How SGP.32 is powering the shift to device-first IoT operations

In this episode of the IoT Now Trending Tech podcast, host George Malim, managing editor at IoT Now, is joined by KORE’s Mandeep Saini, director of product management, to explore how SGP.32, eSIM and remote SIM provisioning are changing the way enterprises connect and manage large IoT fleets.

From moving beyond carrier lock-in and enabling frictionless connectivity orchestration to creating a true single pane of glass, they discuss why IoT management is shifting from SIM-centric to device-first operations. They also examine the need for governance, the evolving role of mobile operators and satellite connectivity, and why strong partnerships, common standards and device certification remain essential.

Looking ahead, George and Mandeep explore how software-defined connectivity, APIs and agentic AI could make IoT operations increasingly autonomous. They explain why connected devices are becoming continuously evolving pieces of software and what enterprises must do to manage changing profiles, detect abnormal behaviour and operate deployments reliably at scale.

Transcript

​[00:00:00] George Malim: Hello, I'm George Malim, the managing editor of IoT Now, and welcome to the IoT Now podcast. I'm delighted to be joined by our guest today, Mandeep Saini, the director of product management at KORE. We're gonna be having a long discussion on how the world of IoT connectivity is developing and the latest innovations.

But before we get into that, let's look at what's been happening recently in the news. We've both selected a couple of stories that caught our eye. So I'll turn to Mandeep first and say, what's caught your attention, Mandeep? 

Mandeep Saini: I'm very much interested in, uh... There was a story that Transforma released about, uh, they were forecasting that AIoT or essentially AI on IoT, the connections were gonna be growing from about 1.8 billion at the end of 2024 to over 11 billion by 2035. So almost 10X. I find that interesting, you know, from our perspective, and what interests me is that what AI at the edge actually is gonna do for the [00:01:00] operations. If you look back, um, traditionally when you deploy a sensor, typically, like, it's deployed once and you leave it alone. You generally don't really touch it, aside from a couple of firmware updates here and there, but it's generally, like, hands-off. But now if you've got AI on the edge and AI device enabled, it's like a living piece of software in the field that is always changing. Like, if you're gonna using different models, you're probably gonna have a lot more firmware updates. So it's constantly changing. So the, how customers are gonna operate the AIoT at the edge is gonna change.

So when it comes to KORE, what I wanna be clear about or where we sit is that we're not the ones that are actually gonna run the OTA campaigns of, like, updating firmware or the OS on the software. That's the customer's domain. Our job is to make sure that when that does happen, it works, at least from the connectivity side. Making sure the device is always online, has enough coverage, and the window that the customers want has as high avail- uptime as possible. So that's where we really fit in and something that I wanna touch [00:02:00] on that I think a lot of customers or people just don't know about is connectivity is actually one of the first places that we can detect a bad actor or a bad device. So by the time, let's say, the customers is... If they pushed an update or the AI is going rogue or something, they may not know about it until later or when their invoice comes in, 'cause then maybe the usage went up. But we should be able to capture that very quickly and proactively tell you, like, something's about to go bad, right?

We could compare against, like, past updates and stuff. So it gives us a lot of opportunity to help the customers ensure that their deployments are working normally after every update. So I think, you know, from the customer side, it gives them a lot of possibilities, but it opens up opportunities for us to help the customers actually achieve their results in a much more scalable way 

George Malim: So that's really interesting and, you know, to see that kind of 10X increase, I think just demonstrates the pace at which things are moving.

Speaking of the pace at which things are moving, I've been interested in an automotive IoT story from analyst firm Omdia. I've been tracking a couple of different [00:03:00] pieces of their research. One that came out slightly earlier in the year was that automotive data traffic is set to rise from 30.7 exabytes last year to 135.4 exabytes in 2035. Now, that seems absolutely enormous in terms of the increase. Um, and, you know, whilst the firm initially said it's likely to be driven by integration of infotainment systems in newer vehicles and rising consumer adoption, many of which will depend on 5G connectivity, I think there's been a perhaps better explanation by more recent research from the firm which says that the automotive vertical itself is experiencing rapid growth with connections forecast to exceed one billion by 2035. So it's a billion connections doing 135 exabytes, and that seems still, you know, a big chunk of additional data traffic. But, the reason behind it being huge numbers of vehicles, so I think that's interesting. I also think it's interesting that Omdia singled out key three trends. First of which is 5G dominance. It reckons 89% of [00:04:00] automotive modules will use 5G by 2035. I guess that's down to the fact that 5G's clearly gonna have the longest lifespan of most of the technologies available today, and given the life cycle of a vehicle, that carries through. Obviously it's great to see more connected vehicles, and I think that paves the path towards V2X communications, which is vehicles everything, and that'll be driving advanced connectivity, um, requirements. Don't have anything to say on automated and assisted driving. I think that's probably delayed and needs a lot of regulatory work to go beyond the kind of basic use cases we see at the moment, so I don't think that's happening. But I can see that automotive's clearly leading the charge in terms of use cases.

