Episode 25: Orange’s AI Super App, Nvidia’s Leadership Question and ISAC’s Defence Play
Intelligence & Noise
Telcos are looking for new AI value: Orange’s AI super app puts multiple frontier models under one operator-led platform, Nvidia’s leadership question raises wider concerns about key person risk, and ISAC is moving from trials toward defence use cases. Peter Jarich and Tim Hatt explore Orange’s free and paid tiers, the burst of ISAC activity from US grants to Telefónica and Elisa deployments, and what survey data says about costs, spectrum, privacy and the business model. They also cover voice AI for customer care, including KDDI’s work and India’s TRAI rules on human support, BT’s pause on its Starlink rollout and takeaways from the Doha event.
Hosted by Peter Jarich and Tim Hatt, every week we cover what's moving in mobile, connectivity, and emerging tech and why it matters.
Read the full transcript
Read the full transcript
Peter Jarich (00:00)
Hi everyone, and welcome to Intelligence and Noise, the GSMA Intelligence podcast that sifts through a week’s worth of news and tries to sort through the noise to bring you some intelligence. If you’ve been listening this long and you just figured out that’s where the title comes from, congrats! I was this many years old when I figured that out. We do that through the lens of the GSMA CEO newsletter.
It’s an email newsletter that goes out once a week. The GSMA Intelligence team is an integral part of putting that together. It looks at two stories that our Director General thinks are important for you to understand if you’re a CEO. We go through one of those stories, as well as one that didn’t quite make the cut, to tell you why we think maybe it should have made the cut and why it’s something you should know about. As usual, I’m here with my friend and colleague Tim. Tim heads up our... I don’t ever say what you do, Tim. Is that it?
Tim Hatt (01:06)
It’s just something out there, Peter. Something out there.
Peter Jarich (01:10)
I mean, if you know GSMA Intelligence, just search GSMA Intelligence and Tim Hatt, or GSMA Intelligence and Canadian. You’ll figure it out. Canadian Tim,
Tim Hatt (01:19)
It’s just Tim from Canada, that’s all you need to know. Yep.
Peter Jarich (01:25)
who is getting ready for Canadian Thanksgiving. When is Canadian Thanksgiving? Yeah, I thought...
Tim Hatt (01:29)
We’ve already had it. We front-run the US. We go early October, you guys go late November.
Peter Jarich (01:36)
More thankful. You were just preemptively thankful for the midterm elections. We’ll see how that works out for Canada. It probably can’t be bad.
As usual, if you’ve tuned in, you know I like to begin some of these with a bit of an anecdote, and I try to connect it to something we discuss at the end. It’s less of an anecdote and more of something I’ve touched on in the past. Some of you may recall that in October (beginning, middle, sometime in October) I was at GTC in Berlin, Nvidia’s big conference and trade show. One of the things I remarked on, across the keynote and a press and analyst session hosted by then-CEO Jensen Huang, was how polished the story is and how polished Huang is.
I think I mentioned that there was a protest from Greenpeace during the keynote, and the aplomb with which he just picked it up, rolled on and kept things moving. And just the ability of one person to talk for almost an hour and a half straight, with a story across so many different product lines, was impressive. But one of the things that struck me, and I was talking to some other analysts about it at the time, particularly in the press and analyst briefing, was that when you force one person to be the mouthpiece for all of AI, it stretches them and pushes them into places where maybe they’re not an expert but think they are. And the folksiness becomes evident, and maybe the inappropriateness of the folksiness.
When I think of the folksy way in which Jensen Huang talked about the AI industry, I very much thought of him as Uncle Jensen: “Sit down, let me wax philosophical about the industry.” But then questions come from the media, such as “How should young people feel about jobs, and what AI might be doing to jobs?” That’s not his job to figure out. He shouldn’t have to answer it. But when you feel you have to, you come up with answers like “Well, I don’t think we should worry about jobs, and young people shouldn’t worry so much.” That comment, “young people shouldn’t worry so much,” feels like the kind of thing your crazy uncle says: “Ah, young people, don’t worry, you’re young. You should be happy and enjoying your youth and not so uptight.”
