As a way of introduction, Ronnie Vasishta, Nvidia's senior vice president of telecom, said enterprise edge computing was not a segment the GPU powerhouse had really engaged with, at least "until now."
Vasishta was speaking in a press briefing towards the tail end of last week, under embargo until Mobile World Congress swung back into action today after a lengthy pandemic-induced hiatus.
Building on an announcement made by Nvidia in April that it was teaming up with Fujitsu, Google Cloud, Mavenir, Radisys and Wind River to develop solutions for its AI-on-5G platform, Vasishta said its first "open" AI-on-5G lab, in collaboration with Google, was in the pipeline.
Vasishta said that when the lab might open and where it will be located would all be revealed later this year. "Industrial companies, systems integrators and network operators will be able to develop and test their AI-on-5G enterprise apps running on Google Anthos using Nvidia's Aerial [5G] infrastructure," he said.
In parallel with the Google lab announcement, Nvidia said it was extending support for Arm‑based CPUs in the Nvidia Aerial A100 AI-on-5G platform.
According to Nvidia, including CPU designs from the company it wants to buy for $40 billion will help businesses "easily deploy intelligent services at the edge by enabling the world's leading OEMs to offer industry-standard servers running highly efficient Arm-based CPUs and Nvidia AI enterprise software with Aerial 5G."
In last week's press briefing. Vasishta was at pains to stress that the Nvidia AI enterprise platform was cloud-native, general purpose, software defined and could be orchestrated to run mixed workloads, whether they were VMs or containers using Kubernetes on bare metal. "This makes the platform wholly consistent with advantages of a commercial, off-the-shelf enterprise solution," he claimed. "We have brought the power of the AI cloud [to] the 5G-connected enterprise edge," continued Vasishta. "This not only brings tremendous, untapped monetization opportunities for operators, which have spent billions of dollars acquiring spectrum, but also to the enterprise ecosystem."
- UK scrutiny chips away at Nvidia's $40B Arm buy
- Fujitsu, Google Cloud, Mavenir, others team with Nvidia
- Nvidia posts bumper Q3, but there may be trouble ahead
— Ken Wieland, contributing editor, special to Light Reading