The Trump administration’s lifting of restrictions on H200-class AI chip exports may look like a turning point in the debate over export controls. But the development carries a deeper meaning, one that exposes the true nature of the 21st-century race between the U.S. and China to dominate in AI.
Jason Hsu, a senior fellow at the Hudson Institute, explains why in a Washington Times op-ed. The administration’s decision to allow exports of the powerful Nvidia H200 chips to China indicates that the AI race has moved into a new phase, one that doesn’t hinge on a single chip. Mr. Hsu says restricting H200 exports would only delay China’s progress on AI, not prevent it.
“At this stage, China’s binding constraints are no longer primarily about access to a particular chip. They are about building and operating the systems that make AI usable at scale: reliable electricity, data centers, cooling, networks and the organizational capacity to integrate AI into real operations,” he writes. “Hardware alone is no longer the only source of leverage. Chinese open-source and open-weight AI models are becoming more capable and more widely available, including to U.S. developers, creating new dependencies that chip controls cannot address.”
