Why Chinese AI Models Should Worry Nvidia, Micron Stock Investors - Yahoo Finance

As Chinese AI models continue to make breakthroughs in key technical areas, the global AI landscape is undergoing profound shifts. Recent reports from The Wall Street Journal and The New York Times hi... Read original →

Breakthroughs in Chinese AI models leveraging MoE and MLA architectures are triggering a structural shift across the global compute stack. Technically, reduced FP8 inference demands erode the necessity for NVIDIA’s Blackwell GPUs and HBM4 memory, directly threatening Micron’s pricing power in high-bandwidth solutions. On the compliance front, tighter U.S. export controls on sub-3nm tools could accelerate China’s adoption of chiplet and near-memory computing, paradoxically strengthening its domestic supply chain. Facing performance leads by models like GLM-5.2 in specialized domains, NVIDIA may fast-track its Rubin platform and loosen software stack restrictions, while hyperscalers like Microsoft and Amazon could hybrid-deploy Chinese-optimized models to cut costs. Over the next 18 months, AI infrastructure investment will pivot from hardware-centric to co-designed algorithm-chip paradigms—sustained Chinese efficiency gains could redirect global AI capex and undermine the valuation foundations of U.S. semiconductor leaders.