Verizon's Yago Tenorio on Designing the Future of 6G

Full Transcript

Don Kellogg 0m10s

Hello, and welcome to the three hundred and fifteenth episode of The Week with Roger, conversation between analysts about all things telecom, media, and technology by Recon Analytics. I'm Don Kellogg, and

Roger Entner 0m19s

with me as always is Roger Entner. How are doing, Roger?

Guest Speaker 0m22s

I'm great.

Roger Entner 0m24s

So, Roger, this week, we have a great guest to talk with us about six g. Iago Tenorio is SVP and CTO at Verizon. Iago, welcome to the podcast.

Guest Speaker 0m33s

Hey. Thank you for having me.

Guest Speaker 0m35s

Hey. Oh, Iago, what a pleasure to meet you here on the on the podcast. You've joined Verizon two years ago and are are making quite the impact. You have made an announcement around the six gs Innovation Forum. Do you want to tell us a little bit about it?

Guest Speaker 0m52s

Yes. Yes. Time flies. Two years. Yes.

Guest Speaker 0m55s

In October, it'll be two years. So we launched the six gs Innovation Forum about a year ago with six founding members, including ourselves. Now we just welcome additional nine members to the forums. So we're 15 strong. And we set it up to look at use cases that we think that six gs will be enabling so that we can design the system bottoms up from there.

Guest Speaker 1m20s

So that's what we're doing.

Guest Speaker 1m21s

So five gs gets derided for not having lived up to expectations. I think every generation gives us a use case that we didn't expect, and it's a bigger success than we thought. But FWA is the killer one for five gs that nobody expected, and it's transforming the entire telecom industry. But what lessons have you learned from five gs as you're approaching six gs?

Guest Speaker 1m49s

That's a very good question. And by the way, there's still like three, four years, you know, that five gs has in it. And I think, you know, what we may see in the next three, four years, it'll be five gs doing the beginning of what six gs will be natively doing. So I think we need to still watch what's left. But you still ask a very good question, what have we learned?

Guest Speaker 2m13s

And I think a couple of things. The first one is reflecting on what I just said on the previous question. I think six gs could be the first generation that this industry takes the opportunity to define with the use cases in mind. So think first, what are we going to be doing with it? And then designing the technology, as opposed to just designing the technology based on speed and latency, performance, capacity, and then hoping that someone would come and use it for something.

Guest Speaker 2m45s

That's what we want to change, first of all. Now, more maybe on the technical side, five gs came with some rupture. Remember there were a lot of options that were suggested, like one, two, three, up to seven. There weren't too many, but the industry decided to actually implement too few. As a result, there was a gap between five year NSA and five year SA.

Guest Speaker 3m11s

When you move to five gs SA, it's difficult to aggregate capacity that is existing on four gs and five gs NSA. I think the industry needs to learn from that, and there is a lot of discussion and debate these days about the options. But I think it's hopefully this time we get it right. There are much fewer options. So I'm hoping that we make a good decision and there is no gap this time around, and it's a smooth, smooth way forward.

Guest Speaker 3m39s

And we're starting hopefully early enough to make these changes and adjustments. And I think, as you pointed rightly out, these are not like clear, bright lines between one g and another. Five gs and six gs will merge into each other and evolve from each other. But when I look at the six gs Innovation Forum membership list, it looks like the same subjects, you know, the same suspects all over again. So is that what it is?

Guest Speaker 4m9s

How are we going to do the innovative things that Silicon Valley is clamoring for and tells us that the telecom industry is never delivering?

Guest Speaker 4m20s

Yeah. So you need the usual suspects, as you rightly call them, as I think they are the ones who are going to build the system. But you also need other kind of players, like from silicon players, hyperscalers, security pioneers, test emulation players. Who's going to help you beyond the natural ecosystem, the usual kind of suspect ecosystem, to figure out the use cases? And those are part of the additions to the six gs Innovation Forum that we announced last week.

Guest Speaker 4m54s

So for instance, we went with Samsung, with Qualcomm as well. We put together some trials on Isaac, one of them in Dallas during an international soccer event that happened in summer that Spain happened to win. Congratulations.

Guest Speaker 5m13s

Well deserved.

Guest Speaker 5m14s

Thank you. Thank you. Thank you.

Guest Speaker 5m16s

Well deserved. They were by far the best team. They did really, really well.

Guest Speaker 5m21s

I can only agree with you, but I'm Spanish. So, in Dallas, in the semifinals, we set up a trial. I think it was very interesting because was not limited to drone detection. This time it was about crowd detection on the Fumfest area. So we were able to report in real time where there was a certain density of fans moving across the area.

Guest Speaker 5m48s

Actually, the organizers and the companies that were working on the area, they were extremely interested, which kind of proves the point because we are trying to set up things and stretching the technology to do things that we believe that we can monetise. And that's the whole point. We also did some drone detection to test the bandwidth, and we did a few more things. So we did a prototype working on some wearables, the usual glasses from Meta, Raven. So we deployed some GPUs at the edge of our network, we trained an LLM in real time with the scores and actually all the news that were coming from a game, each game.

