Episode 22: NVIDIA’s Telecom AI Push, Europe’s Satellite Alliance and the 6G Spectrum Race

Peter Jarich (00:00)
Hey everyone, welcome to yet another episode of Intelligence & Noise, the GSMA Intelligence podcast, where we take a look at the week's news, we we sift through it, bring you some things, try and make you a little bit more intelligent, try and sort through the noise and maybe highlight some things that you haven't heard. We do it through the lens of the GSMA CEO newsletter. if you get it, you know what that is. If you don't, just just

trust that it's for for smart people and and we're gonna try and make you smarter by talking about it. I'm usually here with my friend and colleague Tim and I'm here again this week. Tim, how you doing?

Tim Hatt (00:40)
Good, Peter, how are you?

Peter Jarich (00:42)
I am I am good. I was excited last week I was at GTC

Tim Hatt (00:46)
Yeah.

Peter Jarich (00:47)
Berlin. We'll talk about that, but just as exciting and I flagged this. I was right across from Canada House.

I messaged you. You did not send me the secret knock. I I tried to get in. I I could not figure out what the secret knock was. I think it's actually the embassy. I mean, for what it's worth, is that sounds I didn't know what it was, but I believe it's.

was the the embassy there. No, I did not have I I I did not manage to to break in. so we are going to talk about we're gonna do it a bit different this week. Usually what we do is we talk about one of the two stories that got into the newsletter, one of the two stories that didn't, that was suggested by by the team and didn't quite get in there, and then a couple other things that we thought were important.

We're we're gonna ship things around and talk about spend maybe a few minutes on the things that did get in, and then the things that didn't, but probably talk a bit more about GTC because everyone wants to know what's NVIDIA. It's the most valuable company by market cap in the world. So everyone kinda wants to know what what's going on. And as I look back at it, there was a lot there. So we're gonna dig into that into a minute.

You know, as usual, I kick off with some sort of some sort of story. This week I'm coming to you. You probably recognize this is not my usual setup here. Don't have the cool, cool microphone. I'm coming to you from China. I'm here at an analyst conference with a customer. I'm here with about five other folks from the GSMA Intelligence team and a bunch of other analyst firms. so flew out here on on a Sunday afternoon, which is the

Only time I ever really go to see movies. And so, Tim, have you seen Supergirl?

Tim Hatt (02:33)
I have not. How is it?

Peter Jarich (02:34)
No, you have not.

probably because you read the reviews and the reviews said you don't need to you don't you don't need to see Supergirl. It's a kind of you know, y you watch it when you're like, you're not paying the the the the twenty pounds to go see a movie. And I was pleasantly surprised. It was it was actually pretty good. But what I was struck by is was one of those things where it starts with sort of a plot

Y y you think the main character's going off to do something and they're sort of drawn off to this side and that side and and you kind of lose track of of what's what's going on. not a not too dissimilar maybe from from the Odyssey where, you know, w we know Odysseus,

Tim Hatt (03:14)
Yeah. Mm-hmm.

Peter Jarich (03:15)
we know what he's trying to do, but then you retell the tales of all these other things and the plot goes in various folks. But what I think was interesting, I mean except for the fact that I mean I'm not trying to say supergirl as like

Right, for for the record, I am not saying Supergirl is on is is is is on par with with the Odyssey. I mean Supergirl had a dog, I'm not sure Odysseus had a dog, thus Supergirl wins. But it's interesting because you sort of go back and forth, and in some sense you're not really sure what the goal is. Very different from, as I think about NVIDIA and how I think about NVIDIA and their role with telecom. There's a whole lot of things going on here, right? We can talk about Sovereign.

We can talk about AI grids, we can talk about AI-RAN, but at the end of the day, we know what the goal is. The goal is to sell chips, the goal is to sell services. And one way or another, you kind of always know what the story is and how it comes back to that. And I think more than anything, you saw that this year. And we're going to talk about that in a bit, but before we talk about that, because that was one of the two stories in the CEO newsletter, a bit of what took place there before that. The other story, Tim, was something that we've we've hinted at in the past.

