Motorola Solutions, Inc. (MSI) Earnings Call Transcript & Summary

Welcome back. Hi there, everyone. And we can just close the doors at the back, that would be good. So I'm Rob Sanders, I'm the European tech hardware analyst, and I'm delighted to welcome Mahesh here, the CTO of Motorola Solutions. MSI for those of you who don't know, is a leader in public safety. Today, I'll go through some questions mainly focused on the tech road map. But if you have any questions, please do raise your hand and let's get started. So Mahesh, maybe just for the benefit of those people on the webcast and for those people in the room, just -- I just said you're a leader in public safety and security, but maybe just a quick introduction from your side in terms of the company.

Sure. So Motorola Solutions, we separated from Motorola Mobility, the cellphone company, so we're not that. And when we split in 2011, we were pure-play public safety. Since about 2018, we have been added to that portfolio with the acquisition of Avigilon in the video space, so video and access control and a series of acquisitions over the past, I would say, decade or so. . We are now, I would say, #1 when it comes to 911 CAD records from a public safety standpoint in addition to LMR, critical communications in North America. But separately, we are #3 in video security across the board as well, video security and access control globally, excluding China. And the key value here is because we are so laser focused on enterprise security, public safety and networks that support both. We are able to connect and bridge those 2 elements appropriately and that is really where we are as a company.

Great. Yes. So maybe let's dive into the sort of what you call now mission-critical networks, the kind of networking side. Maybe starting with the tech road map. So what is the sort of long-term strategic plan for technology around connectivity. For example, the transition to converged solutions from LMR and how is that -- how is the rollout of your APX NEXT portable radio, for example, played out, which integrates 4G, for example?

Sure. So APX NEXT adoption has been incredibly strong. And it's also important to point out that the majority of our revenue today does come from the APX original family of radios. So we have a lot of fleet conversions that are in progress. Now and there's a good runway ahead of us. But adding to the appeal of this converged connectivity solution, adding LMR with LTE, 4G and beyond really is the ability for us to data-driven applications into the mix. And so we have done that via apps for the APX NEXT platform. 90% of our APX NEXT customers today subscribe to this app suite at bundle, and that's significant for us, and we're continuing to add more functionality there. We recently introduced the SVX body-worn assistant which combines a remote speaker mic, a body-worn camera and an AI assistant in one device specifically for the APX NEXT device. And what that does is it adds to the app ecosystem in terms of the AI assistant capability but also extends what we are able to do with a mission-critical network and the mission-critical radio the officers carry today.

Got it. And what about 5G? Obviously, you've integrated 4G. Is there something -- is that something your customers are demanding that you add into your road map?

To be honest, there's no incoming request for it. I mean, we track the evolution of these standards, and we make sure we are up to speed there. If you think about just spectrum and you think about the use cases that our customers care most about. Some of the challenges with 5G, and obviously, we can radios multiplex between 5G, 4G, et cetera, as appropriate, penetrating buildings going through basements, et cetera, the number of -- the amount of coverage we have with 5G. There are a lot of challenges there. There are benefits as well, higher bandwidth, lower latency, et cetera. But the way we look at it for our customers and the bandwidth that they need, given the resilience, the demand, the ability to work in any given situation the solutions that we have today from a converged standpoint is starting to make an impact, and we are continuing to add capabilities as we migrate to 5G and other solutions.

Got it. And then let's talk a bit about direct-to-device services, satellite-based connectivity in remote areas. How will your portfolio evolve to capitalize on these technologies?

So when you think about satellite, it's an interesting space, but it's worth mentioning that when you talk about satellite in the context of direct-to-device, there are a few constraints there. One, it is low bandwidth. So today, most direct-to-device capabilities are largely text messages and very low bandwidth communication, not even sufficient for voice, and we're trying to do more than voice today from a data standpoint. The second is the number of subscribers, the number of users who can occupy a small space. So if you have multiple officers, multiple responders responding to an incident, the ability for you to leverage solely satellite direct-to-device, it becomes a real challenge. And the third is, oftentimes, you need a clear view of the sky in order for you to get decent connectivity. So all that said, satellite is actually a powerful story for us, but perhaps not from a direct-to-device standpoint, but really coupled with our infrastructure, our LMR infrastructure. So we've been continuing to evolve our infrastructure with the DCRE base station, adding capabilities to the infrastructure, but that's also a good point for us to enhance connectivity rather than just it being purely wired, there may be an opportunity for satellite there as well. But -- so we do believe from a connectivity standpoint, it's a resilient story, more than pick one connectivity mechanism for us.

