Huawei Technologies will be installing a Cloud and AI Business Group in Korea this year in a bid to penetrate the Graphic Processing Unit (GPU) server market currently dominated by NVidia.
The business group may later on be spun off for further growth, according to industry sources on April 3.
Huawei is currently recruiting experts for this new organization, including former and current employees and executives from NVidia.
In August 2019, Huawei unveiled its Ascend 910, calling it the world’s most high-quality AI processor. Ascend 910 recorded a speed of 256TeraFLOPS. This is about twice the speed of NVidia’s GPU Tesla V100. Ascend 910 is made via the 7nm+ technology, while Tesla V100 is with the 12nm.
A month after revealing Ascend 910, Huawei announced the Atlas 900, saying it was the world’s fastest AI training cluster. The key purpose is to study weather forecast, autonomous driving, oil exploration and space observation, among others.
In January 2020, Huawei announced the Kunpeng 920, which is a server CPU aimed at big data calculations and dispersed storage. It was based on the ARMv8 architecture, and formed of the 2.6GHz 64 Core. Power efficiency was improved 30% in this model, according to Huawei. On the same day, Huawei released the Taishan server series, which is run by Kunpeng 920.
In 2019, Huawei created an umbrella Cloud & AI Business Group at its headquarters. Carriers, BG, Enterprise BG, Network& Solution, along with ICT businesses all belong to this group.
Huawei Korea is formed of three different divisions: Carriers, Enterprise and Consumer. The new Cloud & AI BG in Korea will be placed under the Enterprise division.
The Elec is South Korea’s No.1 tech news platform.