The semiconductor has become the new oil — a strategic resource whose supply chains determine national economic competitiveness and military capability. The US CHIPS Act ($52.7 billion), the EU Chips Act (EUR43 billion), Japan's ESPA, and South Korea's K-Chips Act represent the largest coordinated industrial policy interventions in a generation. But unlike previous technology races, this one is defined not by who can innovate fastest but by who controls the chokepoints in a globally interdependent production chain.
The Economy-Security Nexus
Kim and Rho (2024), in the Journal of Chinese Political Science, analyze the US-China chip war through the lens of the economy-security nexus — the growing entanglement of economic competition and national security concerns that is reshaping technology governance across Asia. Their analysis shows that semiconductor export controls are not merely trade policy but geostrategic tools designed to constrain a competitor's technological trajectory.
The implications for Asian economies caught between the US and China are severe. Countries like South Korea, Taiwan, and Japan host critical segments of the semiconductor supply chain — design, fabrication, equipment manufacturing, packaging — and face pressure from both sides to align. The chip war is fragmenting what was once a globally integrated innovation system into competing blocs, each pursuing technological self-sufficiency at the expense of the efficiency gains that integration provided.
Chokepoint Theory
Liu and Lin (2025), in the Journal of World Investment and Trade, develop the concept of techno-geopolitical chokepoints — specific stages in the semiconductor production chain where a small number of actors control essential capabilities. ASML's monopoly on extreme ultraviolet lithography equipment is the paradigmatic example: without EUV machines, no entity on Earth can produce cutting-edge chips, and ASML is a single company in a single Dutch city.
The chokepoint framework reveals that semiconductor sovereignty is not achievable through investment alone. A country can build fabrication plants, but if it depends on foreign-controlled equipment, materials, or design tools, its sovereignty is nominal. True technological self-sufficiency would require replicating capabilities that took decades and tens of billions of dollars to develop — a cost that even major economies may find prohibitive.
The CHIPS Act Assessment
Hyatt et al. (2024), in Digital Policy, Regulation and Governance, assess the US CHIPS and Science Act against the reality of semiconductor production economics. Their analysis finds that while the CHIPS Act represents an important signal of strategic intent, its $52.7 billion investment is modest relative to the scale of the challenge. TSMC's single Arizona fab costs approximately $40 billion. Building a genuinely self-sufficient semiconductor ecosystem would require sustained investment orders of magnitude larger than current commitments.
The deeper challenge is talent. Semiconductor fabrication requires specialized engineering expertise that takes years to develop and is concentrated in a few global clusters — Hsinchu, Seoul, Eindhoven. Capital can be allocated by legislation; expertise cannot. The CHIPS Act addresses this through workforce development provisions, but the gap between current capacity and strategic ambitions remains substantial.
The chip wars illustrate a broader truth about innovation policy: in a globally interdependent economy, national innovation strategies must account for the entire value chain, not just the segments within national borders. Sovereignty over a single stage of production provides limited advantage when every other stage depends on foreign actors whose cooperation cannot be guaranteed.
The Talent Bottleneck
Beyond capital investment, the semiconductor sovereignty race faces a talent constraint that money alone cannot solve. Advanced chip fabrication requires specialized engineers with expertise in materials science, chemical engineering, precision manufacturing, and process integration, skills that take years to develop and are concentrated in a few global clusters. TSMC alone employs over 70,000 people, many with decades of fabrication experience that cannot be replicated through training programs.
The CHIPS Act allocates funding for workforce development, but the timeline for producing experienced semiconductor engineers is measured in decades, not budget cycles. South Korea and Taiwan have built their semiconductor workforces over 40 years of continuous investment in education, apprenticeship, and industrial learning. Replicating this capability in new locations requires patient, sustained commitment that democratic political systems with short electoral cycles may struggle to maintain.
This talent constraint has implications beyond individual countries. The global semiconductor workforce is a fixed resource in the short term. When the US, EU, Japan, and India all simultaneously pursue semiconductor expansion, they compete for the same pool of experienced engineers, potentially driving up costs and slowing progress. The chip wars may ultimately be won not by the country that invests the most capital but by the one that develops the most talent.