I remember talking to a few of the analysts and saying: as you push someone into those spaces, and they’re so invested in the industry, they go from being Uncle Jensen to crazy Uncle Jensen. That’s part of what happens. But it also happens when you have someone who’s been the CEO of a company for so long, founder and CEO of a company that has evolved so much over time, that you have one person in place with no one either feeling capable of, or wanting to, say: “Maybe these are not the answers you need to be giving. Maybe you need to say, ‘We’re here developing the technology. Those larger philosophical questions of how people should feel about their jobs, I get it, and we’ll work through this as an industry,’” as opposed to telling people the answers to all of their questions.
Clearly no one was there telling that story. No one was doing that check. And I think that maybe says something about where we are in the here and now. That’s not to say I had any inkling of what was going to happen. I think a lot more people did if you were following the news. But it’s an interesting data point, and we’ll talk more about it later.
That said, let’s move on to the news. Tim, two news items made it into the CEO newsletter this week. One was a bunch of stuff around ISAC and the defence industry. The other was around Orange IAs, and “IAs” feels like someone getting AI wrong, like spelling AI incorrectly. The French oftentimes misspell acronyms just because of the weird way they do things. What is Orange IAs?
Tim Hatt (06:24)
Well, as you say, we’ll come on to ISAC, but Orange IAs came out of a series of product announcements that Orange made in France last week. It’s essentially a super app that gives people access to different frontier models and AI services. For our listeners, a super app is an app that bundles in a bunch of different services. WeChat in China is a good example. You see some mobile money apps working the same way. But what Orange is doing here is bringing in access to different frontier models, be they Claude, ChatGPT or a whole bunch of others, with a free tier and a separate paid tier.
The reason I think it’s interesting, Peter, is, one, that it gives customers a choice of different AI. Two, the network, at least in the premium tier, can select what [algorithm] is best to manage energy consumption. And three, it positions the operator as an agnostic platform. If you’re going to push AI monetisation on the consumer side, which is earlier days than the B2B side, this is a way to do it, to get yourself in the middle. I think it’s an interesting play. Plus, we’ve seen it in China and India, and we’ve written about that in our tokens research. Now we’re seeing it in Europe, and this is the first good example.
Peter Jarich (07:42)
Yeah, that’s completely it. As much as people talk about the token economics and the token bundling taking place in China, a big part of the solution was really just “here’s access to a whole bunch of models that you can spend those tokens on.” It’s like selling video game tokens without access to the video games. What am I going to spend these tokens on? If you give them the models, they can do that. So yes, it’s interesting. And for those folks out there, in France AI is “IA.” So in case you think I’m just becoming crazy Uncle Peter and talking about acronyms, I am not.
ISAC. I think it’s interesting because I’ve had some interesting conversations around ISAC, and there are a whole bunch of people on both sides of the fence: is this going to be a great opportunity? Is it hype? Is it 6G looking for a solution? I’ve had a few back-and-forth, probably ill-informed LinkedIn conversations with people saying that industry organisations like CTIA and GSMA are hyping things like ISAC, and I have to remind them I don’t think we’ve said much about ISAC at all. If you’re looking for who’s hyping it, feel free to look to the rest of the industry. And every now and then the industry has my back and proves me right.
Last week was a good one. I think four things took place, so I’m going to go in order. On November 5th, ETSI came out with a new report on the importance of bundling ISAC with edge compute. Interesting, and it makes sense: to enable autonomous operations and distributed intelligence, you need a framework for integrating computing. Then... oh, I take that back, it was the third. Two days before that, on November 3rd, the US announced $50 million-plus in grants for developing ISAC use cases, across 11 teams. On November 10th, Telefónica talked about a private 5G deployment for NATO, which included, amongst other things, ISAC. And on the 11th, Elisa announced a joint venture to develop drone detection privately. A whole bunch of stuff at once. Tim, why do you think we see all of this at once as we go into the end of the year?
Tim Hatt (10:26)
Well, I think for a couple of reasons, Peter. Number one, particularly for the European stories, there’s a clear link with defence, and that’s a topical issue for many governments right now. It’s interesting to see mobile operators putting network infrastructure and ISAC use cases in there to support it. It’s no coincidence, Peter, that the examples are Telefónica in Slovakia and Elisa in Estonia and Finland. Those are countries very near the Eastern Front of Europe, near Russia. That’s where the tensions are.