The fact that bunch of European operators, the big four European operators, came together and they'd suggested in the past that they were going to bid on some satellite spectrum and it looks like they finally said the said said the quiet part out loud or or just just acknowledged

Tim Hatt (04:43)
It did.

Peter Jarich (04:44)
what we already knew.

Tim Hatt (04:45)
They they made it official. so so the news here is is a joint MOU and planned by, as Peter says, the the big four operators in Europe, Vodafone, Telefonica, Orange, and looking across Dutch Telecom and looking across to the MSS planned auction that is likely to happen in twenty twenty eight, and that they will jointly bid for the portion of spectrum that is available for European companies.

And I think it's important because two reasons. One, it speaks to sort of solidarity among the telcos about jointly creating a neutral access layer for D2D, which they all know is going to be a big part of network strategy going forward. And two, it reminds us of the competitive dynamics that are implicit within this auction because it's set up so that a third of the spectrum goes to the Irish Squared sovereign constellation.

A third to European companies and a third to the open market, i.e. Starlink or AST or anybody else. And so this shows kind of a preemptive measure to get ahead of it. So I think it's important, Peter, we take that into account.

Peter Jarich (05:52)
And and and and and just and just

j just taking a step back, w when we say a third's gonna go to the IRIS² constellation with European operators, the other third though is to new entrants, correct? So it's it's it's one third for European operators, one third for one third for new entrant type operators, one third for for anyone and as brother who who could be out there. So so there's no reason

Tim Hatt (06:10)
Correct, correct, correct. And

Peter Jarich (06:13)
to think that the European operators couldn't bid on two chunks, right? I mean I think I think the expectation is that last third is gonna go to someone with deep pockets.

Tim Hatt (06:22)
That's right. It's on the open market and therefore subject more to to you know the bidding pressure, I think that you're gonna see. Of course, you could see that on the European portion as well. But I think, Peter, what's significant here is there's only 30 megahertz in this auction. And so when you look at that, there's gonna be more demand than there is supply. And if this consortium does secure that third, the question becomes how many SATCOs can serve that? Is it one or two?

Probably not more than that. And so this is an interesting print to play, given everything we're seeing with SpaceX who came out with the results last week.

Peter Jarich (06:56)
And there's no indication that this would be the case, and given the fact that these are the four biggest operators in Europe, right? But you could see another consortium come I mean, you could see you could see a bunch of other

Tim Hatt (07:08)
You could. You you you

Peter Jarich (07:10)
operators come up and say, Hey, yes, we're gonna put together a viable bid separate from them.

Tim Hatt (07:17)
I think you could see that. You could also conceivably see more operators joining this consortium. and and so I don't I don't think we can rule that out. I think anything's in play here.

Peter Jarich (07:27)
we can go back to what we saw in the US, right? Where what we saw in the US about the three operators come together and sort of say, Hey, we're going to figure out a way to share this in a pro-competitive way that looks at a mu bunch of different constellations, you could see plenty of different operators. It it's I guess just going back, I mean, from you know, the whole concept of an

the various commercial folks and various administrations outside of Europe have already complained about this and and said that they don't they don't like the way it's taking place. But is it too much to say that we have to look at this at least through the lens of what did or didn't happen in Ukraine, like as of as of recently,

Tim Hatt (08:10)
Yeah.

Peter Jarich (08:10)
where where you have a country who is depending upon satellite services for defense and to protect themselves and then

For some reason it doesn't work and there are various news outlets saying that the one reason it doesn't work is 'cause one band made that decision that it wasn't gonna work.

Tim Hatt (08:34)
You're right. Ukraine is a flashpoint. And and and we actually had a media sort of round table last week to to release our our Q three satellite tracker and and this question came up, I think this in sovereignty. And and it's interesting, like the story you get from SpaceX, from Musk himself, has actually changed. Where the original problems were seeing, no, that's the military's decision, I've got no involvement in it, to, well, actually we can throttle speeds if

commercial satellite use is being turned into military use without our permission. And so you don't really know which it is. I think suffice to say, there is a big sovereign imperative among European policymakers to ensure their connectivity is safeguarded.

