A RAND Corporation paper asks whether U.S. export controls on advanced semiconductors can build a U.S.-led global artificial intelligence ecosystem. The title poses a question that frames an ongoing policy debate.

The analysis, titled "Can Export Controls Create a U.S.-Led Global Artificial Intelligence Ecosystem?", joins RAND's continuing work on U.S. technology and industrial strategy. RAND is a federally funded research and development center based in Santa Monica, California.

What is RAND actually asking?

The framing pairs "export controls" with "ecosystem" in a single interrogative. It treats hardware restrictions as one variable inside a larger system. RAND's analysts appear to question whether the existing regime can deliver outcomes beyond slowing competitor access.

The U.S. export control regime on AI chips rests on rules administered by the Commerce Department's Bureau of Industry and Security. The rules restrict shipments of certain advanced processors, semiconductor manufacturing equipment, and electronic design automation software. China remains the principal target.

Why does the ecosystem question matter?

Ecosystem dominance in AI requires more than control of leading-edge logic. It depends on cloud infrastructure, model development tooling, advanced packaging, intellectual property licensing, and deep talent pipelines. RAND's framing shifts U.S. policy from a defensive posture to a constructive one.

Semiconductor companies operate across dozens of countries. The same wafer often crosses borders several times before reaching a finished system. That interdependence is the point that industry groups have pressed on Washington for years. Coordinated allied controls are harder to evade than unilateral ones.

What does the existing regime look like?

The Biden administration issued the first round of advanced AI chip controls in 2022. It tightened the rules roughly a year later. The current administration has reviewed the regime. Reports indicate further restrictions on AI software are under discussion at Commerce.

The CHIPS and Science Act authorized federal subsidies for domestic semiconductor manufacturing. Foundries including TSMC, Samsung, and Intel announced new U.S. fab projects under the program. Construction at leading-edge nodes remains the binding constraint on added domestic capacity.

How does RAND approach the question?

RAND has produced multiple prior analyses of semiconductor industrial policy. Its work has covered CHIPS Act implementation, advanced packaging capacity, and the resilience of Taiwanese and South Korean fabrication clusters. The organization contracts with U.S. government agencies.

This latest paper enters a crowded field of think tank commentary on AI controls. The Center for Strategic and International Studies and the Information Technology and Innovation Foundation have issued competing analyses over the past year. RAND's contribution reframes an old debate rather than answering it.

Where could the policy go next?

The central RAND question has no clean answer. Export controls can constrain competitors over the short term. Building a global AI ecosystem requires sustained capital spending, allied coordination, and patient supply chain work measured in years rather than quarters.

If RAND's framing takes hold in policy circles, the conversation may shift from how strictly to enforce chip restrictions toward whether the United States is doing enough to build a competing AI stack. The paper supplies a question the next administration will have to resolve.

-- Grace Kim