Chinese artificial intelligence company DeepSeek has partnered with Huawei Technologies to develop programming tools for Huawei’s Ascend AI chips, strengthening efforts to build a domestic alternative to Nvidia’s hardware and software ecosystem.
DeepSeek said it is open-sourcing programming infrastructure for Huawei’s Ascend platform, including computing and communication libraries designed to help developers build and optimise AI applications on the Chinese-made processors.
The collaboration also includes a high-level programming language called TileLang. DeepSeek said the tool is designed to make AI-chip programming more efficient and simplify code development while still allowing developers to access the performance capabilities of the underlying hardware.
DeepSeek described TileLang as offering a simpler programming model than Nvidia’s CUDA, the software platform that has become central to developing applications for Nvidia’s graphics processors.
The partnership goes beyond software. DeepSeek and Huawei have jointly advanced a supernode system built around 128 Ascend 950 chips, with the companies working to improve both computing performance and communication between the processors.
This is significant because large AI models require enormous amounts of computing power. Connecting multiple processors allows AI systems to distribute workloads across many chips, but efficient communication between those chips is essential to avoid bottlenecks.
DeepSeek’s announcement comes about two weeks after Huawei unveiled its next-generation AI processors and supernode computing systems. Huawei has said it expects its AI computing systems to see wider use for model training in the coming year.
The software push reflects a broader challenge facing Chinese AI companies. Building competitive AI hardware is only part of the equation; developers also need programming tools, libraries and software ecosystems that make those chips practical for large-scale AI development.
Nvidia has a major advantage in this area because its CUDA ecosystem is widely used by AI developers. DeepSeek and Huawei are therefore attempting to address the software side of the challenge by making more tools available for Ascend processors.
DeepSeek has also been moving toward greater use of Chinese-made chips for AI development. Earlier reports indicated that the company planned to increase its reliance on Huawei processors for training future AI models as domestic chip availability improves.
The partnership comes amid continuing US restrictions on China’s access to advanced computing chips and semiconductor technology. Those restrictions have encouraged Chinese technology companies to develop domestic alternatives across both hardware and software.
The success of the effort will depend not only on the capabilities of Huawei’s processors but also on whether developers adopt the new software tools and whether they can achieve competitive performance at large scale.
For now, the DeepSeek-Huawei collaboration represents another step toward a more self-contained Chinese AI computing ecosystem, with software development becoming as important as the chips themselves.
