Home · Technology · Sep 30 archive
DeepSeek and Huawei Partner to Develop AI Chip Software
Confirmed
In Short: Chinese tech firms DeepSeek and Huawei have partnered to develop programming tools for Huawei's AI chips, aiming to reduce dependence on Nvidia.
Chinese artificial intelligence company DeepSeek has partnered with Huawei to develop programming tools for Huawei’s Ascend AI chips, as Chinese technology companies accelerate efforts to reduce their dependence on American chipmaker Nvidia.
In a statement, DeepSeek said, “To build a new generation of independent, self-controlled GPU software ecosystems, the first priority is establishing a high-level language that is universal, easy to program, and still capable of reaching the hardware's full performance potential.” The firm added that TileLang meets this need and is simpler to use than Nvidia’s CUDA.
DeepSeek and Huawei have developed a new toolkit based on the open-source GPU programming language TileLang, to allow for rapid development of AI for Huawei chips that is not dependent on Western hardware or software libraries.
The partnership includes programming tools, computing libraries, and communication software intended to make developing AI applications on Chinese-made chips easier.
Together, the two firms have established a new “supernode” solution formed from 128 Huawei Ascend 950 chips and worked to establish new Huawei-optimised programming tools.
DeepSeek revealed that it had worked with Huawei on a supernode solution using 128 Ascend 950 processors.
DeepSeek said its collaboration with Huawei focused on improving both computation and communication within the 128-chip system.
The companies have not published comprehensive independent benchmarks establishing how their jointly developed 128-chip system performs against equivalent Nvidia-based infrastructure.
DeepSeek has committed to open-sourcing the programming tools alongside libraries that cover compute and communication.
This is important for Chinese AI labs because to make Huawei chips work for training runs they need to efficiently network and share data as a cluster.
Huawei dominates the Chinese chip market and is officially backed by the government as the go-to hardware option for Chinese AI.
The initiative comes two weeks after Huawei unveiled its next-generation AI processors and computing systems, positioning them as alternatives to American technology.
Background
These developments underscore the growing competition in the AI sector, with Chinese firms like DeepSeek and Huawei challenging Nvidia's dominance.
What's confirmed
- Chinese artificial intelligence company DeepSeek has partnered with Huawei to develop programming tools for Huawei’s Ascend AI chips, as Chinese technology companies accelerate efforts to reduce their dependence on American chipmaker Nvidia.
- In a statement, DeepSeek said, “To build a new generation of independent, self-controlled GPU software ecosystems, the first priority is establishing a high-level language that is universal, easy to program, and still capable of reaching the hardware's full performance potential.” The firm added that TileLang meets this need and is simpler to use than Nvidia’s CUDA.
- DeepSeek and Huawei have developed a new toolkit based on the open-source GPU programming language TileLang, to allow for rapid development of AI for Huawei chips that is not dependent on Western hardware or software libraries.
- The partnership includes programming tools, computing libraries, and communication software intended to make developing AI applications on Chinese-made chips easier.
- Together, the two firms have established a new “supernode” solution formed from 128 Huawei Ascend 950 chips and worked to establish new Huawei-optimised programming tools.
- DeepSeek revealed that it had worked with Huawei on a supernode solution using 128 Ascend 950 processors.
- DeepSeek said its collaboration with Huawei focused on improving both computation and communication within the 128-chip system.
- The companies have not published comprehensive independent benchmarks establishing how their jointly developed 128-chip system performs against equivalent Nvidia-based infrastructure.
- DeepSeek has committed to open-sourcing the programming tools alongside libraries that cover compute and communication.
- This is important for Chinese AI labs because to make Huawei chips work for training runs they need to efficiently network and share data as a cluster.
- Huawei dominates the Chinese chip market and is officially backed by the government as the go-to hardware option for Chinese AI.
- The initiative comes two weeks after Huawei unveiled its next-generation AI processors and computing systems, positioning them as alternatives to American technology.
What's still developing
- As one of the most prominent Chinese AI developers, DeepSeek has expressed a public desire to reduce its dependence on Nvidia and in July Reuters reported it could develop its own chips to achieve this aim.
- On 17 September, Huawei announced a new roadmap for its upcoming flagship chips the Ascend 960DT and Ascend 960PR.
- In addition to the leap in raw performance it is promising through the chips, Huawei is set to launch its new HiF4 data format and Peerium Computing Architecture with the Ascend 960 family.
- It has benefited from strict export controls on Nvidia chips entering China, which Beijing has only recently eased.
- Developers use programming languages to tell computer chips which calculations to perform and how to manage the information required for those calculations.
- FCRF Launches CP-FRM to Build India’s Next Generation of Fraud Risk Professionals Nvidia’s dominance in artificial intelligence extends beyond manufacturing powerful processors.
- Its CUDA software platform has become an important part of how developers build, train and run AI models.
- CUDA allows programmers to write software that uses Nvidia’s graphics processing units, or GPUs, to perform large numbers of calculations simultaneously.
- Over time, developers have built applications and research tools around this software ecosystem.
- That creates an additional challenge for rival chipmakers.
- Even if another company develops a capable AI processor, developers may find it difficult or expensive to move existing applications away from Nvidia.