Peter Jarich (09:18)
Yeah, I mean this has to be I mean th if nothing else, this was for all the folks that were saying, Hey guys, we need to watch out for this. This is probably them going, this. Like this is like this. This is what we're talking about. And this is the kind of thing

Tim Hatt (09:32)
Yeah.

Peter Jarich (09:33)
that that that we need to be careful of.

Tim Hatt (09:35)
And and and who would have thought 10 years ago that a satellite spectrum auction would be the talk of the town? we were not this wasn't on the radar. As much as direct-to-device (D2D) was was was a possibility, the fact that there's this much tension and industry preemptiveness going on speaks to the the critical value that it is. So I don't think the topic's going away. You're right to draw the read across to the US, but really really interesting eighteen months ahead of us.

Peter Jarich (10:05)
Cool. anything else or should we move on to the other topic that sucked

Tim Hatt (10:10)
Let's go

Peter Jarich (10:10)
all the oxygen out of the air last week?

Tim Hatt (10:12)
let's go to NVIDIA and and fly over to Berlin. So give us Peter, you were there. You were there on on

Peter Jarich (10:19)
Yes.

Tim Hatt (10:19)
the halls and and everything. Give us give us a sense of the vibe, first of all.

Peter Jarich (10:24)
I mean it's it's it's it's definitely if you can find the time and and get there and and folks can go it definitely any of these events, and I think there's another one coming up in DC. the each obviously are going to have a different a different feel, right? So the the big one in in Silicon Valley and Santa Clara is obviously, you know, everything, right?

The DC one I feel is gonna be more policy. I think you're gonna see some different angles. obviously in Berlin, you know, this is becomes de facto, sort of all the AI industry coming together. And so it's everything from AI. You've got, you know, folks from from very narrow use cases to data center builders to to everyone in one place.

And so it's it's exciting just to see the breadth of the industry, all the verticals being treated. Obviously, big European focus, big big focus on industrial. as always with these things, you get a keynote kicked off by by Jensen Huang, the CEO and founder, which is, you know, these things run from an hour and a half to two hours of

Generally one man on stage speaking nonstop with a few video breaks where he goes off and I'm sure has has a guy that hydrates, does don't know what what what he does. but it's it's impressive because when you do have something like that where it's one company that kind of sets the agenda, that keynote actually becomes so important because it is the agenda for the whole show. I mean, it's their show, so you get a feel for that. And and it sort of sets the stage right away.

this year, you know, pretty clear. And and, you know, I will say I think it is wonderful in a use case of use case of AI. You can sit there and take my notes the whole time, right? Look back at, okay, here's here's here's the key messages I see. But you know, we all now have a have a a a friend who is fueled by tokens who can back that up and say, okay, what did you think were the big topics and go, okay, I missed that or I didn't miss that or you missed that.

big ones obviously we're here for Europe. We're here for Europe. we've got European investments, and here's European model makers. And you know, this is how we are supporting Europe's agenda to make sure that we can support you with homegrown capabilities in terms of models and in terms of data centers. A big focus on open,

Tim Hatt (13:01)
Mm-hmm.

Peter Jarich (13:02)
right? In particular, open models.

obviously even well before the acquisition of Hugging Face, NVIDIA was big out there in terms of the open side of things. I think there's an interesting angle there that didn't necessarily get played up a lot in the keynote, but came out in some conversations, but just really focused on the open side of things. what's interesting is I think also taking a step back. Actually, I'll come back to so security, agentic security. Watch a new product, open shell, or a new service called Open Shell.

and around that highlighted, hey, this is all about agentic security. Now is the right time. There's there's this is the perfect time for agentix security. And and I said in a blog post, I said if I'm if I'm if I'm if I'm hugging face, I'm probably thinking, no, the the right time for agentix security was probably a year ago or at least six months ago before this, but it it does mean it's particularly important. a lot of hardware innovation, they rolled out some some new platforms.

a big integrated pod. So they've obviously sold these big scale solutions where you add a whole bunch of cluster, you know, you cluster a bunch of servers. They sort of pre-package that in the supply chain to make it easier to deploy. redefining physical AI. So again, it was Germany, industrial was important, but sort of saying that physical AI is more than I think this is about w we keep hearing physical AI is a focus on the from the industry and focus from NVIDIA and trying to take a step back and say,

It's not just robots, don't don't just don't, you know, we mean this. This is a huge opportunity. And it's not just robots, it's robots, it's cars, it's medical imaging, it is mobile network base stations, anything where AI is having an impact on interacting with the physical world. So a reset there.

