Huawei is determined to gradually absorb the market share that keeps Nvidia in China. Until just a few months ago this last company monopolized Something more than 90% of the Chinese market of the chips for artificial intelligence (AI), but after the entry into force of the last US sanctions package its leadership is all likely to be compromised. Even so, Nvidia has a very important asset that is helping her defend her presence in the Chinese market: CUDA (Compute Unified Device Architecture).
Most of the AI projects that are currently being developed are implemented on CUDA. This technology brings together the compiler and development tools used by programmers to develop their software for NVIDIA GPUs, and replace it with another option in the projects that are already underway it is a problem. Huawei has Cann (Compute Architecture for Neural Networks), which is its alternative to CUDA, but for the moment CUDA dominates the Chinese market.
These are the two great buzas of Huawei to beat Nvidia
Huawei wants to snatch the leadership in performance in AI applications from NVIDIA. His most ambitious proposal right now is the chip Ascend 910dwho seeks to overcome the performance of the GPU NVIDIA H100. However, this Chinese company has also recently presented its chip Ascend 920a solution that is clearly destined to occupy in the Chinese market the gaps that it will leave The H20 GPU of Nvidia.
This proposal will enter large -scale production during the second half of 2025 using 6 NM integration technology that have presumably developed elbow with Huawei elbow and SMIC. It is evident that to grow in the market is essential have good hardwarebut it is also crucial to position itself strongly in training the great language models, in Inference processesor, better yet, in both contexts.
“Training is important, but it only happens a few times. Huawei focuses mainly on inference, which will ultimately give us access to more customers.”
This declaration of Georgios Zacharopoulos, a senior researcher of AI who works on the acceleration of inference in the Huawei laboratory in Zurich (Switzerland) clearly reflects The effort that this company has made for years to dominate inference: “Training is important, but only a few times. Huawei focuses mainly on inference, which will ultimately give us access to more customers.”
Inference is broadly the computational process carried out by language models with the purpose of Generate the answers which correspond to the requests they receive. In any case, the information we have reflects that the GPU Ascend 910D will allow Huawei to compete with the chips for the most advanced NVIDIA both in inference and in training.
The US response to the steps that Huawei is not taking long to arrive. And is that the Department of Commerce has approved a resolution whereby no country on the planet can buy the GPUs for the Ascend de Huawei. According to this American institution, this Chinese company has produced these chips using US technologies illegally, so its export outside the country borders governed by Xi Jinping violates the export controls of the Department of Commerce.
In practice to the US it will cost a lot to control the commercial flow of the GPUs for Huawei outside China, especially when these semiconductors They go to allies of the latter country. Its strategy to exert pressure on countries interested in getting the Huawei chips is to announce fines, the possibility of revoking export rights, and even establishing criminal consequences.
In Xataka | In a low voice, China has begun to remove some tariffs from US products. Your concern: the chips
In Xataka | China’s domain of rare earths has nothing to do with geography: it is born from 39 university programs
GIPHY App Key not set. Please check settings