I also think it’s interesting that these are examples of operators solving real-world problems. To your question of whether this is technology looking for a use case, yeah, kind of, like everything is in a way, but this isn’t “let’s wait till 6G.” This is being deployed now as a trial. When we zoom back to look at where 6G goes, it’s an interesting proof point of how we get there. What do you think?
Peter Jarich (11:35)
No, I think extending that, right? I mentioned the first thing was coming out of the US, funding 11 different projects with a pretty broad set of players. You’ve got some big players that we know: Nokia, Ericsson. You’ve got some smaller network folks, like Parallel Wireless. You’ve got some folks who are more on the ISAC side of things, like Tiami, and then a whole spate of research universities: Northeastern, Arizona State, Michigan State, all over the place. You see San Diego, which probably puts it close to Qualcomm, which is great. Virginia Tech tends to do a lot in the telecom space. I think this goes to your point: let’s look at these use cases and see what’s possible.
I can tell you about some of the conversations I’ve had with folks who have questions about it. Part of it is the business model, and we’ll get to that. Part of it is trying to understand the technical side. I’ve had tons of conversations that go, “Look at what’s taking place right now. Look at some of the demos from earlier in the year that we talked about. Clearly, we can do some tracking and positioning with 5G networks in the here and now.” That’s great. Then you’ve got some folks going, “No, no, no, we need tons of spectrum and millimetre wave.”
And you look at what’s taking place in China, where the government’s view seems to be that positioning and sensing are important but you can’t really do it unless you have tons of spectrum. They’re going to use millimetre wave, which, by the way, they haven’t released for comms yet. So you’ve got that side of things, and then the question of how much of a hit it takes to your capacity. Does this mean I lose 10% of network capacity because I’m doing sensing along with comms? Is it 25%? I’ve seen numbers all over the place and I don’t believe any of them. [I’m sure some folks are being disingenuous, but I don’t know who they are, so I’m not singling anyone out. I just think it’s still early days and we don’t know.]
And if you add in things like sensor fusion, ISAC combined with other types of sensors, then what does that mean? All of this, for me, is about the experimentation to figure out what it looks like in the long run and how it could work, which I have a hard time faulting.
Tim Hatt (14:17)
Absolutely. And we’ve both heard conversations, haven’t we, through talking to carriers over the last few weeks and others. I was just in Doha talking with representatives from the US government about how ISAC looks in practice. I know Ericsson is one of the leaders here and very vocal on it. But a big point is that 6G has got to be a network that is agile and senses. There’s a defence application, there are biodiversity applications, urban planning applications. The key point, Peter, is that we’re seeing it deployed now to front-run 6G, so you can test it ahead of the network architecture coming through.
One other point, Peter. We often talk about technologies and their evolution; let’s put the monetisation angle out there. In Europe, excluding the US, which is the biggest military spender, defence spending in 2025 was, I think, around €450 billion, at least among EU countries, between 2% and 2.5% of GDP. Now, if you take, hypothetically, a 5% slice of that, you’re getting into the tens of billions. European mobile service revenue is [€]150 billion, so you’re coming up on something like a 5% area. I’m not saying all of that is [accessible], but it’s not small money. And that’s why we talk about defence as an opportunity.
Peter Jarich (15:46)
Yeah, defence is an opportunity, and we clearly know where we are in 2026. If it’s getting discussed in the Nordics, I’d imagine conversations in Doha are just as heated, just as interesting, just as much of a priority. What I think is interesting, though, comes from our latest network transformation study. We’re going to put the dashboard out before the report, because the report is sitting... yeah, I’m responsible for the report, which means it’s on a to-do list well below a whole bunch of things. But the data’s in, and I’ve been sitting on it for a while. We dug into ISAC with a couple of questions in a couple of places in the study.
What I think is interesting is when we ask operators, and these aren’t just a bunch of randos, these are telco network decision makers distributed with quotas across the world, so we get fairly representative results from people who know what they’re talking about, what the biggest use case is, defence comes down towards the bottom. They’re looking at industrial, logistics, transport, mobility, smart cities. So I think telcos are seeing opportunities outside defence. Which means, to your point, if defence can be a decent-sized bump on their services, maybe there are other service revenue opportunities.
