By Mustafa Mayar, Spring 2026 Marcellus Policy Fellow

Semiconductors now sit at the center of U.S. national security because they power advanced computing, artificial intelligence, advanced weapons systems, and much of the civilian economy. Recent supply-chain shocks and intensifying competition with China have pushed Washington to treat chips not only as commercial goods but also as strategic assets. That shift is understandable. The problem is that current U.S. policy often collapses two different goals into a single framework. Resilience concerns how the United States maintains access to critical inputs, reduces dangerous dependencies, and strengthens its domestic industrial base. Denial is about how the United States and its partners can slow China’s access to a narrow set of frontier technologies with military and strategic significance. These goals can reinforce one another, but they are not the same and should not be treated as interchangeable.

This paper argues that the United States should pursue selective semiconductor denial rather than broad decoupling from China. The case for this approach begins with the structure of the semiconductor industry itself. U.S. and allied leverage are real, but it is concentrated in specific chokepoints rather than spread evenly across the full supply chain. The most important chokepoints include electronic design automation (EDA) tools, advanced semiconductor manufacturing equipment, and frontier fabrication capabilities. These are the parts of the system where outside access can shape the development of cutting-edge chips and where Chinese substitution remains most difficult. Broad decoupling would weaken the conditions that make this leverage effective by accelerating China’s self-reliance drive, straining allied coordination, raising costs for U.S. firms, and adding fragility to an already concentrated system.

The paper recommends a restrained semiconductor strategy built around five steps. The United States should limit the strictest controls to frontier capabilities, review semiconductor controls regularly, build domestic resilience without pursuing full autarky, coordinate allied controls around specific chokepoints, and preserve civilian flows where the security case for restriction is weak. This approach does not assume that the old model of frictionless globalization can return, nor does it deny that some parts of the semiconductor system now require tighter control. It argues instead for a managed and selective form of fragmentation, one that protects U.S. security interests without turning the entire semiconductor economy into a permanent site of confrontation.