Got it. And obviously, you've recently made a very notable deal, $4 billion deal to acquire Silvus. Silvus, as I understand it, and correct me if I'm wrong, is giving you this mesh network capability. I mean today, it seems to be used for unmanned aircraft drones. It's taking obviously more into the defense space. So what's the strategy for integrating that so-called MANET technology into the broader technology portfolio?

Sure. So the space in general is MANET, the T silent, don't ask me why. It's mobile ad hoc networking is what it stands for, and it's really mesh networks for high bandwidth, low latency, highly secure communications. And some of the key properties of MANET specifically for the defense applications in this case is that it has to work in really harsh battlefield scenarios, this includes low probability of detection, low probability of intercept, supporting anti-jamming capabilities, et cetera. And Silvus as a company with their MANET technology really grew up and proved themselves out in Ukraine, where this is one of the toughest challenges. So as a very proven quantity, they have been able to expand not just from radio data communication in a mesh network topology, but really address drones more broadly. Drone communications in that type of space requires all of that, low probability of detect, low probability of intercept, anti-jamming capabilities, et cetera. And today, Silvus supports more than 100 manufacturers of drones out there in autonomous systems. Anduril, for example, is a good example of one that uses Silvus. And the advantage here is that as we go not just from a single drone solution, the world in this space is starting to expand to swarms of drones. So this also has one of the highest node count in terms of the mesh nodes that it can support within one network. And from that standpoint, there are some powerful things that are capable with Silvus. We do believe that there's another part of this business, which is not just spectrum dominance as they call it, for low probability of detect, low probability of intercept. It's also spectrum monitoring. So being able to understand what are the emitters out there from the standpoint of RF? What's emitting RF out there, including is there a drone that's out there that's emitting RF that you want to detect and geo locate. So this also comes into our public safety story where we can start to look at this not from a communication standpoint, but also from a drone detection standpoint, and that's the spectrum monitoring piece of Silvus. So we see this as an expansion opportunity, first and foremost, as we go into new defense which is really driven by autonomy, drones, et cetera. But we also see this as an opportunity for us to eventually get it into public safety, starting with things like drone detection.

And how is your channel today? I mean, is there a cross-sell play into those defense customers? I mean, obviously, when I think about you today, it's more public safety, different kind of customer set. Obviously, one very large customer, the Department of Defense. I mean how -- is that those relationships already exist? Or is Silvus bringing those relationships?

So I would say that one of the core thesis from our side in acquiring Silvus and before Silvus, almost $900 million of our revenue came from the federal space, the government space. So we did have presence. We had the contacts. We had government affairs and lobbies relationships. We have contacts across the board that allow us to bring Silvus, which is a technology-heavy, strong platform, coupled with our sales access and our go-to-market capabilities. Bringing those two pieces together, we believe that there's really good synergies there.

Okay. And what are the risks you see migrating to cloud-based architectures. Obviously, I've seen this with RAN, for my coverage, Ericsson and Nokia. How do you mitigate these risks and challenges around interoperability, for example?

So in the context of mission-critical networks, I would say, probably the biggest hesitation from the standpoint of customers is that they're used to our infrastructure, our LMR infrastructure exceeding five 9s of availability. And so when you think about cloud platforms today and cloud platforms, regardless of what application that they're running, you're talking about sub-5 in many cases, sub-4 9s of availability. So there's a level of redundancy and capability that's required there for us to be effective. That said, when you think about failover as an example, where you can failover from on-premises to the cloud in the case of a hurricane, and this happens frequently in places like Florida, where we -- our networks have had to failover into the cloud because the on-premises infrastructure had to go -- went down because of lack of power, flooding, other situations. So cloud as a failover mechanism for on-premises now adds to that five 9s of availability. And so that is the story that our customers are liking and we're pursuing in this space.

Okay. And of course, we've already gone 10 minutes. We haven't used the AI word, which is -- but we should, of course, ask, I mean, just from a technology perspective internally in terms of improving your portfolio, how are you leveraging AI and ML to enhance your network performance. Obviously, we've seen AI coming into the RAN space. I mean, how do you see that improving your network performance reliability, security, et cetera?