But it also points to some things that I hope this new set of deployments and trials can solve. One of the biggest concerns is high deployment costs. How much is it going to cost me? Is this new spectrum? Is this new RAN? Do I have to add new sensors? Then there are things like privacy and security. For me, that called out all the things that weren’t necessarily addressed in any of these announcements. Maybe it’s not their place, but hopefully we’ll figure out the spectrum implications: what spectrum works, what the limitations of different bands are, what it means for capacity planning, data ownership, privacy, all of that. I’ve talked in the past about how there’s lots of work on how we need to be thinking about this. And then, on the back end, cybersecurity.
I was talking to a customer on the networking side just last week who said some of this is basic hygiene in terms of data. If this is highly critical data, we need to be doing all the things we normally do to protect our data. Then interoperability across different suppliers, and mostly the economics, as you called out. What’s the operational side of things? I’ve said this before and I’ll say it again: I’ve had conversations that were so inspiring with some real ISAC innovators, who say, “Here are all the cool things we can do.” And I say, “Well, how do we make money?” And they say, “Man, don’t bog us down on that. You’re going to stop the innovation if you think like that. You can’t think too narrowly.” And I’m like, yeah, but if you’re not making money for operators, you’re going to have a hard time selling it to them, and “build it and they will come” isn’t really a great business model. So I think there’s a lot we still need answers on, but it’s good to see.
Tim Hatt (19:16)
Yeah, absolutely. We will follow it closely.
Peter Jarich (19:18)
Let’s talk about what didn’t make it into the CEO newsletter. Could have, maybe should have, not our decision to make. There were two stories that didn’t get in. The one we’re going to talk about, well, let’s get to that in a second. The other: Australia is working to tighten some AI safety rules and has put out a plan to establish an AI regulator.
I think it’s timely. There’s the question of AI regulation, whether you’ve got frontier model makers saying “someone regulate us,” or people thinking that a new regime in the US might actually put limits on some of those model makers and how that impacts them. You’ve got Australia, which likes to be first on some of these thorny issues. Maybe you can do it when you’re a smaller country. Maybe it’s the goodwill of all the kangaroos. I don’t know, it’s the quokkas, I think.
Tim Hatt (20:23)
Ha ha ha.
Peter Jarich (20:24)
It’s the impact of the quokkas. Who will think of the quokkas? I mean, Australia, you’ll remember, was the first country to put in a nationwide ban preventing under-16s from accessing social media. So they do have a role in this. Again, it’s probably best that it didn’t make it into the newsletter. It’s part of a broader theme. The other story that didn’t make it in is also somewhat AI-related, and we’re going to talk about it: voice AI for customer care. Two stories here, Tim. What were they?
Tim Hatt (21:03)
Yeah, two stories, as you say. The first is out of Japan: KDDI, the operator, working with a partner by the name of Altius Link. And secondly, the Indian telecom regulator, TRAI for short, introduced rules permitting operators to use AI-based agents for their contact centres 24/7, but with specified times for human oversight.
Peter Jarich (21:31)
There has to be 16 hours of daily human support.
Tim Hatt (21:35)
That’s right. So if you call at two AM, you might be dealing with just AI. If you call during daylight working hours, you’re okay. But I think it’s interesting, Peter, and I’m interested to get your take here. I think people often forget about calls because nowadays they’re generally unlimited. People try to do a lot of things online. But when you look at the numbers, call volumes to call centres are still quite high for troubleshooting. And the reality is it’s still a very high cost base. So this has been one of the first areas where we’ve seen a lot of telcos deploy AI. About 45% of the deployments, according to our tracker, cumulatively through Q3, have gone into customer care. So it’s an obvious place to look. The question, and I’d like your take on this, is: does it work? That may seem like a basic question, but it’s a key one.
Peter Jarich (22:30)
What I think is interesting here is that across the things we’ve talked about, there are two threads. With Orange IAs, we’re talking about an opportunity to make money from AI. And here we’re talking about applying AI to save money and/or provide better service. On the one hand, it completely aligns with what we’re seeing. For those of you wondering why Tim keeps throwing out data, and why every now and then it spurs me to call out some data: this is all research we’d love for you to take advantage of. If any of this seems interesting, and you’re a customer, check it out on the site. If you’re not, reach out to us.