Cool. Right, let's move on to our main discussion. We've got several topics to cover. The first one is the evolution of connectivity management, and we're looking at how having a made-for-IoT specification in the form of SGP.32 is making it easier, um, and to connect and manage large fleets of [00:05:00] devices.

So, Mandeep, what's the realistic target over the next few years, and what benefits does KORE see from frictionless access and orchestration of connectivity? 

Mandeep Saini: Yeah, great question. So I'm gonna actually touch on something I mentioned earlier, and, you know, I think what SGP.32 really helps is, um, you know, if I give an example, right? In the old world, if you deployed something, you're typically locked into whatever provider you have. So you have one decision, one shot to get it right, and if you get it wrong, then you gotta do a truck roll or go through some really difficult migration work. Which, you know, it was very difficult and no one wanted to do it.

So as I mentioned earlier, like SGP.32, the devices and the SIMs itself are really becoming like a piece of software that's running on the device. So these IoT devices, you can configure them and change them as they go.

And then you were talking about cars, so I'm gonna bring that back in. The best analogy I would say is, you know, let's say I have a car that's like 16 years old, has no smart features. It is very dumb. So if there's any updates that I [00:06:00] require, I'm not gonna get any, like say the dashboard or anything. The newer cars now, they have the flexibility. Like if they get, let's say, something on the dash wrong or they wanna change some software, now they have that capability 'cause their cars are now connected to the internet. So now you can... let's say there's a bug or a feature update, or they wanna add some other things to the dashboards, that they can now do that.

So it's a constant living thing that's always changing. So again, like you know, that's how SGP.32 is gonna transform operations. And what you deploy on day one isn't necessarily what you end up in a year or two, right? 'Cause we can always learn and adapt. So for KORE and timing, I think customers, if they aren't already, they should start thinking about how to move to SGP.32.

And even if they don't have a solution that will take advantage of what the standards provide, you know, that's totally okay. But if you start now, then by the time, let's say, your use cases do change or you have to make a different decision, you should already be prepared to make those updates versus, let's [00:07:00] say, you continue down the path of being static, being locked in because that's what you've always been doing. But let's say in two years, three years, when 32 becomes a standard, you're stuck and then you gotta do a truck roll to go and update it. So I do think this year we're gonna start seeing some pilots and some testing of it, and maybe even some rollouts. But next year I do definitely think it's gonna start to ramp up 

George Malim: Yeah, I think it's right on the cusp of becoming, like, the mainstream activity.

Yeah. It's really interesting. Um, but it creates a challenge actually, because connectivity management has traditionally been carrier centric and, you know, you'd have the automakers doing deals either with big global connectivity management platforms, but more typically with global carriers and big brands like that.

What does the new ecosystem look like as carriers effectively lose their lock on customers, but at the same time they also get the opportunity to access an expanded marketplace? 

Mandeep Saini: Yeah. So I wanna start by saying, like, the carriers to us, like, all these tier one carriers around the world are very critical to the success of KORE, and we do anything we can to [00:08:00] ensure that there's success on all fronts, right? And so I would frame it less as carriers losing or losing their lock-in. I think it's more, um, the unit of competition is gonna be changing. And as you mentioned, the TAM is also gonna increase.

So, let me touch on those separately. So the lock-in used to be s- you know, like, you're fixed. So like, I'll use an American carrier. So let's say you were on AT&T, there's no way T-Mobile, Verizon or anyone else could come in and take your business even if they were better in the area that you were in or offer services that you had.

So the TAM was very small. Like, they couldn't capture existing customers unless they did, like, a truck roll or a campaign to switch. But so what SGP.32 unlocks is that now the TAM increased. So devices that were before unreachable or unattainable by a tier one carrier, now they can, right? So the retention now is really on coverage, price and quality in my mind.

So, if what you have today meets all three, but later on, let's say a [00:09:00] competitor releases a new update or they upgrade the towers or something else, I think that really um, increases competition for everyone in the ecosystem. So I, think the ecosystem is generally gonna be bigger for everyone, and frankly, like I mentioned, our success is reliant on the carrier's success as well, and we can't achieve the success without the carriers, and with the customer use cases too, like, not everyone is gonna want a roaming solution. So in order to offer localised solutions, we definitely have to partner with the carrier. So they're never out of the loop. They're always part of the value chain.

