Episode 3: Sovereignty, Satellites & AI: The New Infrastructure Race

Peter Jarich (00:00)
Hi everyone, and welcome to Intelligence and Noise, the podcast from GSMA Intelligence. This is week number three. So by now I shouldn't have to introduce myself, but I will. I'm Peter Jarich, I'm [with] the GSMA Intelligence team, and I've got Tim with me. Hi, Tim.

Tim Hatt (00:16)
Hey Peter, how you doing?

Peter Jarich (00:18)
Good. For those of you who don't know, Tim heads up our commercial content and consulting. So, as usual, we're gonna talk about what's taken place recently in the week, what we're seeing in terms of our research, how it links into that, what we think is important, and maybe what's important but people aren't paying attention to. So we're gonna get to that in just a second. Also talk a bit about where we've been and where we're going.

All of that. But before we get there, as is my want, I'm gonna kick off with a quick story that hopefully sets a bit of how I'm thinking about the week.

So, early this weekend, I was out for a bike ride, came through a small town. As I came through the town, I could feel that the handling was a bit off my bike. Couldn't tell why, so I went to pull off to a side road to check it out. But when you pull off quickly onto a side road, pulled up quickly because my rear radar told me there was a car not too far behind.

You pull off quickly and there's no air in your front tire, you may know what happens. And so basically splayed out across the road. For those of you who remember the great Tour de France where Lance Armstrong blew a tire and managed to somehow stay upright, I'm not Lance Armstrong. It was interesting 'cause it was fine. Shook off the driver behind me, who got out and looked at me, like, "Are you alive? Do I need to help you up?" And I don't think she believed me.

But what's interesting is, as I did that, I recognized that, you know, even if I wasn't, there were so many things connected to me, right? I had my phone connected to me, which probably could have sent an alert or contact to someone. The front bike computer has a sensor that says if you've crashed. That sensor is connected to my rear radar, power meter, the gears.

And it reminded me, as I sat around icing various things later on, of an analyst conference back with Qualcomm a number of years ago—and this may have been a decade ago—where the question was, you know, how many IP addresses do we think there are? How many IPs are there in a given household? What's the average gonna be? We think it's gonna be seven, you know, in a few years.

And really they're talking about connected devices. I think at the time there were a bunch of panelists going, like, "One?" Like, who thinks about that in terms of connected devices? And then really, seven.

And if you think about it now, I mean, two things I took away from that. One, I can't even count, Tim. I mean, I don't know if you have any idea how many connected devices, but when I just think of my bike alone, it had rear radar,

Tim Hatt (03:02)
Yeah.

Peter Jarich (03:03)
the bike computer, the gears, the power meter. Now, they're not all connected to the internet, right? But they are connected to the head unit that's connected eventually to the internet. And I think about the house in terms of, like, those numbers are just crazy, right?

But also, this was interesting because, from the analyst perspective—and there were a lot of telco analysts—it was sort of like, "That's a really weird way to think." I think it was a reminder of sometimes we need—and I think sometimes in the telco space, we don't always think in the same way that folks from, necessarily, a pure tech background think—and we sometimes need to get outside of those views.

I think for me that's particularly important because it links to a couple of things we're going to talk about today. One, in terms of an upcoming event, which will hopefully get some outside-in views, and two, a bit about perhaps the biggest outside-in big thinker, Elon, and the thing we can't all stop talking about, none of us can stop talking about: the SpaceX IPO.

So, a bit of a painful way to get into that thinking, but a reminder to everyone: wear your helmet. It's a good—so let's kick it off. CEO newsletter. What resonated? It's a little bit easy here because it's a big AI-focused week, right, Tim? I mean, you actually authored it this week. You actually were—or not authored it—you were the one who took the lead in finding the stories.

So remind me the two sets of stories that went through.

Tim Hatt (04:38)
So we had a story around security and looking at how that is being impacted by AI, for better and for the risks with that. And then, secondly, we had a story on the China and India AI push and a lot of data center investments going through hyperscalers and telco partnerships.

Peter Jarich (05:00)
And so I think, if I think about—obviously, all particularly important, right? We've been doing a lot of work in security. Security is particularly important. I think we've talked a lot about how the double-edged sword of AI around security: it makes it easier to launch attacks at the same time, it makes it easier to—or hopefully to—defend against them.

