-
Nvidia is lauded for its AI chips but said its GPUs can be used for 5G acceleration too
-
NTT DoCoMo is leading the way with its Nvidia-based 5G vRAN deployment
-
Softbank will use Nvidia to deliver vRAN and AI on the same framework
While Nvidia may be known as the artificial intelligence (AI) chip leader – with its H100 chips – the firm wants to remind people that its Graphics Processing Units (GPUs) can be used for other tasks such as Layer 1 5G acceleration too.
The company told us that Nvidia Aerial 5G is now live in the DoCoMo network in Japan. "This is the first deployment of a GPU-accelerated vRAN," Ronnie Vasishta, Nvidia’s SVP of Telecom, said. The company noted that the virtualized radio access network (vRAN) deployment is also open RAN compliant.
Aerial, by the way, is an application framework for building software-defined, cloud-native 5G networks, accelerated by Nvidia GPUs.
Softbank also recently announced at Computex 2023 they will be using the Nvidia Grace Hopper "superchip" in a MGX server chassis with Nvidia Aerial to deliver vRAN and AI on common infrastructure.
Stefan Pongratz, analyst and VP at Dell'Oro Group, noted that Nvidia, along with Fujitsu, is powering one of the first deployments of 5G GPU-based acceleration with the DoCoMo rollout.
"We should start seeing more meaningful accelerator revenues over the next year or two," Pongratz told Silverlinings.
But as Nvidia said, it is at the forefront of the acceleration movement with its GPUs.
Want to discuss AI workloads, automation and data center physical infrastructure challenges with us? Meet us in Sonoma, Calif., from Dec. 6-7 for our Cloud Executive Summit. You won't be sorry.