“`html
CNBC AI News, July 20th – Zhou Hongyi, founder and chairman of 360 Group, recently released a video interpreting Nvidia’s return to China with the H20 chip as a sign of ongoing AI competition between the U.S. and China.
From an industry perspective, Huang’s statements convey several key signals. Firstly, the global AI chip market is being reshaped. While NVIDIA remains a leader, the pace at which Chinese companies like Huawei are catching up is remarkable.
“Huawei has only been developing GPUs for a few years, while NVIDIA has been at it for 30 years,” Zhou Hongyi noted. “Although Huawei hasn’t surpassed NVIDIA yet, the speed at which they are advancing is noteworthy.”
On July 16th, NVIDIA CEO Jensen Huang, when asked about the potential for Huawei’s AI chips to replace NVIDIA in AI training, responded that it was only a matter of time.
“The point is, we’ve been doing this for 30 years, and they’ve been doing it for just a few. And the fact that they can tell you how powerful they are, says something.”
“No one has put more energy into building the AI ecosystem than I have. I’ve been doing it at a very high level, at a scale that’s difficult to execute. And the fact that Huawei can be compared to us already, speaks volumes,” Huang stated.
Huang also expressed hope that more advanced chips would find their way into China due to continuous technological advancements. He emphasized that the Hopper architecture, used in the H20, is already excellent. He further added that NVIDIA will continue to develop even better technologies over the coming years, regardless of specific product restrictions in China, the products available will improve with time.
Notably, in June, Huawei founder Ren Zhengfei told the People’s Daily that concerns about chips were unwarranted. “By using methods like stacking and clustering, the computational results become comparable to the most advanced levels.”
He acknowledged that Huawei’s single chips still lagged behind the U.S. by a generation, but added, “We use mathematics to compensate for physics, non-Moore to compensate for Moore, and cluster computing to compensate for single chips. In practice, the results are still sufficient.”
Ren believes that China has opportunities in the mid-to-low-end chip market, with dozens or even hundreds of Chinese chip companies working diligently. He also emphasized the greater potential in compound semiconductors. Specifically, regarding silicon-based chips, “We use mathematics to supplement physics, non-Moore to supplement Moore, and utilize the principles of cluster computing to meet our current needs.”
He pointed out that software is not a bottleneck, being merely mathematical, graphical symbols, and code built from advanced operators and algorithms without restrictions.
“`
Original article, Author: Tobias. If you wish to reprint this article, please indicate the source:https://aicnbc.com/5182.html