But I think it's worthwhile digging into more, at least for me, the more important one probably is the investments in China and India, really with a focus on building out capacity. And I think, you know, there's a couple things for me there right now. And I think it was highlighted particularly interestingly against the backdrop of, you know, connecting with some operators earlier in the year, of talking about how, you know, they had spare sort of data center capacity that they weren't using. They built out a whole bunch of AI capacity, had it, you know, running traffic, had it being used right away.

Right. And it was interesting 'cause talking to another analyst, they sort of positioned it as this idea that, you know, the model makers are chasing every single GPU coming out of NVIDIA to try and lock up that capacity as soon as they can. So we're in a place where it's not necessarily that hard, if you build out capacity, that people are gonna take it off. So it makes sense from a telco perspective that they're taking part in that.

But I think it's a little bit more than—it's a little bit more than them just sort of following those opportunities, because I think we're seeing a couple of angles. One, this also goes on hot on the heels of what we've seen from China launching AI-related plans for consumers and enterprises that are, you know, essentially selling tokens, right? You could consider this—I think a lot of people are thinking about this, and it's probably a little bit simplistic—but thinking about as metering.

You know, we've lost the metering battle in terms of selling gigabytes, selling traffic. Consumers sort of generally are going towards, you know, all-you-can-eat or limited plans. So does this allow you to actually get back to a place where you're making money based on consumption?

That seems to be the way, and then I think that—and the overall—am I reading that correctly in terms of how people are thinking about it? Because I think they're missing something, Tim. But is that fair in terms of what you're seeing?

Tim Hatt (07:21)
Yeah, I think that's fair. I think there's a capacity point and then there's a commercialization point to what you're talking about. I think from a capacity level, the sheer scale of the investment size tells a story in and of itself. China's got a nearly $300 billion package that is spread over multiple years, designed to build out capacity to service their burgeoning AI economy. The mobile operators are a key partner in that. We see the same in India.

Reliance is partnering there with, you know, hyperscalers. We see it with other Indian telcos as well. So I think there is a point of India and China forging the capacity ahead for their domestic AI ecosystems to compete with the U.S. and Europe.

On the commercialization side, we do see a new approach, as you talked about, with, you know, the tokenization and trying to introduce a new unit of measurement for how people consume AI. I think that will be a really interesting development to see if people understand what that means and how they translate that into their own calculations of whether it's worth it or not. But we're in the early stages of that.

And then, just last point I'd make: on the enterprise side, I think there's a lot of interest in monetizing that capacity. We talk about things like AI factories, specialized data centers. We're going to be putting out some research on that.

And we think for the mobile operators alone, that potentially has a 5% uplift to their revenue. And so if you have a cloud business in the low single digits or mid-single digits right now, and you're looking for an uplift there, this is a really interesting opportunity.

I think these two countries don't always get talked about in conversations, but I think we're seeing them that big. What do you think?

Peter Jarich (09:10)
Well, and that's why I think—as much as we can look at this as interesting and sort of as part of a larger trend—you can only think that so far, right?

I mean, China, as you pointed out, in China the operators are central to how enterprise services are delivered and how cloud services are delivered in a way that that's not the case in other places, right? I mean, you know, the operators in China are very much like the—you know, are much like the AWS or Googles in some way of what you see in other markets. And so it does make more sense for them.

Same—you are seeing that connection in India, and also there's more of an interest there. So I think that's one way we're probably looking at this a little simplistically.

Two is this story of the Chinese operators with these token plans, right? I think the simplistic way, and the way I've seen a lot of folks talking about this, again, is, "Look, you know, we found a way to come up with a new thing we can sell in an incremental way. Great."

But the reality is that it's doing a disservice to the fact that, you know, it's building on the fact that the Chinese telcos have already built out some capacity that they can leverage, and that they're doing more than just selling a bundle of tokens. They're selling, in some cases, packages that include access to a bunch of different models and other things on top of that.

So it's not just as simple as we're selling tokens. And that's probably good because—and it feels like, as much as we all want to, you know, and we know AI is going to transform it, is transforming, you know, the way people work, the way people live, the way people generate content, et cetera, et cetera, et cetera.

That, you know, we're in a place right now where—I don't want to say token generation is at its peak, but we've seen that we've gone from the place where you've got folks like Jensen Huang saying, you know, "I want my engineers spending as much on, you know, tokens or a share of their salary on tokens," just to say, "Look, here's what I call being productive," to the fact that we've got companies—

And that includes telcos—who want to run things and see the token costs are outrageous, right? You see pushback from companies saying, "What was it, Uber that said earlier in the year that they blew through their AI budget in the first several months of the year?" Because you've got this sort of token waste, which happens when people are playing with these tools plus using inefficient tools.

