U.S. Tightens Controls on Nvidia AI Chip Exports - Let's Data Science

On May 31, 2026, Reuters reported that the U.S. Department of Commerce issued new guidance clarifying that export license requirements for advanced AI semiconductors apply to entities headquartered in... Read original →

The U.S. Commerce Department’s May 2026 guidance shifts export controls from geographic to corporate nationality, effectively closing the offshore subsidiary loophole Chinese firms exploited to access NVIDIA’s Blackwell/Rubin and AMD’s MI350x chips. Technologically, this throttles China’s AI training throughput, forcing reliance on subpar alternatives and accelerating domestic Chiplet and advanced packaging R&D. Compliance burdens surge: every transaction involving Chinese-headquartered entities—even overseas—now requires licensing, extending delivery lead times by over 30%. Strategically, Huawei Ascend and Cambricon will capture vacated domestic AI training demand, while Samsung and SK Hynix may expand compliant foundry roles to hedge U.S. policy risk. Over the next 12–24 months, the global AI chip supply chain will bifurcate into three tracks: a U.S.-led closed ecosystem, a China-centric self-reliant bloc, and an ambiguous intermediary zone. This move isn’t escalation—it’s institutionalization of semiconductor decoupling.