So if you take away all the capabilities that we've been from a mission-critical network standpoint from voice to data to other even video and such transmitted through this network. The ability for 2 responders or 2 public safety professionals to be able to talk to each other reliably is incredibly important. And that voice communication is incredibly important. So it starts with really how we use AI to enhance audio, specifically speech. And if you're in a scenario where there's a big accident, there's a fire. The noise around you is incredibly immense and crazy. And there's the ambient sound you're not going to be able to use your iPhone to make a call and have the other person on the other side actually understand what you're saying. So the ability for us to leverage AI to isolate voice, to enhance voice and to be able to transmit this through a resilient network so that the other person can actually hear and understand what you're saying that's one of the first areas where we apply AI effectively. Separately, we also leverage it for network monitoring. We have the capability to understand what coverage looks like, where perhaps feedback from device, the telemetry data that we get in terms of battery data, signal strength data, et cetera, allows us to diagnose capabilities and such. So I think that's the second area where we actively apply AI. And the third is these public safety networks are cybersecurity targets. And the number of incidents year-over-year has grown quite tremendously. What we have seen is we have seen strong uptake in our Managed Detection and Response, MDR capabilities for public safety networks. We've seen about 114% growth year-over-year in uptake there. And what is interesting there today is with AI, we roughly process across our customer base, about 1 billion incidents a day, cyber incidents a day and 99% of those incidents are automatically triaged leveraging AI. And a small 1% of those are actually looked at more carefully either through a second pass with AI or a much smaller portion looked at by human analysts to see how that can be mitigated. So that's how we leverage AI starting from resilient quality going into making sure that resilience in the network is maintained by understanding potential faults. And the third really is making it secure.

And just adjacent to that, have you started to use. I mean, I don't know how much software development you're doing presumably quite a lot. Have you started to see the benefit. The productivity benefits of AI to actually getting time to market and perhaps lowering your cost as well?

So we're experimenting with every single tool that is out there from a developer productivity standpoint. We were early adopters of Copilots. We have leveraged tools like Cursor, Sourcegraph, Windsurf, et cetera. We are starting to now hit a stride in terms of how those tools can actually improve productivity in a meaningful way. The challenges that we have to keep in mind here is that in our space, resilience and availability of our solutions matter a lot. So testing is a big part of it. So one of the areas where we're actually applying AI to begin with, is to create automatic test capabilities and really test the resilience of our solutions as a whole. We will start to see, I think, more meaningful productivity in the future, but it really becomes one where we need to decide whether it is savings that we now use to accelerate innovation across the board, and that is a possibility that we see here as well.

Okay. And maybe just switching -- you mentioned earlier, the SVX remote speaker mic, that product cycle. Can you just give us some more stats in terms of upgrade, progression of the installed base, share gains and the AI features you just mentioned around voice, how have they been received, for example?

Sure. And just to maybe remind you as to what SVX is. SVX creates a new category for us. It goes from us talking purely about body-worn cameras to really body-worn assistant. And it is coupled with our APX NEXT device. It brings 3 devices into 1 or 2 physical devices into 1. And those 2 physical devices are a remote speaker mic, which most officers leverage when leveraging a radio and the second is a body-worn camera. It combines that into one device. There is manageability benefits there. There's maintenance benefits, reduces the weight of what an officer needs to carry. The second is that from the standpoint of having complete access to information that is stored, a body-worn camera is supposed to see and hear everything that the officer sees and hears. A remarkable number of officers wear ear pieces when they are talking on their radio. So the body-worn camera doesn't naturally hear everything that the officer hears, which ends up being critical from an evidentiary standpoint, also when you're trying to author reports and assisted narratives after the fact, being able to help with that you don't have -- the data is not complete. The SVX taps into all of it. It taps into all your LMR audio, whether it's in the ear piece or it is heard out loud. We also have access to 911 audio and data, and we have access to computer-aided dispatch data as well. Bringing all of that together, the SVX is now a complete evidentiary platform. And the third is we have significant TCO advantages compared to any competitive product out there, not just from the standpoint of it being one consolidated device, but we also leverage connectivity in the APX NEXT radio. So it leverages the broadband connectivity and the redundant resilient connectivity that's available on the radio, both from the standpoint of not having an agency or a customer having to subscribe to multiple data plans. It's one data plan, but it's also a more secure and available solution. So that is really what SVX is. We launched SVX in July of this year. So it's just been a bit more than a month at this point in terms of uptake. The uptake has been incredible. The customers that have placed orders so far are, for the most part, are not existing body-worn camera customers of ours. So we are seeing some competitive flips there. And we have actively worked to make sure that, that pipeline continues to grow, and we are seeing strong adoption and growth in that pipeline.

Got it. That sounds very encouraging. And in terms of the video security access control area, how are you integrating AI, ML, video analytics into that space. Is that something where there's some interesting innovation cycles coming through?