Going back to that network transformation work (you can tell I’m finally getting around to actually paying attention to it): when we ask telcos, big picture, not down to specific ISAC use cases, not down to specific instances of how you’re deploying AI agents or what’s holding you back on SA, all things we ask, but what’s your goal with AI? What’s the driving business goal? Saving on OpEx comes out on top, followed by improving customer experience, by a large margin over supporting new services or saving on CapEx. So this definitely aligns. You’ve got both of those things: saving on OpEx, so AI rather than people and systems rather than people, and improving the customer experience.
It also aligns on that third part, that the focus isn’t on supporting new services. So 100% seeing this align, and a little disappointing, because I do think there’s such a huge opportunity on the telco side to be part of the “how do I generate revenue” side rather than just “how do I save money.” But it totally aligns with what we’ve seen from other operators in how they’re looking at AI, so it’s not too surprising. In conversations with some of our folks on the events side, they’ve said they’ve seen a lot of interest at MWC coming up in sponsorship from folks on the voice side. So I think we’re just going to see voice, whether as a consumer application or as customer care. I think there are some other big-picture stories here, but does that make sense to you? Am I just becoming...
Tim Hatt (25:16)
Yeah, I...
Peter Jarich (25:17)
crazy Uncle Peter?
Tim Hatt (25:18)
No, no, [not crazy at all]. I think you’re talking sense. Going big picture, I think this is the direction of travel, just like electric cars are the direction of travel, even if not many people have an electric car now. I think we’re going to get to a point where most calls are automated. My question on the big picture, and whether it’s good, is this: when I call banks, insurance companies and airlines, and get through the automated system, I’ve got to tell you, it’s still coming up wanting. It’s a little like, if you’ve seen Seinfeld, the great cult show in the US, a great episode where Kramer poses as a movie booking agent and mishears the audio, and answers when somebody is saying something completely different. Frivolous example.
The same applies here. I think there’s a reason KDDI is going into so much detail on what they’re doing with the AI, and that’s really important. Otherwise you only get one chance, and if you turn people off, they’re going to churn, or they’re going to rebel.
Peter Jarich (26:24)
I think people recognise there’s an opportunity, and it’s one of those things where there’s a huge opportunity here if you get it right. Maybe to date we haven’t quite gotten it right. You could say, what’s new here? Haven’t we been doing this for a while? Haven’t we had IVR, haven’t we had voice prompts? It just hasn’t been great, which has turned people off. How many times have we all just hit zero, hit zero, hit zero: “I just need to talk to a person.”
Tim Hatt (26:53)
Exactly.
Peter Jarich (26:53)
How do we get to a point where it makes sense? For me, the big story, again, you’re right, is the direction of travel: AI is going to transform how we talk about customer care. The interesting part, and how I think some of this links together, is that whether it’s this or the Australian side of things, regulation is going to play a role. Whether it’s “you have to have a person available at this time”...
Tim Hatt (27:10)
Yep. Sure. Yep.
Peter Jarich (27:13)
...or, as in another part of the work going on in India, setting out the fines and the expectations for how quickly people will have their issues solved. So I think we’re going to see that connection between regulation and, to some extent, trust: if I know this stuff is getting dealt with, maybe I’m going to trust it more. If not, that goes all the way back to the uptake question you pointed out. And ultimately, we know the cost savings are the important part.
What I’m not completely sure about, and I’ve mentioned this before from conversations with some operators, is the cost of rolling out AI systems that can do this at scale, with the quality that’s needed. That means developing the AI infrastructure and the investment in it. Are we there yet? How good is good enough in terms of a model to deliver good quality for customers? That’s going to determine how quickly you roll out and where. Maybe in some markets it makes more sense, and in others it makes more sense to put people at it, which is good for people who need jobs.
Last, the Willie Nelson memorial segment: On the Road, On Our Minds. You were in Doha, and I think you talked a bit about it last week. Any last thoughts from Doha itself? I think you were still there for another day. Did anything exciting happen?
Tim Hatt (28:52)
Well, it was a solid event overall. As you said, it was the first time co-hosted with the ITU Plenipotentiary, which brought a more varied audience: policy makers as well as industry professionals. But fundamentally, Peter, what was interesting is that the Middle East sits as an overlap between the US, Europe and China, so you get both viewpoints. AI is moving fast, there’s lots on satellite, and still a lot around the usage gap in North Africa. So, very successful discussions.