Something that I'll bring in that you didn't kind of mention but I think is really important to mention here is that we talk about carrier lock-in and competition. Something that I foresee happening is satellite, especially companies like Starlink, they're creating a new dimension of how customers will build a solution. Now, they're not necessarily reliant on just cellular. They can also add backup or satellite. So I kinda summarise that, it kind of is increasing the competition, so these different players now have to compete on different things beyond just, you know, price. Because [00:10:00] now a customer can easily flip or switch profiles and redesign their solution to meet their needs whatever that may be 

George Malim: Yeah, I think it's an interesting one because it is about meeting needs and being able to do that in a more flexible way than what's gone before.

But I think some people got carried away with marketing a couple of years ago and suggested that, you know, kind of eSIM was gonna be like, you know, you're gonna change when, in your scenario that you were saying with Verizon and AT&T. You know, If Verizon's offering slightly better performance in a given location, you're gonna instantly switch to that, uh, you know, switch across and then switch back when you're in a different area and you're gonna have these numerous changes happening all the time. And I think it was kind of a dead end to suggest that operators of 300,000 devices really want that level of continuous change. Do you think that's, you know, the reality is yes, there's more flexibility, but no one's gonna go crazy and- 

Mandeep Saini: Yeah

George Malim: continue to change. 

Mandeep Saini: I agree there's more flexibility- Yeah ... and I would strongly discourage any customer who would say, "Oh, I'm gonna flip my half a [00:11:00] million devices hourly or daily or whatever it is." Because there's always risk, right? And we do our best to ensure that everything happens correctly and without issue, but there's always chances, right? Like, if your device is moving, it could lose coverage, right? So, I would say that every change is something that customers have to plan and monitor and verify, and as I mentioned, there's always gonna be risk, right? So how I'd re- reframe it is, yes, there's flexibility, but what customers need to have is governance, right? A policy that defines what should happen under what conditions so the switching or changing is minimised and only happens when it needs to, and the way I can actually talk about this is so let's say, um, talking about the risk, right? So let's say you have a medical device in the hospital that you're monitoring, let's say, you know, a pacemaker or something, uh, that a patient has.

So if you're rolling these updates constantly and if one of those fails, then imagine what that means, right? So if that patient has a medical emergency and it's in the middle of an [00:12:00] update when, you know, it shouldn't have been done 'cause there was no need for it, there's an inherent risk. So just because you can do it doesn't necessarily mean you should do it.

George Malim: Okay, moving on from that though, what is actually possible and practical now with eSIM and remote SIM provisioning? 

Mandeep Saini: Yeah. So, I'll be honest, like a lot of what SGP.32 provides was technically possible before, right? With the earlier standards like 02 and 22. You could buy a global SKU, it could be a roaming SKU and either through, depending on your product, either using 02 or 22 standards change the profiles, or you can buy an eSIM that has two multi-profiles and have your devices switch. But all these options had compromises or some asterisks about, you know, dos and don'ts, and there were some always limitations. So there's always risks. And what 32 really makes possible is that it kind of addresses a lot of the pain points that 02 and 22 had. Like 22, you had to have a device, a human had to do it. With 02 there were some limitations, there was more lock-in, and 32 kind of addresses all that. So, the end result is still gonna be the [00:13:00] same with 32, but what 32 really makes possible now is, like as you mentioned, is more flexibility, but it makes it a little bit easier and a little bit less, or less friction for the customers to perform the operations that they need to perform.

George Malim: Yeah, that makes really good sense. One term that I've been seeing this year, probably since MWC in Barcelona in the spring, is the SPOG or single pane of glass. How does remote SIM provisioning contribute to a true SPOG offering? 

Mandeep Saini: Yeah. So for us, we've been in the business of a SPoG or single pane of glass for as long as I can remember. So it's nothing new for us. I would think that, you know, SGP.32 and eSIMs and RSPs and all that is just a continuation of what we're doing. We're not changing direction. We're just adopting new standards and adapting to customers' changing use cases. So our job is to make sure that whatever standards that we use, whatever solution the customer have, how do we make it simple? And the goal we have is, whether it's a SPoG or otherwise, is how do we abstract the complexities away? So whether you've got a global deployment with dozens of carriers [00:14:00] or you're using MZ-based products or whatever it is, that we make it very, very simple to that. But what SGP.32 does provide us is that it gives KORE the ability to go deeper into the stack of the solution, right down to the eSIM, to be able to, you know, talk to the eSIM, tell it to switch, and the eSIM can talk back to us. So there are definitely some changes, but to the customer's perspective, it should be not too different. If anything, it should be easier. 

