Huawei chips continue to fill Nvidia’s gap in Chinese AI research centers

Huawei chips began filling voids made by Nvidia in the Chinese AI research centers. Many Chinese customers are now choosing the Ascend processors over the US SoCs. The reason is not efficiency but availability, following the U.S. trade controls.

According to a report, an AI research center under the Chinese Academy of Sciences (CAS) has been using Huawei Ascend AI chips instead of Nvidia. It is the third case where a Chinese sector opted for native components rather than foreign ones.

A Hong Kong-based Centre for Artificial Intelligence and Robotics (CAIR) has recently introduced its new AI model named CARES Copilot 2.0 in the medical world.

It aims to help doctors, mainly surgeons with surgical planning, generating diagnostic reports, and retrieving files of similar cases. The model is an upgraded version of its predecessor (equipped with Huawei Ascend 910B AI processors).

The Hong Kong-based AI research center says that it has upgraded many of its medical AI models with Huawei AI chips. Liu Hongbin – the executive director of the CAIR believes Huawei AI chips can basically satisfy the lab’s research needs.

CAIR and Huawei’s partnership began last year in August. Both parties penned down a memorandum of cooperation to jointly work on imposing AI use in surgery.

Huawei Ascend AI chip (Image Credits: Huawei)

On November 22, Liu Hongbin had a small talk with the media about its new AI model. During the interrogation, he also talked about its bond with Huawei:

“At that time (during cooperation), we felt that in the foreseeable future, due to geopolitical reasons, the US would definitely still restrict us from using high-end chips. As a research institution, we needed to have alternative plans. Since then, the capabilities of Huawei’s AI chips have quickly advanced.”

Liu added that Huawei’s progress in AI chips is visibly fast. He said, earlier, the center used to take one or two months to complete the entire training. Although the same amount of data took only a week to sum up the training with Huawei chips.

The executive mentioned that the previous-gen model that runs on Ascend 910B performance is quite close to Nvidia A100. Meanwhile, the Copilot 2.0 uses Huawei’s Atlas 800T server. CAIR’s growth shows that Huawei chips are constantly filling the gaps made by Nvidia in Chinese AI research centers and other companies.

Although Nvidia still leads when we talk about ecosystem. But in terms of achieving functionality, one can trust Huawei as it can easily meet the requirements, says Liu.

Huawei chips continue to fill Nvidia’s gap in Chinese AI research centers (Image Credits: Huawei)

|| Source ||