So I think it's important to note that we have been leveraging AI in video security for over a decade at this point. Over 80% of our cameras has AI built into it. At the edge, we take advantage of the latest advancements in silicon accelerated AI processing technology. And we have a fairly sophisticated and tenured AI team that specializes in computer vision and also AI more broadly and more recently, generative AI capabilities. To that edge compute capability, we've also added a whole lot of AI capability right into our network video recorders. This is -- these are devices that record video on-premises but we've also added the same sort of capabilities through to our cloud solution. So in the cloud, our Avigilon platform has 2 pieces, Unity, largely operates on-premises and Alta, which operates in the cloud. We have introduced Alta Intelligence, which is a capability that provides AI to all cloud-connected devices as well. A few years ago, we acquired a company called Calipsa. Calipsa was a cloud analytics platform, which has now become our cloud generative AI platform as well and Alta is benefiting from a lot of those capabilities. And lastly, in addition to AI as a SaaS service that is cloud delivered, we're also introducing managed services that sit behind all those detection mechanisms. When AI detects something that is important, who responds. Typically, only customers who are in the Fortune 500 are above have an SoC, the security operations center that is staffed and manned to be able to respond to events. If you want to democratize that and increase the appeal for AI and analytics, you want to make sure that there are people who can indeed respond to those events. We introduced a solution called Alta Protect for our cloud platforms, where we have active monitoring support, people monitoring support that can attach to it on the back of the SaaS AI services that I mentioned. We also have another solution called Envysion, which is really used by quick-serve restaurants, think of customers like Chipotle, et cetera, who leverage this type of solution, they too would like monitoring capabilities, and we're coupled monitoring in the same fashion as Alta Protect into the Envysion platform, and we see strong adoption there as well. So across our AI breadth, whether it is AI on device, majority of our devices have that. In fact, a disproportionate majority of our devices have that On-prem storage, cloud also benefits from AI, and we are introducing both generative AI capabilities on premises and in the cloud. And then finally, to take advantage of all of this to make sure that the action cycle is complete. We are -- we have introduced services that get coupled with those AI detection capabilities as well.

Got it. And that those services, just -- you mentioned managed services. It made me think that, that might be a lower-margin business, but obviously, it's a way to create -- generate demand.

Correct.

So -- but you are willing to build out people as well to support their growth?

And in this case, we leverage third-party services to do that for us, and there are quite a few of that. And we are increasingly seeing that AI is playing a bigger role where the number of -- just like what I mentioned with our managed detection and response framework for cybersecurity, the vast majority of events never need human eyes on them. And so this is where AI makes a huge difference.

Got it. And the last group, sort of area of your product is command center. That area, what I'm interested is sort of areas like predictive analytics, those kinds of things. And when you look at that portfolio, are you concerned perhaps that best-of-breed competitors are potentially more nimble than you as you look at that portfolio?

I don't believe so. And it is because over the past decade or so, we have assembled key assets both organically and inorganically to attack every part of the public safety workflow. And the biggest trend we are seeing right now is that as you look at a full workflow, public safety incident workflow, roles that are participating in the execution of that workflow are being consolidated. And as you see that consolidation is a key trend, whether that is because of understaffing or other capabilities, that is the opportunity that AI delivers and having an application platform that actually addresses every part of that workflow gives us that unique advantage to consolidate those roles. And so we see that as actually a powerful differentiator for us today.

Okay. And I'm interested in your NLP capability, neuro-linguistic programming for processing unstructured reporting. Obviously, you're dealing with all sorts of different data which can be hard to kind of synthesize. So how are you doing that? And that seems to be an area where you could potentially create a lot of benefit for the customer.

So we dipped our toes in natural language understanding about 6 years ago almost. Since then, we started in the 911 space where we had some of our biggest customer base, and we started with transcription. So the ability to transcribe live audio through the course of 911 calls. We then went from 911 transcription to translation, being able to translate languages as well. We have then taken that transcribed data and built upon it and said, what is being said, how can we action based upon what is being said, whether that is disambiguating addresses. There are many recent cases somewhat recently where a mother called 911 on behalf -- sorry, a daughter called 911 on behalf of her mother. And they were in different locations. The cell phone was the one that registered a location, but the cell phone belongs to the daughter, not the mother. And through the course of the conversation and analysis of the conversation, our AI capabilities flagged to the call taker that perhaps the person who requires assistance is not in the same location as the person who is calling for help. And so that was immediately clarified. Responders were sent to the right location. That really tells you how are -- we have built up our core language understanding capabilities. That has now expanded into other parts of our solution. We launched what we referred to as Assist. Assist for us is this AI platform that touches every part of both our public safety and our enterprise security solutions and Assist is that glue that really binds together the workflows across these different solutions. And as I mentioned to you before, we see the ability for us to touch the various parts of the workflow as a key advantage in being able to stitch together and optimize these different roles, user roles appropriately. Assist is a key way for us to go do that, and that is a key AI investment that we have made. And that leverage is not just language understanding but it also leverages multiple other AI capabilities in the background to bring that together, both structured and unstructured data.