George Malim: That's really interesting. How important is a strong ecosystem to enable the benefits of RSP and eSIM? It strikes me that you need lots of different experts collaborating to actually deliver on the promise.

Mandeep Saini: Oh, it's absolutely critical, right? So we could have definitely gone down the path of done everything ourselves. But, you know, in order to have success releasing this, you have to work with strong partners. So we do have partners for our SIMs, our applets. We have our carrier partners. We have partners to, you know, build these SIMs, right? So, I think the key to success is always focusing on what you can do really well and tapping on the shoulder of [00:15:00] others or partners that can help you fill that complete solution rather than you building it yourself. And the other part too is that with this ecosystem is the standards, right? So making sure that we're all following the same standard. Like, if you've got a device that's not following the standards, it's nothing more than just a document, and it may or may not work. So we do have experts at KORE that know the standards, and we work with customers to make sure that if they're gonna adopt these new technologies, that what does the standards require? We do have services to help certify devices as well. So I think certification will also become very important. And again, we do some of that, but we also have partners around the world to help us and the customers make sure their devices are certified and will work on these networks. 

George Malim: That's really interesting.

So just to sort of close the loop on SGP.32, do you see it as just another connectivity standard or does it fundamentally change operations? 

Mandeep Saini: I wouldn't say it's either, either/or because I think standards are gonna evolve all the time. And I'll use an example. Like, you're gonna see this all the time, right?

So with [00:16:00] Wi-Fi, what we had with Wi-Fi, let's say, 10 years ago, it's very different, right, than it is now. And it's not because there's a regulatory body or someone that's pushing standards for the sake of it. It's because the needs are changing of the consumers Wi-Fi case. And the same thing goes for connectivity, right? So let's say, you know, 15, 20 years ago, the bandwidth or the use cases were very different. Now, it's changing, right? Customers want flexibility. They have higher bandwidth use cases like your vehicle example earlier on.

And another example I can give you is, let's say, you know, smart home products. In the past, there were all these different technologies and standards for communication, and now, the consumers pushed the companies to come together and create a new standard called Matter.

So, yeah, I think a lot of these standards, I wouldn't say it's being pushed by anyone. It's really just they're naturally coming up as the needs of the customers change, and I honestly think it's gonna change again, right? Again, as use cases change, as we learn more about what's possible, what's not, [00:17:00] you know, so maybe in the next couple years we'll see another standard come out as well 

George Malim: Yeah, that sounds really likely the way things will go .

Let's move on to, another topic which is still looking at the single pane of glass and, you know, trying to draw on your expertise actually having had that offer for many years.

What makes a true single pane of glass? Because my perception is that you've got people dealing with different functions on different screens, maybe different people in different locations all trying to operate a device fleet efficiently and keep the connection up and running smoothly, and it seems kind of very fragmented and hard to keep track of across the whole device estate.

So could you talk me through, you know, the ingredients behind a true single pane of glass?

Mandeep Saini: Yeah I think to have a true single pane of glass, at least in the SGP.32 world, one thing that it comes down to is that you need one correlated timeline per device that holds all the information about what's going on in the real world. Because as you mentioned the question that an operations person or team is gonna ask, you [00:18:00] know, let's say is this device... You know, why did this device go quiet? Is there a network problem? Is it a same state problem? Is there a firmware issue? Is gonna be vastly different than someone else that may just wanna know is this device working or not? Is it online? So there's many different layers to providing what information the customer needs. And I really wanna push on why the device is the anchor for going forward, especially with SGP.32. Historically, connectivity management was very SIM centric, and that was for good reason too. The SIM was the only thing that never changed. It was very static as a natural anchor. But with SGP.32, that's breaking that, right? The software on the eSIM, the profiles, they can all change through the life cycle of the devices or the customer's use cases change.