Tim Hatt (14:47)
Yeah, yeah.

Peter Jarich (14:49)
But then what I thought was really particularly interesting in terms of a of a reset was it kind of felt like a philosophical question at some point.

Towards the middle of or actually the beginning of the the keynote, Wong said, Hey, we have to ask ourselves what is AI? And and NVIDIA's put out some I think it's blogs and they've talked about before of AI as this as this layer cake. Everyone likes to think of layer cakes. back in the day T Mobile in the US talked about the spectrum layer cake, now they've got the layer cake of of AI, which was really about this idea of you've got energy and power and land.

And shell as sort of this base layer, and then you've got compute, which is the silicon, and then you've got the infrastructure which is built upon that compute, and then you've got the LLMs, which are trained on and live on there, which power applications, right? So you've got sort that base foundation, energy and land, and you've got chips, and you've got infrastructure and LLMs which power applications. And there's your sort of five layers. But what was interesting is okay, great. Thanks, thanks for telling us what AI is or what you think it is.

But the important part for them was to say, look, we're in more than you think of us as a chipset player. But to make us work, you need the land, you need the energy, great. And he kept coming back to the need for energy. But yes, we're delivering compute, we're delivering chipsets, but we're delivering the infrastructure. We're actually developing models, which means that that we're actually developing and sort of so showing that they're more than just a chipset player. And I think it was a big, you know, you're gonna continue to see this from them of.

making sure people actually see what they do. So that was the big sort of picture. There were

Tim Hatt (16:25)
Yeah. Yeah.

Peter Jarich (16:27)
some telco specific messages which we can go into, but that was that was enough that was that was a enough

to to to give me enough to think about for a while anyways.

Tim Hatt (16:38)
No, no, no, super interesting summary. And and and there's a lot to unpack there. I mean, I just coming back on a couple of things you said, 'cause i the point you made about the layers of AI, and I know Jensen Huang made this point clear. You've got your power, you've got your land, you got your shell, etc., right up to this top of the stack. It seems to me that NVIDIA is really trying to embed itself up and down that value chain as the poster child of AI. I know they talked about AI as the next infrastructure cycle. Like

From a telco perspective, if I'm an operator, I guess, what do you think about AI as an infrastructure cycle and having to disaggregate that from traditional 5G or 6G cycles? Like, do you think we're moving into more of a different world in that sense?

Peter Jarich (17:22)
Yeah, I mean I think I think I think they know they have to take part in that cycle. I think the difficult part is they kind of know what those other cycles look like. Right? They know what those other cycles look like in terms of okay, it's gonna come every ten years. I kind of know what I'm going to do. and that's a really good point because they also make those decisions and then go forward on it. Right. And I've had this conversation a couple of times over the past is that we've talked about open ran and the reason why.

the the you know general wisdom is the accepted belief is that open ran didn't really succeed because or it was off to a bad start because by the time it came around the five the big 5G investments were already made. And so if you had been earlier and maybe were viable before it could have been chosen in those early 5G decisions, but it missed that and thus you kind of missed on some of those big opportunities. But that kind of points to the fact that you make your decision and then

These things last for a while. And yes, you could integrate those open-brand solutions, you could integrate them alongside things, but regardless, you sort of you have those big chunks of decisions of how the technologies are made. AI, however, is is those AI investment decisions and the way AI is developing this is is not following that same cycle, right? Now, as the question there is, are we investing in AI to support our radio access networks, which is a large share of the network CapEx for mobile operators?

Are we investing in it to support data center needs? I think it just ends up getting integrated into that. And so I think it's it's it it is a different conversation, but then comes back to the question for telcos of particularly around the radio access network, is how quickly do they jump on it? And how why and is it gonna actually be a good investment?

