Experts tell Congress hardware, chips and export controls are central to U.S. AI lead

Witnesses repeatedly told members that U.S. leadership in artificial intelligence depends not only on models and talent but on hardware, energy and secure supply chains.

Samuel Hammond said the U.S. lead is “downstream of our massive advantages in AI hardware and data centers” but warned that those advantages are tenuous. Hammond testified that China added “over 400 gigawatts to their grid last year” and argued that without controls on hardware exports the U.S. could be leapfrogged.

Hammond and other witnesses urged stronger export controls on advanced AI chips and described legislative proposals and enforcement tools. During questioning, Hammond referenced a Senate proposal he described as the GAIN AI Act, which he said would give U.S. companies a right of first refusal for certain chip sales. The witnesses and members also discussed reports of third‑party routing and smuggling of advanced chips through intermediaries in Malaysia, Singapore and Taiwan.

Committee members raised Taiwan’s central role in chip production and asked witnesses about the proportion of advanced AI logic chips produced there. Hammond testified that for the most advanced AI logic chips, “it’s over 90%,” and noted U.S. plans to expand manufacturing capacity, including new foundries in Arizona.

Members and witnesses also discussed the national security consequences of broad access to advanced chips. No new sanctions or export rules were enacted at the hearing; members signaled interest in follow‑up oversight and possible legislative responses.

The subcommittee received no classified briefings during this session and did not vote on export measures. Witnesses recommended continued export controls, monitoring of supply chains and investment in domestic fabrication and energy infrastructure to keep AI critical infrastructure onshore.