Skip to main content

Nvidia hedges its 5G bets between Intel and Arm

AI
(Image credit: Pixabay)

Nvidia is adding support for Arm CPUs to its Nvidia Aerial A100 AI-on-5G platform, which is designed to help businesses deploy intelligent services at the edge.

At MWC 2021, Nvidia announced the Arm version of the Nvidia Aerial A100 platform will combine 16 Arm Cortex A78 processors with the upcoming Nvidia BlueField-3 data processing unit (DPU).

Notably, during a conversation with SoftBank CEO Masayoshi Son last year, Nvidia CEO Jensen Huang pointed out that the Arm acquisition (if approved) will allow Nvidia to bring its AI expertise to “the most popular edge CPU in the world".

Nvidia argues that the new self-contained, converged card will enable businesses to run edge artificial intelligence (AI) workloads over 5G virtual radio area networks (vRAN).

“We’re bringing together two worlds — AI of computing and 5G of telecommunications — to create a software-defined platform for AI on 5G. Now supporting Arm, our growing Aerial platform is accelerating AI-on-5G everywhere,” said Ronnie Vasishta, SVP of Telecom at Nvidia.

AI at the edge over 5G

Nvidia explains that its AI-on-5G platform gives enterprises an option for deploying intelligent services at the edge. Commonly cited use cases include precision robots, automated guided vehicles and such. 

The MWC announcement is about bringing more choice to users by adding the Arm-powered option in the Aerial A100 AI-on-5G platform, which already supports x86-based Intel CPUs.

Nvidia’s new BlueField-3 A100 DPUs, expected to go on sale in the first half of 2022, are optimized for 5G connectivity and developed especially for AI workloads, and include Nvidia’s AI software library and Aerial 5G SDK. 

“BlueField-3 expands the Arm ecosystem by offering a combination of Nvidia accelerated computing and Arm performance to enable network providers more choice in how they create and deploy 5G systems,” notes Chris Bergey, Infrastructure SVP and GM at Arm.

By offering the ability to deploy BlueField-3 A100 with either x86 or Arm-based CPUs, Nvidia argues that both network operators and businesses will be able to roll out software-defined 5G base stations and AI applications that can be easily upgraded in terms of features and performance.

Mayank Sharma

With almost two decades of writing and reporting on Linux, Mayank Sharma would like everyone to think he’s TechRadar Pro’s expert on the topic. Of course, he’s just as interested in other computing topics, particularly cybersecurity, cloud, containers, and coding.