Chinese companies that are dedicated to artificial intelligence (AI) They already have at their disposal a wide range of GPUs designed and manufactured in China. Huawei is the best positioned company thanks to its Chips Ascend 910 family, but in the country led by Xi Jinping There are many other specialized companies in the tuning of hardware for which they also have great potential. Metax, Biren Technology, Moore ThreadsINNOSILICON, ZHAOXIN, ILUVATAR COREX, DEGLINAI OR VAST AI TECH ARE SOME OF THE MOST IMPORTANT.
“China is dedicating mass resources to the implementation of emerging companies specialized in the development of GPU. Do not underestimate them.” This warning was probably not ignored. He was aimed at the US government and came from someone who knows well what he says: Jensen Huang. The general director of Nvidia He pronounced these words Last year, during the celebration of Computex, and it is evident that its intention was to prevent US administration about the consequences that They would have the sanctions that seek to stop the technological development of China.
Hardware is no longer the problem; It is CUDA
Li Guojie, a computer scientist from the Chinese Academy of Sciences that is considered an authority in China, has confirmed What we have just seen: GPUs for the most advanced Chinese, such as Huawei Ascend 910 chips that I mentioned above lines are comparable to current NVIDIA solutions in terms of calculation capacity. However, the authentic strength of the company led by Jensen Huang is not its hardware; is its ecosystem CUDA (Compute Unified Device Architecture).
“Deepseek has had an impact on the CUDA ecosystem, but has not completely overcome it because barriers persist”
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, who aspires to an important portion From this market in China, it has Cann (Compute Architecture for Neural Networks), which is its alternative to CUDA, but for the moment CUDA dominates the market.
The most surprising thing is that not only Chinese companies need to part with this software; The entire AI industry wants to end Cuda. This is at least what Pat Gelsinger saidthe former director general of Intel. In December 2023, this executive wet and explained what was the official position of his company in the context of the AI sector. “The whole industry is determined to eliminate market CUDA (…) We see it as a shallow and small pit (…)”, Gelsiter defended in the framework of the event “Ai EveryWhere” held in New York.
According to Li Guojie“China must develop an alternative system to achieve self -sufficiency in AI (…) Deepseek has had an impact on the CUDA ecosystem, but it has not completely overcome it because barriers persist. In the long term we need to establish a set of software tool systems for controllable that exceed Cuda“. This is undoubtedly one of the great challenges that China faces in this area. Perhaps Cann manages to break down gradually and manages to finally impose himself, but while he persists to some extent the development of AI in China will continue to depend on the US.
Image | Moore Threads
More information | SCMP
In Xataka | We can forget an AI without hallucinations for now. The general director of Nvidia explains why