competitors—all while maintaining global stability. As Miller noted, the goal is not just about making the goods, but “trying to make the rules that make the goods.”
The dialogue then pivoted to economic security, regulatory policy, and domestic political imperatives, such as increasing US manufacturing. Miller observed that because the collective manufacturing output of nations like China, Japan, Canada, and Mexico is so vast, the US inevitably will remain dependent on these foreign supply chains. However, he stressed the importance of “supply elasticity.” For goods with multiple sources or those where production can be easily scaled, reliance is less of a concern. In contrast, semiconductors are far harder to source domestically; therefore, the US must focus on building capacity for specific constrictions that pose the greatest risk to the supply chain.
The event concluded with a Q&A session, where the audience largely focused on how US regulations and export controls might affect the global AI race. Miller argued that historical controls have significantly altered the playing field, suggesting that without them, China might have already secured a position as a global leader in AI. While he acknowledged that the US would be an outlier if it weren't taking steps to protect its interests, Miller ended with a powerful reminder that an enduring advantage of the United States is not just the hardware, but the people. He observed, “The transfer of talent from China to the US has been the greatest loss for China and greatest win for the US and we want that to continue.” In an era defined by technological competition and fragile supply chains, Miller’s insights suggest that the future of the AI race may depend as much on people as it does on processors.
(Image 1: Chris Miller is a professor at the Fletcher School at Tufts University and a nonresident senior fellow at the American Enterprise Institute; Image 2, l to r: CMIST Director Audrey Kurth Cronin, Chris Miller, and Mark Kamlet, University Professor of Economics and Public Policy and Provost Emeritus; Image 3: Cronin and Miller during the February 26 Scientists & Strategists event; Image 4: Kamlet speaks during the moderated discussion; Image 5, l to r: Cronin and Miller; Image 6: Kamlet was the discussant for the event)