Peter Jarich (29:26)
Cool. I’ve got one story, you’ve got one story. Let’s talk about crazy Uncle Jensen, like we said we would at the beginning.
Tim Hatt (29:31)
Yep.
Peter Jarich (29:32)
What struck me last week is that, on the one hand, obviously we’re all talking about this because it’s a bit of hot gossip, someone high-profile getting caught. It’s like the week or so when everyone was focused on the kiss cam, and you recognise you shouldn’t be having an affair and go on the kiss cam if you can avoid it. But it was a lot more than that, because it’s the industry. It’s one person who’s the face of the industry. We’ve seen more over time. Obviously there are questions of geopolitics.
For me the bigger question is: if you’ve got a company saying “we knew about some of this, it’s fine, business as usual,” and yet you hadn’t disclosed it to shareholders, what does that mean for your reputation and your legal exposure to shareholders who might say, “If you knew about this, you probably should have told the people who are invested in you”? So we’re just seeing a lot more. Now we’ve got calls for him to maybe... so yes, hopefully you know by now he’s resigned. There are calls for him to potentially come back, and discussions that he may consider coming back. There are just so many layers to this. To some extent you want to throw up your hands and go, “Don’t even think about it, there’s too much going on.”
Tim Hatt (31:17)
Yeah, absolutely. And things move very fast, don’t they? There’s rearguard action to get him reinstated, a lobby for that. There’s the US government, which is a lot more circumspect. But I think the key thing for me, Peter, is this notion of key man risk. So much of Nvidia’s brand, IP, gravitas, everything is tied up in one guy. And now they’re thinking, “Man, what do we do?” The guy had apparently 60 direct reports. I don’t doubt he’s a genius, but the question now, and I think this is why there’s such a vacuum when looking at future leadership, is how do you fill that? We see it with Meta and Mark Zuckerberg, and obviously with SpaceX and Elon Musk and his other companies. Another example.
Peter Jarich (32:07)
Yeah, I think there are two interesting parts there, and I wonder about this, and we’ve had this discussion before. If you look at SpaceX and Meta, you’ve got Zuckerberg and Musk with a big chunk of the controlling votes. From a share standpoint, I think it’s something like 3% for Huang, which is not nothing for a huge company. I don’t know where that sits in terms of voting, but I think there’s a bigger question here.
Take something like SpaceX. You’re going to want Musk in charge because, to some extent, it’s still a growth company. They’re growing out the AI side of things. You don’t know what it’s operating as, pushing into spaces. Think about Anthropic: we still don’t know what it’s going to be when it grows up. It’s pushing very fast. Is Nvidia that same kind of company, in a space like Anthropic, building out new businesses and markets that haven’t existed before? Or is it an Apple, which says, “Okay, I developed an iPhone, I’m going to develop a new iPhone, I’m going to iterate marginally on the next thing within a fairly well-known market,” and even when new markets come around, like the AI side of things, says maybe we’ll just work with other folks? What kind of company is it?
Because on the one hand, if it’s that sort of SpaceX/Anthropic, then yeah, you need that guy in there to guide the market. We have a visionary leader here. If it’s not, then you need a CFO, one who says, “This is a predictable, stable market. We churn out more chips and develop new chips at a...” And maybe we pull back on the pace.
Tim Hatt (34:13)
Mm-hmm.
Peter Jarich (34:13)
Maybe we pull back on the pace and say these things are selling just fine as they are, we don’t need to constantly make them better. I saw some interesting reporting mentioning that you’ve invested in a whole bunch of companies that are buying your chips. Every time you roll out a new generation of chip, you’ve just [depreciated] your investment, because these people have bought a whole bunch of the old one and you roll out something new. Maybe someone coming in from more of an operational standpoint says, “Okay, maybe we think about this business differently.” That’s the interesting part for me: what kind of company are they? Again,
Tim Hatt (34:54)
Yeah. Yeah.
Peter Jarich (34:56)
there’s so much to talk about. That said, speaking of SpaceX, do you want to talk a bit about SpaceX and Starlink, and why it may not be coming to us here in the UK anytime soon?
Tim Hatt (35:04)
Yeah, well...
Peter Jarich (35:05)
...
