Chinese AI company DeepSeek has teamed up with Huawei Technologies to release an open-source suite of programming tools for Huawei’s Ascend chips, a move aimed at loosening Nvidia’s grip on AI software infrastructure.
The announcement, made Wednesday via DeepSeek’s official WeChat account, includes TileLang, a high-level programming language designed to simplify development for AI chips. DeepSeek said TileLang offers “a simpler programming model” than Nvidia’s CUDA while still allowing developers to access the hardware’s full performance potential.
Alongside TileLang, DeepSeek released underlying libraries built for high-throughput training and inference on Ascend chips, including DeepGEMM Ascend for matrix multiplication kernels and DeepEP Ascend for large-scale communication across devices. The company also said the two firms jointly advanced a “supernode” solution based on 128 Ascend 950 chips, optimizing both computation and communication.
The release builds on Huawei’s recent hardware push. It came about two weeks after Huawei unveiled its next-generation AI processors and supernode computing systems, and one week after Huawei said it would ship its next-generation Ascend 960DT chip in Q1 2027, three quarters earlier than previously planned. Huawei Rotating Chairman Eric Xu claimed Ascend chips already hold a larger share of China’s AI chip market than Nvidia, though no figures were given.
DeepSeek has been moving toward domestic hardware for months. Its V4-Flash AI model was optimized specifically for Huawei chips, and the company reportedly plans to install more than 160,000 Huawei Ascend chips at a new data center in Inner Mongolia. The partnership illustrates how Chinese tech firms are adapting to US export controls that have restricted access to Nvidia’s most advanced processors.
Nvidia’s CUDA platform has powered AI development for more than a decade, and no Chinese rival has broken that hold. The key question remains how much the new tools can narrow the performance gap that keeps developers tied to Nvidia’s ecosystem.