So it doesn't feel like it's genius to say we're going to get to a place where—I mean, we're seeing models built on the fact that we need to focus on token generation, token generation, and how do we scale this because it's gonna grow exponentially—to the fact that, yes, as more and more people use these tools, there'll be some scaling.

But the amount of tokens per person, either as models get more efficient or people are being—you know, people have to sort of cut back on capacity—it feels like we're building for a world that we're gonna look back and go, "Well, that was really silly to expect that token use was going to continue on that trajectory when the models weren't quite great and people didn't know what they were doing."

So that is my worry there a bit. It was a long-winded worry, but I do feel like we're building for a future where everyone is spending tons of time and money on tokens inefficiently. And I don't know that that's a really smart way to think about the future.

Tim Hatt (12:48)
To say nothing of the fact that if you compare the token system to other units of measurement, monetary or otherwise, with a megabyte of data or a gigabyte of data or currency, they have accepted values because they are universally used.

Right now, people forget the token economy is, as you say, it's an early economy. Pricing is set individually by operators. That will change by operator, that will change by country, that will change if you're a business, that will change if you're a consumer.

So my question is, do we start to see big variations, swings in the price of a token over the next year as you scale up this amount? I think there's a lot of questions.

Peter Jarich (13:26)
I mean, it already is, to the extent that I think, you know, clearly it's understood that the costs of generating tokens in China have been different than elsewhere.

I think you hit something that's really interesting there because when we think about some of these plans, right? And again, we put out some analysis around the beginnings of what's taking place in China, thanks to our colleague Radhika Gupta, who put together most of that. I think it's a really good review. We're going to follow that on with some work around sort of the pros and the cons.

And one of the things I'm kind of worried about is it makes sense, right? Some of these packages include access to a bunch of different models. But what that means is that if I run one sort of use case across different models, token usage might be different across those different models.

So do I—how does a consumer—I get that, barely, right? But how does a general consumer go, "Whoa, whoa, whoa, wait, wait. I, you know, my friend just did the following thing and he did some cool things, and I tried the same thing and I blew through all my tokens, and now I'm done for the month."

It's a fun one, but we have more important—we have other token-related things to talk about. So what didn't—the things that didn't get picked, Tim, I think included super app and mobile money stuff, particularly around MTN and Ant. I know that's important, but which of these do you think was the more important of the two that didn't get picked?

Tim Hatt (15:03)
So I think the most important of the two that didn't get picked is around the sovereignty side. Both are interesting, but I think here's what's interesting about the sovereignty side, for my money.

So number one, this is a mega push, particularly in Europe. I think the first point is it's very much a joined-up initiative between the private sector and governments, particularly in Europe. The announcements we saw that we were looking to potentially feature in that newsletter were around European carriers, Telefónica, Deutsche Telekom, that are partnering with U.S. cloud companies to ensure sovereignty for cloud AI as well as satellite.

Now—

Peter Jarich (15:46)
And to be—and just to, for those folks, it was the news Telefónica worked with Google, Telefónica's work around NTN, and then it was actually DT around Palo Alto on security.

Tim Hatt (15:55)
That's exactly right. And so I think the point that's interesting for me is not so much the partnership between these companies because Google, AWS, Microsoft all have existing businesses in Europe, extensive cloud businesses. It's the mechanism of control.

And so if you look at the announcements, whether that was the Telefónica-Google partnership in Spain, Deutsche Telekom with Palo Alto in Germany, the interesting thing here is the partnership is giving encrypted keys to the operators to manage the sensitive data within their own cloud.

So that whilst you can still have workloads hosted on U.S. clouds, the sensitive portions of those for highly regulated industries like banks or healthcare or defense or energy or pick a sector are encrypted and under the control of the operator on their cloud so that it aligns with the rules of their respective governments.

And I don't know what your thought is on this, Peter, but I think it's interesting because with the whole sovereignty question, it's not just about data residency and control, it's about competition. And if you have governments insisting on that control, does that potentially trigger efforts to restrict competition in the U.S. of European companies? And so I think this is a way to try and balance those things.

And it's interesting, but what's your reaction to it?