Tim Hatt (19:17)
Well, and just data point from our from our analysis and our surveys, you know, operators tell us they're spending about two and a half percent of their capex on AI. NVIDIA is a bit higher as you as you might expect, but that's going up. I think another interesting threat, Peter, is on the monetization side. We see from our tracker and telco AI that 50% deployments are increasingly moving towards monetization. 50% of new deployments last six months moving that way.

Did you hear anything on tokens, tokenomics and how that is shaping up?

Peter Jarich (19:52)
that's a really good question. So if we look at and let me just think of like what the big what the big news was from a telco perspective. So I think and and it was said I think in in some of the the emails afterwards with some of the analysts, like hey telco kind of got top billing and everyone's gonna say that Virgo got top billing, but it really did. Right. And not top necessarily industrial probably got more.

But for one of many verticals, you know, telco got a lot of focus. and you saw a lot of that through you saw a lot of that through announcements from Nokia. let's not forget NVIDIA has a billion dollars for two point nine percent of of the company invested in in Nokia, so so there's clearly a a connection there. And then with a lot of operators, some from Cisco and obviously NVIDIA itself.

And you know what's interesting from from the operator standpoint is, you know, it was a fairly large breadth. Obviously, a lot of European operators, right? But you had T Mobile in the US, you had DT, you had Elisa, you had BT, Orange, and then also some in from SoftBank. And a lot of this was you definitely saw a broad set of of what these folks were doing.

Right. So T Mobile Lisa was really talking about delivering these field trials of AI-RAN. And to be fair, this this was focused on how we power RAN assets with AI. Right. So T Mobile's doing it. You've got Elisa doing it in Europe. Elisa being a sort of multi-workload launch, meaning it's not just RAN, it's voice data, but you're doing video analytics. So this idea that you could do more than just powering the mobile side of things with also different use cases.

DT was talking obviously about physical AI, industrial robotics, but validating some of the throughput claims that that NVIDIA has made. And as a reminder to folks, think we called this out before, you saw some of those throughput claims, particularly from Nokia, as they're talking about what you could do with AI-RAN, saying you could get they have a roadmap to get to a hundred percent throughput gains.

Right. So you're going to totally get more value from your spectrum. People, I think, a lot of operators kind of like, well, that's just not possible. That's just that's and so I think showing how this works is particularly interesting. you had Orange looking at distributed voice AI inference. BT talking about actually using some of the models to inspect field infrastructure. So you saw the whole breadth here, which I think is interesting. To go back to your question though.

The monetization side of things, right? I mean, I mean, we're seeing that this works, great. We're seeing AI-RAN works. We're seeing some proof points for those efficiencies. Great. We're seeing how maybe we could support field inspections for the infrastructure. The monetization angle was sort of under the radar, right?

Tim Hatt (23:07)
Mm-hmm.

Peter Jarich (23:07)
Yes, you could kind of see this, the Elisa thing talking about.

multiple different workloads. You go, okay, cool. We could actually use this for things like those video analytics. And maybe we could monetize that, right? Or you could look at some of what Orange was doing with AI, with with inference and voice, and you could say, okay, cool, maybe there's some distributed applications there. But very much the monetization story was much more about the efficiencies. And to be fair, and we've called this out

That if you look at folks like DT at their AI Investor Day, they were talking about from an expectation standpoint of what they're gonna what they're gonna make from AI, that they saw was like two and a half times the the value from efficiencies than it was from new revenues. So it

Tim Hatt (23:57)
Yeah.

Peter Jarich (23:58)
is the right story, but you didn't see a big monetization push. You didn't see really the token beatering come up much at all.

I think it came up in a panel session that I wasn't at because I asked that question of some of the no of the NVIDIA folks, and they said, Hey, yeah, we actually called it out in the panel. But I was like saying, like, this guy's this would be this would be the place where I'd expect to be talked about, if nowhere else. And it feels like it it really wasn't the topic as much as it was the efficiency. But again, if that's where the telcos are focused, and if they have a hard time leading some of these big claims, that was still the right message for them to be telling about talking about.

