Washington’s decision to ease restrictions on Nvidia’s H200 exports to China signals more than a routine adjustment in trade policy. It reflects a growing divide over how the United States should compete in artificial intelligence.
On one side, Nvidia CEO Jensen Huang has increasingly argued that diffusion—how broadly AI is adopted across the economy—will ultimately determine leadership. On the other, Chip War author Chris Miller maintains that computing power remains one of the few strategic choke points the U.S. still controls.
This article examines how these competing visions collided ahead of the H200 easing, why diffusion theory is gaining traction in Washington, and what China’s slower pace of AI adoption suggests about the real effects of U.S. containment.
Far from a simple race for better models, the outcome of the AI competition may hinge on who shapes the conditions under which AI spreads.
cwnewsroom.substack.com
On one side, Nvidia CEO Jensen Huang has increasingly argued that diffusion—how broadly AI is adopted across the economy—will ultimately determine leadership. On the other, Chip War author Chris Miller maintains that computing power remains one of the few strategic choke points the U.S. still controls.
This article examines how these competing visions collided ahead of the H200 easing, why diffusion theory is gaining traction in Washington, and what China’s slower pace of AI adoption suggests about the real effects of U.S. containment.
Far from a simple race for better models, the outcome of the AI competition may hinge on who shapes the conditions under which AI spreads.
Jensen Huang Backs “Diffusion”—Chris Miller Pushes Back on the H200 Easing
This week, Washington quietly crossed a line that could reshape the global AI race. By allowing Nvidia to sell its H200 GPUs to approved Chinese companies, the Trump administration reignited a long-running debate over whether the U.S. should prioritize containing China—or focus instead on...
cwnewsroom.substack.com