Nvidia Shifts AI Chip Production from China to Focus on US Demand
Nvidia has halted production of its H200 AI chips for the Chinese market, reallocating manufacturing capacity at Taiwan Semiconductor Manufacturing Company (TSMC) to prioritize its newer Vera Rubin architecture. The move reflects tightening US export controls and waning demand in China, where regulatory uncertainty persists.
The Vera Rubin chips, designed for high-performance AI systems, are in strong demand among US tech giants like OpenAI and Google. Meanwhile, proposed US restrictions could cap Chinese firms at 75,000 H200 units each—a fraction of potential demand from the country's tech leaders.
This strategic pivot underscores the growing bifurcation in global AI supply chains, with advanced semiconductor technology increasingly concentrated in Western markets. The ripple effects may extend to cryptocurrency mining hardware, where Nvidia GPUs have historically played a dual role in both AI and blockchain computations.