Tim Hatt (24:37)
I it's interesting because that in a way chimes with what we see from the operators and when you ask them about how important is ROI and AI investments and they say, Yeah, it matters, but it's definitely not the only thing that matters. this is a long term play. And so it's interesting to hear you talk about that. One other thread you mentioned at the start was around how this was very much a European focused agenda and supporting

European governments, European companies, mobile operators, and others among them. How much did the sovereignty dimension come into it? And and was there an explicit mention of how that how sovereignty would be satisfied?

Peter Jarich (25:20)
Yeah. that was another thing where, you know, coming out of the keynote, for example, didn't really like I came away going, it felt last year, because last year's GTC, the European event, was was done in Paris. It's like I kind of feel like it was muted compared to last year. And then I could go and, you know, again, upload the up upload the video or, you know, put the video into your favorite AI tool and say, hey, how much they're like, yeah, there was one mention.

You know, it was explicitly mentioned once in in in the keynote. now I I asked the some of the NVIDIA folks and and I talked to the Nokia folks in in a briefing saying, like, guys, you know, why do you think this is? And you know, one answer which I think is completely fair is there's a lot to talk about. There's a lot, you know, so so you can only talk about so many things. I think maybe it's a little bit implicit. It might be part of the, you know, just part of the atmosphere of the way that that you think about this.

I think it could do to be a little bit more explicit, because I think there's a couple angles there that I think are particularly important from a NVIDIA perspective, that people don't necessarily understand. So one of the things that Nokia folks called out was, you know, the push on open models and the push on you know, the NVIDIA's focus on open. Is open is how you're gonna

See some of these sovereign models launch anyways, and how you're gonna see sovereign solutions. And so de facto open sort of is part of the sovereign solution. I don't think you can say one equals the other, but it's a good connection there. But what I thought was particularly interesting was, and you saw this buried in some of the messaging from Nokia, is that if we are running agentic automation, we're running some of these agentic tools for autonomous networks, where are you gonna run those?

Right. Because we're talking about control, we're talking about control of the networks. And so as much as there has been a focus on how do we support operations, how do we support some OSS as well as BSS? But in this case, let's talk OSS because we're talking operations from a cloud infrastructure. Who owns that cloud infrastructure becomes more and more important if we're thinking about this as a sovereign solution. Right? Because

You know, you can get to a situation where I could host my OSS and I could host operations on you know on on public cloud. But if if someone sort of starts to think of that or or an administration goes, Whoa, whoa, whoa, whoa, whoa, do I really trust that? Then you wanna have a backup or you wanna have a solution. And if I'm beginning to do, and again, if if I believe that.

Most people do that higher levels of network automation require agentic AI. Then I'm gonna want to have a way to host that my own, or a way to at least run that that doesn't rely on public cloud. So I think that was I think for me that was the interesting part. The rest of the stuff, of course, you were gonna see, right? You know, there was new models for AI-RANs, some of the new aerial models. I think a 2U and a 4U that accompanied the the ARC.

Pro and the and the ARC machines that were launched before. There were 6G demos, all the stuff you would expect. But I think that, you know, that big focus on AI-RAN proof points. We talked about this. Here you go. And here's the breadth of how it's being used. And then that sort of sub-angle of how do we think about autonomous operations? What security do you need for that? And how does that fit into Sovereign? I think that was.

Again, it was a little bit under the radar, but I think particularly

Tim Hatt (29:13)
Mm-hmm.

Peter Jarich (29:14)
important going forward. And we're gonna see that. I think we're gonna hear more of that. Who knows? I mean, I think particularly DC's coming up, GTC, DC, so I think regulation may be a part there. And then of course CES will be next year. They did something last year, we'll see it this year. But it was it's still a lot. I'm still digesting a

Tim Hatt (29:33)
So much.

Peter Jarich (29:33)
half of it. Yeah, including the

If you haven't gone and watched the if you haven't gone and watched the the keynote, it's it's out there. If nothing else, skip ahead to the end, watch the last five minutes for the recap with a Eurovision style robotic song and dance that recapped the whole thing. There's even a shout out to David Hasselhoff, 'cause it was in Berlin. so yeah, that was yeah, creativity.