Peter Jarich (17:23)
Yeah, I'd say—I think obviously, and it sounds trite, and it sounds like it's a punt to say this whole space is evolving, right? But, you know, if I look at what was going on across all of those, like, I personally think the most interesting part was the NTN, was the broader NTN angle, right?

In part because some of what you're talking about, right, of how we manage keys, how we manage access, how we manage control, who controls things, has really been the—you know, the cloud side of things has really been the sovereign story, right? The sovereign cloud, how do I build out a sovereign cloud? How do I make sure? How do I have sovereign compute? The sovereign capacity.

And I think to some extent it's, you know, I look at this as access, right? But making sure you can control and you have some control over that access and that connectivity, I think is particularly interesting.

But I will admit, even as I look at it that way, the NTN angle side of being that—there is also, as we go forward, really also the compute side of things. It's just: where does that compute live?

Tim Hatt (18:28)
Yep. Yep.

Peter Jarich (18:30)
And can I have compute, right? Because one of the value props, right, is if we think about NTN being a connectivity solution, it's just essentially delivering connectivity in one way from point A to point B, but you've got a ground station in there, right?

And so I think in some cases then you get rid of the need for a ground station where—and I think we just saw this, it was just early this week, I think it's on Monday—that I just saw an announcement from Terma on a project in Sweden and Denmark.

It's called the Bifrost Project because why not come up with—you're in the Nordics, why not come up with a cool Thor-related name? But it's all about how do we process data? I mean, particularly in this case, defense, on satellites.

And the positioning was about sovereign, because hey, we don't have to go back, we don't have to go down to any ground station, no matter where that ground station is. Don't worry, this is going to be on the satellite. You can get rid of—and that's interesting how I think, I think really for me, the satellite side, which we saw, I think it was the Telefónica part, is an angle that we don't necessarily talk about a lot.

But I think as we've got folks who are building out these massive, massive, massive satellite capacities, we're obviously gonna talk about—

Tim Hatt (19:58)
And we're going to, when we come to the SpaceX IPO, there's a sovereignty angle there. But I think the interesting thing—you hit on it when you talk about, you know, the partnership there, which happened to be with OQ Technology here, which is a satellite company out of Luxembourg.

There are a lot of European suppliers who are setting up their business and set of controls to cater to the sovereignty requirement as a competitive USP against Starlink.

And again, I think that, you know, something lost in the news, Peter, I think last week, which was completely overlooked by the billions of articles on the SpaceX IPO, was the fact that the company has some big-time regulatory uncertainty in India. And that's, you know, the largest country in the world now. And part of that is centered on the sovereignty question.

So I think you're right. It's very much, you know, sovereignty is not just about AI, it's about satellite and the open questions around how you ensure that mechanism and control are live ones.

Peter Jarich (21:03)
Yeah. No, I think—and that's a good segue into the part of the podcast. We're gonna call it "On the Road and On Our Minds" and eventually our marketing focus are gonna get a Willie Nelson theme. We'll see if we can get Willie to join the podcast at some point.

Because for those of you who know, "On the Road Again" is a Willie Nelson song and "You're Always on My Mind" is—and it really makes sense to have Willie. So if you know Willie, tell him—tell him where him—

On the show.

That leads into the—we'll talk a bit about where we're going and upcoming events that we think are important for folks to know, but let's get into the SpaceX of it all. Like everyone, everyone has been talking nonstop about it. We obviously put out some early—I'm gonna kick off with something a bit from, you know—well, just for folks who are on the same page, right?

SpaceX had their IPO. Kicked off at $135 a share. I think right now, as we record this, it's at $160, so it's a 20% bump or so.

But I think one of the things—and this was a big chunk of what we talked about in the analysis we put out—I guess it may seem really, really, really pedantic or in the weeds, but sort of this idea of the TAM, right? Of what is the opportunity for the business as stated, you know, in their filing.

I think it's particularly interesting that they quoted this, like, $28.5 trillion TAM, of which 2% they're actually currently active in. I'm sorry, sorry, $2 trillion. Sorry, not 2%, 7%, 7%. So $2 trillion, which is the space economy, right? The launches, everything with satellites, the connectivity side of things.

And that's even being a little bit generous, right? The $2 trillion is kind of rounding up. It's including a little bit of the mobile side of things, but they're not really—you know, Starlink Mobile really isn't fully operational.

So, yeah, that's 7% of that TAM that they're actually—and I think that's the part that I think puzzles me the most. It's a business that says, "Here's these things that we will do. We're not doing them yet."

