Huawei and DeepSeek Take Aim at Nvidia with New Programming Tools

DeepSeek and Huawei are working on tools for Huawei's Ascend AI chips, which may pose a threat to Nvidia's dominance in the field. According to DeepSeek's WeChat account, the company is developing programming tools for Huawei's Ascend CPUs. This endeavor is significant because it shows that Nvidia's competitiveness goes beyond its GPU performance.

CUDA gives developers a framework for creating Nvidia-compatible apps, but DeepSeek and Huawei are developing an alternative with TileLang, an AI chip-specific high-level open-source programming language. TileLang offers a simpler programming model than CUDA, which may boost development efficiency and simplify code.

The collaboration between DeepSeek and Huawei is part of a larger effort to develop an AI computer environment independent of Nvidia. This is evident in Huawei's recent announcement of its next-generation AI CPUs and supernode computing systems, which are expected to be widely used for model training next year.