Why AI's Infrastructure Boom Masks Fragile Supply Chain Dependencies
Harvard economist Willy Shih explains how the AI buildout exposes the uncomfortable economics of global manufacturing that policymakers continue to ignore.
Last verified:
The Supply Chain Blindspot Behind AI Infrastructure
According to The Verge AI’s recent interview with Willy Shih, Harvard Business School economist and author of Where Does My Stuff Come From?: The Powerful Economics Behind Global Supply Chains, the AI sector’s explosive demand for computing resources is colliding with a reality that most policymakers have yet to internalize: global supply chains operate on principles of economic interdependency that cannot be quickly reversed through tariffs or reshoring mandates. Shih’s analysis suggests that the current political moment—marked by renewed protectionism, trade tensions, and reshoring rhetoric—fundamentally misunderstands the structural economics underlying advanced manufacturing.
Why First Principles Predict Policy Failure
Shih’s predictive track record offers a lens for understanding why reshoring promises rarely materialize. According to The Verge AI, when covering Foxconn’s 2019 proposal to build an LCD manufacturing facility in Wisconsin, Shih immediately identified a fatal flaw: the absence of upstream suppliers. An LCD fab requires not just a factory building, but an integrated ecosystem of glass suppliers, chemical manufacturers, and specialized infrastructure providers. Without these components already embedded in the regional economy, a standalone fab becomes economically unviable.
This principle extends to the current AI infrastructure buildout. The rush to produce AI chips and accelerators (GPUs, TPUs, and next-generation silicon) depends on a similarly complex web of suppliers, materials, and logistics networks. According to The Verge AI, Shih’s analysis reveals that policymakers “insistently ignore” these structural constraints, leading to policies that look sound on paper but fail in execution.
China’s Durable Advantage Is Not About Labor
A critical insight from Shih’s framework, discussed in The Verge AI interview, is that China’s continued dominance in advanced manufacturing is not primarily attributable to cheap labor—a misconception that shapes much Western policy thinking. Instead, Shih emphasizes the accumulated supply chain density: China has invested decades building the ecosystem of materials suppliers, logistics infrastructure, skilled technical workforces, and specialized fabs that allow rapid iteration and scaling. This ecosystem advantage persists even as labor costs have risen, making China difficult to displace despite political pressure and tariff threats.
For AI infrastructure specifically, this means that attempts to rapidly build domestic chip-manufacturing capacity in the US, Europe, or other regions face a bootstrapping problem: you cannot simply transplant a fab and expect it to function at scale without the surrounding supply chain.
Why This Matters
As the AI sector demands exponentially more compute—driving orders for advanced semiconductors, power systems, and specialized cooling infrastructure—the fragility of globalized supply chains becomes a competitive and national-security issue. According to The Verge AI, Shih’s work suggests that neither reshoring nor continued dependence on existing supply chains offers a simple solution. Instead, the challenge is understanding the deep economics of manufacturing interdependency and building policy around that reality rather than around nationalist sentiment or tariff rhetoric.
For technology companies, chip manufacturers, and infrastructure investors, this implies that supply chain risk in the AI era is not primarily about geopolitical disruption—it is about the slower, more fundamental problem of whether the ecosystem supporting advanced manufacturing can expand fast enough to meet AI demand. That answer remains uncertain, and most decision-makers are still operating from incomplete mental models of how these systems actually work.
Frequently Asked Questions
Why did Willy Shih predict the Wisconsin Foxconn LCD fab would fail?
Shih applied first-principles supply chain analysis: an LCD fab requires integrated upstream suppliers (glass plants, specialized materials, chemical infrastructure) that simply did not exist in Wisconsin. Without the ecosystem, a standalone fab cannot operate economically.
What does Shih's analysis reveal about reshoring policy?
Reshoring initiatives have largely failed because policymakers underestimate the deep interdependencies required to support advanced manufacturing. Building a factory is not enough—you need the entire supply chain ecosystem.
How does AI's computing demand affect global supply chains?
AI's infrastructure buildout is creating unprecedented demand for chips and components during a period of political tension over globalization and tariffs, straining a system that already struggles with complexity that most regulators do not fully grasp.