Chinese AI startup DeepSeek is planning to deploy at least 160,000 Huawei Ascend 950DT AI accelerators at a massive new data center in Inner Mongolia.
If completed, the project could become one of the largest known clusters of Huawei AI chips and represents a major step in China’s effort to build powerful domestic AI infrastructure while reducing its dependence on Nvidia hardware.
DeepSeek reportedly plans to use the Huawei processors mainly for inference, allowing the company to run and serve its AI models at massive scale. For the more demanding task of training advanced AI models, DeepSeek continues to rely largely on Nvidia accelerators.
The project could also provide a major boost to Huawei’s growing AI chip business. However, shortages of advanced memory and other components may limit Huawei’s production capacity and could delay delivery of the full 160,000-chip deployment.
DeepSeek’s huge infrastructure investment highlights how rapidly China’s AI industry is expanding its computing capacity as domestic technology companies compete with OpenAI, Google, Anthropic and other global AI leaders.
