The Five Nine: Do GPUs make sense for AI-RAN?

Nvidia made headlines in October when it announced plans to invest $1 billion in telecom vendor Nokia to develop AI-RAN solutions that use its GPUs. But do GPUs even make sense for AI-RAN?

Nvidia’s announcement with Nokia mostly focused on using GPUs at the base of wireless towers to run both radio access network (or, RAN) applications alongside AI workloads. But while Nvidia seems eager to tap into this new market, there are a few problems.

First, there are a ton of questions about whether the use case Nvidia is targeting even requires GPUs. And second? AI-RAN isn’t just one thing. And plenty of AI-RAN workloads are running today on existing hardware – that is, on CPUs.

It’s a bit complicated, but with a player as big as Nvidia shining a spotlight on the topic, it’s worth taking a deep dive to really understand what’s happening in this space.

With help from Orange’s Director of Innovation in Radio & Environment Atoosa Hatefi and Dell’Oro Group’s Stefan Pongratz we’ll dig into the different flavors of AI-RAN, what would compel telecom operators to spring for GPUs and what the overall market forecast looks like. 

Catch the video at top, listen to the audio edition below on the go or watch this and future episodes on YouTube!

This podcast is written and hosted by Diana Goovaerts. It is edited by Diana Goovaerts and Matt Rickman. Liz Coyne is our executive producer. Special thanks to our guests Atoosa Hatefi and Stefan Pongratz.

To learn more about the topics in this episode:

Check out our full archive of episodes of The Five Nine right here