Tim Hatt (30:00)
May In

NVIDIA may join Canada and entering the Eurovision contest next year. We can use

Peter Jarich (30:08)
I mean I mean they might they

they might they might they might allow them to do that.

Tim Hatt (30:13)
Okay.

Peter Jarich (30:14)
I mean just by the time this comes out, I think we will have had two blogs up and potentially a spotlight on the platform. So go check them out and as always reach out. I'm sure there's tons of stuff I missed or won't even realize till someone asks me asks me about it. so this is normally the time where we

Move over to the the Willie Nelson Memorial. Willie's not dead, I don't think, but call it the Willie Nelson Memorial segment on the road and on our minds. On the road, you know where I am. I'm in China. Tim is is holding down the fort. On our minds. We usually share things that didn't

Tim Hatt (30:56)
Mm.

Peter Jarich (30:56)
even make it into consideration for the CEO newsletter, but this week we're gonna talk about things that made it into the into the newsletter.

Tim, I guess I'll kick it off. This was a I don't want to say slightly minor, but but slightly interesting piece. It was Verizon with Ericsson launching or at least talking up some work that they were doing around seven a 7 GHz field trial for for pre-6G or

Early stage pre-standard proof of concept. I can kind of see why it didn't make it. And it's kind of feels like it's we're gonna see a lot of this, right? We're gonna see a lot of people. I also like I have my own thoughts. Like what why do you think companies talk about this? Why do you think it it feels a bit it feels a bit inside baseball? It feels a bit like

Why do we need to know that you're doing this? What what what's what's what's your thought if you're a marketing person?

Tim Hatt (32:07)
Well, I think if you're from a vendor, so in this case from Ericsson, or or if you were from Nokia or Huawei or or Zeti or any of them, there's a marketing win here in showing you're working with a major customer on a new spectrum band. So we're talking about piloting the use of second 7 GHz spectrum for 6G. and they ran it with a four hundred megahertz carrier for a call in California. I think more important though, Peter, is the signaling effect. When we look at this

We have the GSMA released research this year that is estimating an additional two to two and a half gigahertz of spectrum will be needed for 6G. But one of that, maybe a little bit more, has already been pretty well lined up. but when you get to the gap, you know, this upper six and 7 GHz is squarely in frame. And so the signaling effect, I think, here is to show guys, if you're a regulator, if you're in the ITU and others, this we should be looking at.

Peter Jarich (33:06)
Yeah. And I think I think for those folks who haven't seen it, it's the Vision Twenty forty report. talks about spectrum for the future of mobile connectivity. Yeah, two to two and a half gigahertz, global average required, two point four to three point three gigahertz extra needed in high demand countries. I think that's definitely part of story. I think for me, and part of why I thought it was interesting is the rigor with which they go into what they do.

did is actually particularly interesting, right? So they talk about the base stations installed in the normal shelter, great. The user equipment was a little small house on wheels. So they could test stationary indoor and outdoor, test throughput, latency, stability, line of sight. Like they went through a lot of details about about what they're trying to do there. So I think from our perspective, 100% right, this speaks to the need for extra spectrum.

But I think there is also a bit of signaling there to your point of showing how quickly operators and countries for what it's worth are are preparing for this and what they have to go through to to get ready. and obviously we know that the US and China in particular are both sort of in this race to see who can move forward with it. So it is I I expect that this is the first of many that we'll see over the

over the upcoming and what it wasn't clear was like what the size of the antenna array was going to be because obviously every time we move to a new spectrum band the story from the vendors is that coverage matches your existing grid but that generally only works when that generally only works when you've got a whole bunch of extra fancy antennas. So still some questions out there. What is your what was the other story that didn't make it in there, Tim?

And maybe should or shouldn't have.

Tim Hatt (35:01)
So switching gears completely.