I feel like if you went into Shark Tank and said, "I have this interesting business idea. I am doing these three things, but there's a hundred other things I think I could do. How much money are you gonna give me?" That'd be like, "Prove you can do that."

I mean, is it just—is it—

Tim Hatt (23:32)
You—

Peter Jarich (23:33)
—is it just the assumption that they can—that Elon and Co. can do all those things? I mean, they are moving, you know, they're building out the capacity. You've got OpenAI, you've got—or not OpenAI, you've got everyone but OpenAI, excuse me. You've got Google, you've got Anthropic buying capacity.

So, I mean, what do you see as the—is it the Elon of it all? What's driving people to go, "Yes"?

Tim Hatt (23:56)
Well, let me give you my view and then I'll get your reaction to it.

So I think part of it is straight-up an Elon premium. The commentary and the conversations I've had with some of my friends in the asset management industry and the venture capital industry is that there is a premium for this guy alone. He calls the trends properly. He's called them right. And he is operating in a frontier market in the truest sense of the word.

When you go to outer space, there ain't no bricks-and-mortar stores out there. There's nothing there. That's just you.

So part of it is that. Part of it is sheer scale that they have over any competitors. We know that in the Starlink space, they control 90% of orbital capacity. We monitor that figure every quarter. That's, if anything, widening as they roll out their constellation.

And so they have the capacity lead. And part of it is on a point of tech innovation where a lot of the investment community wants to get in on what they see as an AI gold rush. We'll talk about the Anthropic, of course, the Claude maker, the OpenAI. Both of these companies are likely going to IPO this year with similarly heightened levels of buzz.

And people are seeing SpaceX as the talismanic company that can marry up the satellite and AI spaces with a greenfield area.

Peter Jarich (25:24)
And that's—and I think what I think is interesting there, and if I think about OpenAI, right? I think earlier in the year, OpenAI had a bit of a moment where they were saying, "Look, look, we needed—we need to pull back on these side quests."

They were doing a whole bunch of different things, focused in a bunch of different areas, and said, "Look, we need to focus," and we talked a bit about this earlier, I think it was last week, talking a bit about, you know, the FT article that said, "Look, they're all in on coding and enterprise," particularly as an IPO is coming in.

It's this idea of focus, right? But what's interesting is I think that's set—that's part of what stands—helps to, helps rather, you know, SpaceX stand apart. It just seems to be a series of side quests, right?

It's got, "Well, we've got the satellite business, and we've got the AI business, and we've got the connectivity business, and we've got the, you know—and yes, we in theory, they've still got Grok, right? And we're gonna—and we know that eventually Tesla is gonna be part of—you know, that at some point Tesla's gonna be brought into the fold to make it one big company and grow Elon's wealth as well."

And I think that is part of—I think, as much as you talk about an Elon premium, I think that's part of, I think, the audaciousness of this. It's everything and it's so big. I think it does make it so much sexier and exciting to people, as opposed to a, you know, from a business perspective, maybe not the best thing because it's all over the place.

But from a "I'm betting on the future, which includes AI and space and maybe devices and there's gonna be some silicon making in there"—it does make it a much sexier story. And they don't lose because they're all over the place, because they've always been all over the place. It's part of what the business is.

Tim Hatt (27:09)
100%. And it's interesting when you step back because we are in the start of a new tech era. That's clear, in the AI era.

But if you look back to previous huge kind of epochs or transitions to a new one, like the dot-com era, to a certain extent post-iPhone, there's an interesting question in terms of how the valuations look and how things turn out.

And I think the question of an AI bubble or correction is interesting, but ultimately it's more a question of how you consolidate the scale that we see.

And so I think right now what they're basically asking for is people to take a giant bet on growth prospects that don't yet exist, but that may exist.

But you tell me, Peter, if you think we're wrong. I mean, when we put out our analysis in our spotlight on the IPO, just before it went out, for Starlink Mobile, you know, they estimate that's going to be around $760 billion.

That's essentially equivalent to the service revenues that the mobile operators make right now. Now, we think that's more like 4% to 5% in terms of a service addressable market. In other words, just translated for our viewers, what we think they can actually achieve potentially of that.

Tim, are we wrong on that? Are we missing something here?

Peter Jarich (28:21)
No, but to be fair, I mean, to be fair, that's the definition of TAM, right? That is the addressable market that we're going after. There's no way they're gonna get all that, right?