Tim Hatt (35:12)
Flipping to this side of the pond. Yes, BT in the UK made the announcement last week that they’re going to put the brakes on their rollout of Starlink broadband. And the story that we are not...
Peter Jarich (35:27)
Was it an announcement, or was it just a comment from on high?
Tim Hatt (35:31)
It was a comment. It was a comment from the CEO, Allison Kirkby, who’s a very well-respected and very successful CEO in the industry. But it’s not so much the news that’s interesting, it’s the why behind it. The why is “we want to let the noise die down.” I’m paraphrasing the CEO: the noise around Elon Musk and Ukraine. We think there’s reputational blowback here that we don’t want any part of as an organisation if these types of concerns continue.
One last point on this, and I’d like your reaction: the other reason I think it’s interesting is that Allison is speaking for a larger silent minority. Having had conversations myself with a number of governments and other operators, a lot of people are jittery about this. She’s got the guts to come out and say it and be honest about it, but she’s far from the only one who feels that.
Peter Jarich (36:30)
It wouldn’t have been big news if it was just... if it was crazy Aunt Allison, keeping up the theme here. It wasn’t someone who’s an outlier expressing a fringe view. I think it was well known that a lot of people are saying it. I’ve been saying this, and I’ll continue saying it: when Ukraine was having its issues, and that was impacting their drone detection capabilities and their drones, there were a ton of people cheering. Not cheering because it was bad for Ukraine, but because it was exactly the message they’d been conveying around sovereignty and how much you can trust some of these systems. It may have seemed academic until it wasn’t. So what you heard from Allison is definitely speaking for a whole bunch of folks, which I guess is the role of the middle powers, right? That’s the...
Tim Hatt (37:27)
That’s the role of middle powers. This is what Carney had in mind.
Peter Jarich (37:29)
Speaking for the people who are afraid to say it, or maybe just haven’t gotten around to saying it yet. My other On Our Minds story falls in the same vein. It was an interesting one, easy to miss. A consortium was announced out of Japan: Mitsubishi Electric, MUFG Bank, Development Bank of Japan, NTT Docomo Business and NTT, to look at developing mega constellations. It’s an MOU to investigate the development of a mega satellite constellation, and just an MOU doesn’t necessarily mean much right now. Each has a role: the system design folks, Mitsubishi Electric; the banks, on how to finance this and what it looks like; NTT Docomo Business to look at what the actual use cases are; and NTT on the networking and integration side. Cool.
But for me, the bigger story is that it’s [2027], and we’ve got Starlink with a huge network constellation and huge plans to put more satellites into the sky. You’ve got AST, Amazon and lots of other folks who are essentially incumbents already. And you’ve still got people investigating the development of new constellations. So it’s far from done. We kind of think about this right now as “this is what it’s going to look like, given the number of satellites in the sky from Starlink and the others.”
Tim Hatt (38:59)
What if you’re...
Peter Jarich (38:59)
There’s still a lot up for grabs. There are still a lot of people looking at this and looking to see what the business looks like, potentially with more specialised use cases, potentially for sovereign functions. It’s still early days, if you will, and that’s great.
Tim Hatt (39:26)
Yeah, it’s a long road.
Peter Jarich (39:28)
A long road. With that said, thank you all for making it to the end. I appreciate all of you listeners. I’ve said it before: please, please, please, if you’re enjoying this, go rate us, provide some feedback, provide some comments. Tell us we’re dumb. Tell us we’re not dumb. Tell us what you want to hear about. We’d really appreciate that. In the meantime, we’ll see you next week. Tim, we’ll see you next week.
Tim Hatt (40:01)
We’ll see you then.
Peter Jarich (40:02)
Wonderful. In the meantime, I’d like to remind everyone, when you’re out there searching for intelligence, please, please, please be sure to embrace a little noise and make us part of that. Thank you all very much. Talk to you next week.
Tim Hatt (40:18)
Thanks, everybod
Co-hosts
- 200 reports a year
- 50 million data points
- Over 350 metrics
How can we support you?
Get in touch
Contact the GSMA Intelligence support team for help with your account, subscriptions, or access to reports and insights.
Newsletter
Subscribe to the GSMA Intelligence newsletter for the latest industry news and insights, delivered to your inbox.