Yeah, well, switching gears completely, it's story on quantum computing and and quantum-safe networks. and the news is that Toshiba, company we don't talk about that much, but a very important company in the computing industry, has released a post-quantum platform that can be deployed on the networks of telco operators, data center groups, enterprises, potentially public sector authorities.

to preempt the risk of quantum uh-based security threats. And I think, Peter, what's interesting about this, so this was it it's deployed for customers across a variety of segments. But I think what's interesting about this for me is that it can integrate with existing networks and and therefore doesn't require a brand new greenfield build. Because what we what we hear from our own

you know, operator customers and others is that only about twelve to fifteen percent have actually deployed post-quantum defences right now. So there is a big gap, about a third plan to, but the big barrier to that is do I need to invest a lot in new networks network infrastructure.

Peter Jarich (36:16)
Yeah, yeah.

Yeah, yeah. And I think I think like so many and this is this is I think probably the reason why I think so many people kind of just like like zone out when we talk quantum is that, you know, the announcement was full of you know, just full of all the great acronyms, right? Post quantum cryptography, PQC and you guess your quantum key distribution and your quantum random number generation. I think it's like I

Every every conversation around quantum feels like it's a math problem, like like right before you get to the commercial side of things. And I I I don't think I'd realized really going into this or before this, like that that the way most of these solutions were being set up was requiring essentially a a a new build, right? And that, you know the value prop here was leveraging existing fiber. And it wasn't that you needed a

you know, to connect a hundred customers. I think the example they give is that, you know, previously conventional deployments that if you're connecting a data center to a hundred customers, you need a a hundred central site boxes with a hundred in the field. Right? Well that's not I mean, that's going to limit this to obviously specialized use cases where you really, really, really, really need to be secure doesn't make it very scalable. So the fact that you can actually have one hub

Connecting sort of shared central equipment to to multiple remote endpoints. So yeah, I completely see that being being important. but yeah, surprised that that wasn't already the case.

Tim Hatt (38:00)
But I think it's it's also interesting whenever we talk about quantum, i i it's so easy to even lose this topic given everything is about AI now. And the reality though is there as much as there is an opportunity through quantum, and we've published some research, you can check it out our website, we'll put it in the notes. Under ten percent of IoT devices are quantum-safe. We're gonna extrapolate that to mobile networks, and I suspect the finding will be quite s quite quite similarly small. It's a hidden risk.

And it's a shadow risk. And so these types of solutions, and if we get to a point where there is an alignment on an on a on an algorithm for for post-quant cryptography, is gonna be really, really important because it's one of those things where you kind of don't talk about it, don't talk about it, don't talk about it until it actually matters and you've lost a lot. so I think these news stories are are kind of a reminder to us that it it for operators and others, be at the forefront of it.

Peter Jarich (38:57)
Yeah, it's been it's been on the horizon for a while and it will be on the horizon until it's not, right? Until I think

Tim Hatt (39:03)
Yeah.

Peter Jarich (39:03)
un until all of a sudden people are like, okay, now we need to now we need to pay attention. But better to to get out ahead of things. wonderful. Tim, I think that was our last story. Any last any last stories that we missed or anything?

Tim Hatt (39:19)
No, I think we've we've covered the bunch. We're gonna be releasing a lot of new stuff the next week, so we'll put that in the notes, but covered a lot of ground.

Peter Jarich (39:28)
Yeah, so definitely go check out some of the things that we've called out. be around next week. I'm headed to the US next week where I will be at the Brooklyn Six G Summit. so I think actually I'll be home when we film this next week, but then I'll be off to the Brooklyn Six G Summit as well as Future Net North America, both of which are kind of

conveniently in the same city over the same days. So we still have some some things to keep us busy. After that, Tim, you'll be in Doha later on in the month. So please come back because the year has still got a lot of stuff coming out, a lot of stuff to keep us busy in Alaska to tell you about. As always, send us notes, tell us what you like, tell us what you don't like, tell us what you want to talk a bit more about.

what you'd like to hear from us about leave show notes reviews. I understand from people that actually know what's going on. That reviews are how people find out about podcasts. So so please leave us some if you are so inclined. and with that, let me just say thank you, Tim. Thank you, everyone,

Tim Hatt (40:40)
Thank you, Peter.

Peter Jarich (40:42)
for for joining us. And as always, I remind you.

The that in your search for intelligence, don't forget to embrace a little noise and that's what we're here for. Thank you, I will talk to you all next week.

Tim Hatt (40:54)
Thanks everybody.