I think—and that's maybe fair. Right? I actually—it's probably not fair to the extent that although it's the addressable market, they just don't have the capacity, right? You can do the calculations, you can do the physics.

And what I think is interesting is whenever you say that, right, or whenever we say that, or whenever we say, "Hey, look, you know, they might have an aspiration to go after all global connectivity, but the physics doesn't lie."

And there's only so much, given the spectrum that they have, and given what you can do with satellites, the answer, I think, you know, is always going to be, "Yes, but these technologies are evolving and they're getting better. And have you looked at V2?" You know, yes, they're on V2 right now, but what is the V5 satellite going to do?

Which I think is interesting because earlier this week—trying to marry up the whole space, the satellite of it all—I think it was last week, rather. I guess some initial specs of the first satellite data center from SpaceX came out. I just saw this somewhere.

It's the AI-1 satellite, which just seems—it seems like you're begging for AI references there, which other people have messed up on. But hey, I think they're probably trying to bait people into that.

And what I think is interesting is—and this is maybe coming back to the satellite side of things. So, I'm looking at this right now. It's got a 70-meter wingspan and a 20-meter deployment height, which, if I do my—I don't even need my calculator for this. You get 120 or 140 square meters. If I'm doing that correct—or, sorry, am I doing that correct? 70 times—yeah. I mean, that's that small, right?

I mean, especially since—

They're talking about this as—wait, I'm missing—sorry, I'm missing a number there. Right. Let me take that back. Sorry. That's actually 1,400. Sorry. See, I did need the calculator watch.

Tim Hatt (30:32)
Yeah.

Peter Jarich (30:33)
Wasn't it not long ago that we were all—that the whole industry was all on AST about going, "Your satellite is so huge"? And that was 700 square meters? Because it's twice as big.

And yes, I know the next further AST ones are larger, right? Right. But this is still like 1,400 square meters, and you're gonna need 6,700 of them to deliver a gigawatt.

The IPO has said—the filing said—we're gonna get to 100 gigawatts, which says we're talking about 670,000 of these.

Like, at what point? I know there's been a lot of people—are people talking about what happens when we get—and I know they've talked about how much, like, how much space we have in orbit, but there's a difference of how much space we have in orbit for a small sort of satellite versus a bunch of these things that are taking up a lot of space.

Is that something people are reasonably talking about yet?

Tim Hatt (31:39)
The short answer is not enough. And I think this is another kind of implied question with this whole space race. What happens if China fulfills its three constellations that it's got in there? Guowang, you know, Tianfeng, there are a number of that that are comparable in size to Starlink.

You start to get congestion, you get space junk, you have issues with longevity. What happens if India does that? You govern the rules.

I think these are all interesting questions. And it goes back to the point we were talking about of the Starlink proposition there is really around betting on a growth that hasn't yet materialized.

No one doubts the innovation and the quality of the innovation. You have to respect that. But we get into the sheer physics and the capacity available in low Earth orbit. There are question marks.

So these are things we're going to have to follow.

Peter Jarich (32:32)
Yeah. Yeah.

Tim Hatt (32:32)
—up.

Peter Jarich (32:33)
And I think it was easy to—it was easy to wave off some of these. There's a lot of space up there. Most of it's just empty. It's not taken up when you're talking about these tiny satellites.

But when you talk about, yes, I need to get to 670,000 to deliver what I'm saying in terms of 100 gigawatts, and this thing is not small, and there's many other people out there, it's gonna be something that we're talking about.

And to your point about the Chinese folks, I'm guessing you're gonna be talking about it next week at MWC Shanghai.

Tim Hatt (33:08)
Yeah, on the road, we will be in Shanghai at our MWC event next week. We've got a team of people going out. We are going to be running our satellite summit there. We're going to be running events on 5G and 6G. We're going to have events around eSIM, a whole range of topics.

And I think what's interesting, Peter, about the Shanghai event is, you know, as much as this is an event centered in China—and so you have to take it with a grain of salt, China's China, it's a bit different. It's not a read-across to the rest of the world. The economy and the tech industry work in different ways.

It is becoming more of a regional event as well. You know, last year, 30% of people at the event were from outside of China. I think that would be an interesting number to check in on again this year.

And I think there are a lot of interesting topical questions related to tech and geopolitics that we're going to be looking at. So any of our listeners in China, feel free to get in touch and come to any of our events there. We're very much looking forward to it.