So the SIM is no longer the anchor or stable. It can always change. The device is. So that's why, for me, with the SPOG, we need to start thinking about how do we transition from a SPOG to actually being device first? And with the device being the anchor you know, things stop being like [00:19:00] features. I wouldn't think of things as features ' cause all the dimensions of the solution can be changing, right? Like I said, the profile can change, the IMSI's can change. You know, you can have a different eSIM, but the thing that always stays the same is the device. So what I think with SPOG is that we have to be device-centric and think about how do we provide the right users the right information, versus like a single pane of glass for everyone. It's just not gonna work 'cause, you know, there's so much complexities that not everyone is gonna use or even understand 

George Malim: Yeah, that makes, really good sense, 'cause it did seem that having all the functions of a device, you know, thinking again back to the automotive use case that I was thinking about earlier, how are you knowing that, I don't know, a LIDAR sensor is performing-

Mandeep Saini: Yeah

George Malim: adequately and, or there's 5G or an alternative to 5G when the truck goes into a tunnel, or, you know, how is that all gonna be on a single pane of glass, and how is that useful to an individual operator or an individual process, so, I think there's an interesting, question of how single is the single pane.[00:20:00] 

Uh- Mm-hmm ... and I, and I think it's probably limited by practicality and usability, and I think that's a reasonable thing to end on, for the Spoke section of the podcast today. So let's turn to the final section, which is looking ahead, and, uh, got a very large question for you coming up.

Uh, it's fine, you don't have to bet your mortgage on it or anything like that. Uh- ... the question is, what will IoT operations look like five years from now? 

Mandeep Saini: Yeah, sure. No problem. Um, and I'll give you my honest answer, right? I'm gonna answer this question first by kind of just generally asking a rhetorical question, right?

Let's look at what AI is doing to the software engineering world right now. In that space, in a span of a couple years, it has fundamentally changed how engineers work. The tools, the workflows, and what the person actually does is very different than it was, let's say, you know, before OpenAI released GPT. And now nobody works in software the way they used to, like four years ago. I think the same transformation is gonna come to IoT operations. We're just a few [00:21:00] years behind, but I do think we're catching up, and we can take learnings from what happened in the software engineering world.

But to be very specific, I see there's three shifts. First, I strongly believe that connectivity is gonna become fully software-defined. So what I mean by that is how you provision switch and optimise will be treated like a program, but with 32 or the newer standards kind of helping make that reality.

Second is I believe the platform that we provide, whether it's FOG or something else, will become increasingly more headless. So what I mean by that is users or humans will less and less have a need to interact with the platform. What we'll start seeing is more agentic behaviour and more autonomous decisions coming in. But again, it comes back to, like, customers need to define what these policies are. But, like, this is where things are gonna happen, things are going. And more and more of the work or operations are gonna start happening through APIs, through agents, through policies, rather than someone sitting at a screen all day and doing something.

And then lastly, [00:22:00] following the AIoT story, you know, devices are never finished. They're always evolving. So as the first point I mentioned, they're living software with constant updates. So being device first and having with the operations with these agents, I think, you know, treating the devices as software that can change, I think is gonna really change how customers are gonna operate their solutions at scale.

George Malim: Yeah, I like the idea that instead of, you know, thinking in terms of operators and screens, we're thinking in terms of agentic AI and seamless processes that happen kind of autonomously below the surface. And I guess that's where all industries are going , not just the IoT enablement sector.

Mandeep Saini: Yeah. 

George Malim: But to close off the discussion, what role will standards like SGP.32 play in that evolution and feeding into those very different systems and processes that we've been talking about? 

Mandeep Saini: Yeah. So SGP.32, um, especially let's just talk about agents, right?

So even if an agent makes some changes, whether it's one door or two doors decisions, if something does happen, if we become more [00:23:00] headless, then there's no humans involved. If something does happen, you know, there has to be some jump into the real world. We will have to start integrating more into enterprises' systems.

So let's say, I'll give you an example, right? If let's say someone is using an agent to do a switch under these conditions, and all of a sudden the device stops connecting, we can't do anything, right? But how do we tell the customer? So if we hook into their ticketing system, you know, they should already have a workflow set up to do escalation.

So, there's that potential as well. And 32, like you asked about how 32 is gonna make this, or change things, like that's what's gonna happen. So 32 unlocks some new possibilities with the devices changing on the fly. And if you go into the more agentic workflows, then that makes this more difficult. It's like, what do you do when things go wrong? And then also when you talk about being device first, I call this like a distributor device operations platform. So with things changing so much, how do you actually know if you ask a simple question, "Is my device online," or, "How [00:24:00] much usage was there?" And then let's say in a span of a week or a month, you changed your profiles, you changed your IMSIs like you won't know that. So I think SGP.32 does bring in new challenges, and if you start bringing in agents, it makes it even harder. But I do think that that's where things are going, and being device first is just one way that we're gonna tackle this.

George Malim: That seems like a really good point to wrap up today's podcast. Thanks very much for joining us. This is Mandeep Saini from KORE, and myself, George Malim from IoT Now. But before you leave, don't forget to check out the new report from KORE, "Is Manufacturing Ready for the Next Phase of IIoT?" It's available for free at www.iot-now.com, so do check that out.

Thanks again, and goodbye