Chinese Firms Shift AI Accelerator Orders from Nvidia to Local Suppliers after Water Shortages
NVDA•Chinese companies are shifting a growing share of AI accelerator orders from Nvidia’s GPUs to domestic suppliers, driven by US export curbs and intensive water shortages hindering H2O-based cooling. This trend marks a pivot in China’s AI infrastructure buildout, reducing reliance on foreign chips and boosting local accelerator vendors.
1. Shift in AI Accelerator Procurement
Chinese companies are increasingly diverting AI accelerator spending from Nvidia to domestic suppliers in early July 2026. The shift follows growing constraints on GPU imports and rising preference for homegrown accelerators to support local data centers and AI training workloads.
2. Water Shortages Strain Cooling Infrastructure
Worsening water shortages across China have limited availability of H2O-based cooling capacity, critical for high-power Nvidia GPUs. Firms face rising operational challenges, prompting them to seek domestic accelerators with lower cooling demands.
3. Role of US-China Tech Tensions
Intensifying US export curbs on high-end GPUs have accelerated the procurement of locally produced AI chips. Chinese policymakers are promoting domestic semiconductor adoption, reinforcing self-sufficiency in critical AI infrastructure.
4. Implications for Nvidia and Local Vendors
Nvidia stands to lose a measurable share of its Chinese AI accelerator market, potentially denting hardware revenue. Conversely, domestic suppliers gain market share and investment momentum as China builds out its AI computing ecosystem.
