Chinese artificial intelligence company DeepSeek said on Wednesday it has teamed up with Huawei Technologies to develop programming tools for Huawei's Ascend chips, in the latest sign of Chinese tech firms deepening their ties as they seek alternatives to Nvidia's ecosystem.
DeepSeek announced the partnership in a post on its official WeChat account. The company said it is open-sourcing infrastructure for Huawei's Ascend platform, including a high-level programming language called TileLang along with related computing and communication libraries.
Aiming at CUDA
The move is aimed at the software layer that has long cemented Nvidia's dominance in AI computing. CUDA, Nvidia's programming platform, is widely seen as one of the hardest advantages for rivals to dislodge because developers have built an enormous library of code around it.
DeepSeek said the first priority in building a new generation of "independent, self-controlled" GPU software ecosystem is establishing a high-level language that is universal, easy to program, and still able to reach the hardware's full performance potential. "TileLang was created precisely to meet this need," the company said, adding that it offers a simpler programming model than Nvidia's CUDA, improving development efficiency and simplifying code logic.
Bigger push for independence
The partnership fits into a broader effort by Chinese technology companies to reduce reliance on foreign chip technology amid tightening US export controls. By jointly developing and open-sourcing the software stack around Huawei's Ascend processors, DeepSeek and Huawei are betting that developers will gravitate toward an open ecosystem rather than remaining locked into a single vendor's platform.
Neither company disclosed a timeline for the tools' release or details of the commercial terms of the partnership.
Sources:
- Reuters, "DeepSeek partners with Huawei to develop chip programming tools, reducing reliance on Nvidia", read here
DeepSeek announced the partnership in a post on its official WeChat account. The company said it is open-sourcing infrastructure for Huawei's Ascend platform, including a high-level programming language called TileLang along with related computing and communication libraries.
Aiming at CUDA
The move is aimed at the software layer that has long cemented Nvidia's dominance in AI computing. CUDA, Nvidia's programming platform, is widely seen as one of the hardest advantages for rivals to dislodge because developers have built an enormous library of code around it.
DeepSeek said the first priority in building a new generation of "independent, self-controlled" GPU software ecosystem is establishing a high-level language that is universal, easy to program, and still able to reach the hardware's full performance potential. "TileLang was created precisely to meet this need," the company said, adding that it offers a simpler programming model than Nvidia's CUDA, improving development efficiency and simplifying code logic.
Bigger push for independence
The partnership fits into a broader effort by Chinese technology companies to reduce reliance on foreign chip technology amid tightening US export controls. By jointly developing and open-sourcing the software stack around Huawei's Ascend processors, DeepSeek and Huawei are betting that developers will gravitate toward an open ecosystem rather than remaining locked into a single vendor's platform.
Neither company disclosed a timeline for the tools' release or details of the commercial terms of the partnership.
Sources:
- Reuters, "DeepSeek partners with Huawei to develop chip programming tools, reducing reliance on Nvidia", read here