Peter, I know yourself, you're going to be headed towards Copenhagen.

Peter Jarich (34:19)
Yeah, I'm holding down the fort here, here on the continent. So I'll be going over to Copenhagen for DTW Ignite. It's a good event. I think it wouldn't be going there if it wasn't a good event.

I think obviously, you know, the TM Forum has really managed to carve out sort of its space in terms of automation, which has led into AI.

So it's particularly important. I think there's a lot of other stuff that goes on there from a monetization standpoint. So from the GSMA, our colleague Henry Calvert will be there to talk about APIs and the work that we're doing with Open Gateway. I'll be there doing some work, presenting on some of what we're doing around sustainability.

So it is still pretty broad, but AI and, you know, AI and automation will be the big topics. I'm interested—last year, sort of data platforms took a little bit—got more visibility than maybe in the past. There were a couple of big launches. I'll be interested to see how that's followed up on.

So that will be a good couple days there.

But then, as much—and it kind of links back, as I think about sort of my story from the beginning—of how the thinking goes in sort of the tech industry versus telco industry, and maybe different ways of thinking that kind of come together, I wanna call out an event that I think we're all looking forward to in the GSMA that's coming up later in the year in September.

It's Nova. Nova is our event that's taking place in Napa, and it's really meant to, you know, do a better job of connecting with the tech industry. It's a very different thing from what the GSMA has done in the past.

So I think around 500 folks—yes, that'll be the normal set of telco folks that you would expect. Obviously, our director general Vivek will be there. We'll have Matias, the CSO at AT&T. We'll have Sunil from Senior Metal from Bharti, right? So he's the head there.

We've got Pietro, the CEO of Telecom Italia, there—always a great speaker. But they're mixing this up. We're mixing this up with a good set of folks from, you know, from other industries, from the tech industry, right?

So you've got Wing, the venture capital firm. You've got UnitedHealthcare, you've got Schneider Electric, you've got Box, you've got, you know, a good number of other folks to comment.

And what I think is interesting is that this is one of those sort of closed-door, no-media, that kind of thing. There's gonna be some keynotes, but more than the keynotes, I think I like is this kind of concept of boardrooms.

So they can have smaller discussions on a whole bunch of different topics. I think that's, for me, the best parts of any event are when you have those conversations and you get down and you actually sort of discuss things and figure them out.

Obviously, all the topics, some that you would expect at an MWC-type event, like 6G, like quantum, but going broader, right? Obviously, you know, frontier AI models.

Security and sovereignty is really big, AI infrastructure. So I think a whole bunch of important topics. Great chance to connect with the folks out there. In Napa, it's an easy trip for all the folks from the Valley.

So I'm really interested to see what comes out of this. For that, to sort of get some of the thinking going, we did a blog post for those folks for the event around AI grids.

I think we'll probably be talking about this more in the future, but it goes back to that capacity conversation of ways in which telcos can build capacity to rival some of these big claims.

Our calculation, right, is that if folks are talking about building out 100 gigawatts of AI capacity in the sky, that, you know, telcos could easily be talking about 200 gigawatts if they leverage some of their distributed capabilities.

And so I think it's definitely a space to watch and surely be talking about it in Napa in September.

So Tim, anything else before we send people on their way home?

Tim Hatt (38:42)
I wanna do the grown-up thing and congratulate you on your country's first-round win in the World Cup. We might see each other in the final. It's still mathematically possible.

Peter Jarich (38:51)
Yeah.

Tim Hatt (38:51)
And I'm not gonna rule it out.

Peter Jarich (38:53)
I just—I just have to say, you know, tying with Croatia—sorry, Bosnia. Sorry, they're all my people, so they will—they will—

Tim Hatt (39:02)
Yeah.

Peter Jarich (39:03)
forgive—they will forgive me. I mean, tying with Bosnia was good, but—but—

Tim Hatt (39:11)
I'm not saying we're—I'm gonna give—I’m gonna go one-nil U.S. on the quality of their win. I'm gonna say we save face, but we're saving our best for last.

Peter Jarich (39:15)
Yeah, yeah, yeah.

Peter Jarich (39:17)
That's gotta get—save it. That's really smart. You know, go on—go on it. You don't wanna go out too hard too early. So we will see. Hopefully we've got some good matches coming up, and it's gonna be a long summer.

So thank you all very much. Thank you, Tim, and we look forward to seeing you all next week.

Tim Hatt (39:40)
Thanks, guys.