Deepseek plans to deploy at least 160,000 of Huawei's next-generation Ascend-950DT chips in a massive data center in Inner Mongolia. The chips would only run inference, not training, according to Bloomberg. For the much heavier training workloads, Deepseek still uses Nvidia hardware. If built, this would be the largest known Huawei chip cluster and a real step toward weaning China off Nvidia.

Huawei probably can't deliver the full order for over a year, though, thanks to production limits and memory chip shortages. The memory bottleneck could loosen up. China's top memory maker, CXMT, is turning out small batches of HBM3E for the first time, the high-speed memory that powers many AI processors. But CXMT is still three to five years behind Samsung, SK Hynix, and Micron, all of which are already mass-producing HBM4. Deepseek's order is part of a bigger Chinese government push to grow its own chip industry without falling behind in AI.