DeepSeek plans 160,000 Huawei Ascend chips in Inner Mongolia, the largest known cluster
DeepSeek will deploy at least 160,000 of Huawei's Ascend 950DT chips at an Inner Mongolia data center for inference only, per Bloomberg, while Nvidia hardware still handles training. The split marks the real limit of China's Nvidia substitution — and Huawei is unlikely to fill the order for over a year, with domestic memory maker CXMT starting small HBM3E batches but three to five years behind Samsung, SK Hynix, and Micron.
Source: the-decoder.com ↗
For the much heavier training workloads, Deepseek still uses Nvidia hardware.
the-decoder.com
Why this matters
- → Reveals the hard ceiling of China's Nvidia substitution — inference only, training still depends on US chips
- → Huawei can't deliver at scale for over a year; memory shortage is the real bottleneck, not chip design
- → CXMT's HBM3E is 3–5 years behind competitors, exposing how far China lags in supporting domestic AI infrastruc
China's AI chip reality check