According to a report from South China Morning Post, Wang Tao, Executive Director, Chairman of the ICT Infrastructure Managing Board at Huawei, recently stated at the World Semiconductor Conference and Nanjing International Semiconductor Expo that Huawei’s advanced Ascend 910B AI chip achieves an efficiency of up to 80% compared to NVIDIA’s A100 when training large-scale language models. In terms of specific test performance, it surpasses NVIDIA’s A100 AI GPU by 20%.
Reportedly, leading Chinese companies such as Tencent Holdings and Baidu have also purchased the Ascend 910B chip for conference tasks, indicating substantial market adoption of the chip and encroaching on NVIDIA’s AI chip market share.
Huawei’s self-developed AI chip, the Ascend 910B, adopts the advanced Da Vinci architecture and supports various AI computing tasks such as deep learning and inference. The chip boasts computing capabilities of up to 320 TFLOPS in half-precision (FP16) and 640 TOPS in integer precision (INT8), all while consuming 310W of power.
Reportedly, it’s expected that Chinese tech giants may now be considering a shift towards local AI products, which could pose a challenge to NVIDIA. Currently, China accounts for 17% of NVIDIA’s revenue in the 2024 fiscal year, making the competition in the Chinese market increasingly fierce for NVIDIA.
According to a previous report from Reuters, NVIDIA, in response to the US ban on high-end chips, provided three chips specifically for the Chinese market, with the H20 chip receiving significant attention.
Hence, there are reports of a price reduction for the H20 chip to stimulate demand and address the intense competition from Huawei in the Chinese market. It’s reported that the price of the H20 chip is over 10% lower than Huawei’s Ascend 910B chip. The chip is reportedly to be sold at approximately 100,000 yuan per unit, while Huawei 910B sold at over 120,000 yuan per unit.
Read more
(Photo credit: Huawei)