Guest Speaker 6m36s

Kind of that LLM was trained in real time, but also had all the history of whatever has happened in the history of soccer. And you can talk to your glasses, and because the GPU was at the very edge, the conversation was extremely fast and extremely responsive. You could enjoy the game whatever you wear and engage with your glasses in possibilities, probabilities, in this player, that player, or arguing about anything. The LLM would know everything about the game. It took me a while to train the system to answer the question who's going to win the World Cup, but eventually

Guest Speaker 7m13s

Eventually, it got it right.

Guest Speaker 7m14s

It got it right. Yes, at the end, it got it right. I think those are examples on how we are researching use cases. We are only in 2026, so there's still some way to go. But I think the future use cases that 6E is going to solve for are going in this direction.

Guest Speaker 7m34s

So with ISAC and symmetrical AI uplinks demanding so much bandwidth, what needs to happen on the spectrum policy? Because it looks like to me we need wide contiguous mid band blocks for six gs actually to work as well as it should be in the real world.

Guest Speaker 7m51s

Yeah, you are absolutely right. Just take Isaac as an example. The accuracy, the precision of the positioning, the resolution of the network acting as a radar is proportional to the bandwidth. It's not proportional to the frequency, it's proportional to the actual width of the channel. So that's an important one because at the end of the day, if six gs is standardised as a maximum 400 megabytes channel, that's what you're aiming for.

Guest Speaker 8m19s

But it's extremely challenging to find spectrum real estate that's going to be able to host 400 megabytes channel per operator in The US. So that is challenging. Of course, there is a grade for degradation, so if it's not 400 megabytes, you can always down select the 200 megabytes. But at the end of the day, there are use cases like Isaac, you know, where resolution and precision depends on this.

Guest Speaker 8m47s

Yep. So, you know, with AI really changing the way we use networks, both fixed and mobile, does that require different I'm kinda leading the witness here, does it require a different network architecture? And how do you avoid being just a transport layer? Because I think that's the big fear, that the carriers are getting disintermediated. How do you play that role?

Guest Speaker 9m13s

And you have like a big relationship with Google that you announced recently.

Guest Speaker 9m17s

Yeah, yeah, we do. Yes, it's a really good point. Think to your first part of the question, yes. So imagine AI wearables, for instance. Imagine the way you use AI today, which is prompting.

Guest Speaker 9m32s

It kind of evolves onto a situational context aware AI that is watching everything you see and listening to everything you hear, for instance. It can be video, can be audio, or it can be based on sensors. But at the end, there'll be an agent living on an infrastructure that's not with you and is absorbing and ingesting information in real time. That immediately changes the downlink to uplink radio from a current nine:one, eight:two we are used to, to something that is probably more fiftyfifty or even, you know, with the uplink having the majority of the load. The network is not designed for that.

Guest Speaker 10m10s

The standard is not designed for that. The architecture of the network needs to change. Again, if you're thinking on an AI wearable, you have to think on something that is extremely small with very small ability to dissipate heat or to generate power or to cause a battery. So that also points in the direction of an architecture that favors the device, and it makes the device job much easier. So that's not the one that we have today.

Guest Speaker 10m40s

Something will have to change. It's an excellent example on how understanding the use case may actually help, you know, designing the system.

Guest Speaker 10m50s

For example, you're working very closely with companies like Google. Yeah. You provide services for each other, right?

Guest Speaker 10m57s

Yes, yes, yes. No. Yeah, yeah. And you were going direction monetization, and you're absolutely right. And again, we can continue with the same example.

Guest Speaker 11m4s

So you have a device that is small and is wearable and is nearly invisible and weightless, but that device will not do the inferencing for you. So is the operator just a pipe taking that data closer to the inferencing point? Even if it's just that, I think the operator will have a role in piping the data centers. And as latency becomes more and more important and response time becomes more important, bringing and hosting part of that inferencing infrastructure inside the network. But even if it's not that, you can think of ways in which the network becomes the token provider in a way.

Guest Speaker 11m48s

So I think that kind of distribution of intelligence is part of the answer to your question. The other part are new capabilities like ISAC, again, where you can see the network providing connectivity, but at the same time, without compromise, providing other sources of information like feeding a digital twin. And the use cases are much wider than just drone detection, which in itself is very important. That's why we went direction kind of demonstrating crowd sensing and crowd detection. But imagine a network that is deployed to provide connectivity in a city could be the same very network that is monitoring the vehicle traffic in real time.

Guest Speaker 12m33s

That's a source of monetization in itself.

Guest Speaker 12m35s

Yeah, and a lot of people in the financial industries are looking for where's the next monetization opportunity, and here it is. So what's next for the six gs Innovation Forum?

Guest Speaker 12m48s

We'll keep working. We'll keep I'm sure we'll keep adding companies going in the direction that you correctly pointed out, maybe non traditional companies that can help us understand the use cases. We'll keep developing trial versions, demos, engaging companies, testing interests, we will keep working Direction LA28 because at the end of the day, I think that's the intersection between early readiness and opportunity to demonstrate and display what six gs is going to be.

Guest Speaker 13m19s

Yeah. So, Iago, I think with that, I wanna say thank you for coming on the podcast. Really appreciate it and telling us where Verizon and other companies in the ecosphere are going with six g. Really exciting.

Guest Speaker 13m35s

My pleasure. My pleasure. Thank you for inviting me.

Roger Entner 13m38s

Thanks, gentlemen. Roger, we'll talk next week.

Guest Speaker 13m41s

Thank you.