Just seeing if there's anyone who wants to raise a hand and ask a question, please do. Otherwise, I'll continue. Can you discuss your R&D resourcing? Obviously, it's a very dynamic time with AI. What might have been true 5 years ago is very different today in terms of how you think about where you want to pick your investment when it comes to talent. So how do you see that going forward in terms of where you need to invest?

Look, there's as we talked about before, with things like AI productivity tools and such coming to the forefront, there are certainly efficiencies that we can gain from a development standpoint. But we are a deep tech company in that whether you are talking about RF capabilities, whether you're talking about public safety workflow on our enterprise resiliency capabilities. This requires deep domain knowledge. And so where we invest actively is to get that deep domain knowledge in-house because that's what allows us to understand what are the key customer problems that we need to go solve and how effectively can we solve them. So from that standpoint, whether it is Silvus where they bring to us a great degree of software-defined radio skill set, cognitive radios and MANET technology in the mix, understanding of RF, combined with our own in-house capabilities from an RF standpoint, that is deep technology understanding, everything from software development to physics that we need to understand, and that's an area where we're actively investing in. On the AI side, very similarly, it's not just understanding how a generative AI, large language model, how transformers work, et cetera. It is now being able to understand how you can distribute that capability, not just in the cloud, but all the way into the edge devices appropriately because if the network goes down, your public safety or your enterprise security solution cannot see operating. It has to continue to operate. It has to degrade gracefully if it does have to degrade. That requires deep domain knowledge, and we are investing in that knowledge across both silicon sensing and AI.

Got it. Yes. No, obviously, the long-term dream is all the AI comes to the edge, and we can have edge compute without having to go to the cloud to request from open AI or whoever, yes, it makes sense. Maybe just the last couple of questions given the time. On the IP side, you've had a few infringement issues in the past. I can imagine there are some Chinese companies in particular that would love to have a lot of the IP you guys have. So how do you go around protecting your IP, given that you've had some challenges in the past?

We've -- quite a few years ago, we completely got out of China. We don't have an ongoing operation in China anymore. In terms of our solution set in-house, in terms of our development capabilities and such, we have strengthened our cyber protection. We have introduced processes from a development standpoint. We have a patent and IP protection process within our teams to make sure that we are robust, and we are effective in protecting our IP. In addition to that, we take our IP very seriously. And in the event of an infringement, we do take that to the bitter end, whatever that end is.

Yes. Makes sense. And just -- I shouldn't end this before, but we discussed this transition to recurring revenue and the SaaS model. So how has that transition gone from sort of specialized hardware to a greater proportion of software? And what challenges have you found? And does it mean, for example, more off-the-shelf hardware will be relied upon?

We are actually seeing quite strong recurring revenue growth, in fact, incredibly strong SaaS growth as well. I think we've previously talked about the fact that even in our command center portfolio, SaaS revenue has grown at over 2x that of our overall command center revenue. We are seeing that trend continue. We see ARR growth significantly in this space as well. But it is important to mention that whether you're talking about video security or access control, sensing has to remain on-premises. Oftentimes, it is compute that is distributed away from a server or a typical on-prem appliance. And that is going in 2 directions. It's going to the edge, but it's also going to the cloud. And when you bring those 2 pieces together, we see SaaS growth, not necessarily being at. As we transition to new platforms and new technology, we see that growth not necessarily being inhibited because there's a lower amount of cost hardware that is sold necessarily. I think that trend we see more from the standpoint of as customers migrate from on-premise solution like our Avigilon Unity Solution going into an Alta type of cloud solution, there are fewer servers sold, but then there are more AI capabilities, more edge capabilities that come with it as well.

How do you see silicon in that whole equation? Do you feel the need that custom silicon is going to become more important, less important? How do you see that? Or are you happy to buy standard silicon?

So far, we have not seen a deep need for specialized silicon. We have seen most edge SoC companies really push the envelope when it comes to AI accelerators. And where they fall short, AI accelerators, low-power AI accelerator silicones have sort of bridged the gap. So from that standpoint, we want to ride the curve. AI architectures for a large part, transformers are dominant today, but there are multiple other architectures that are being investigated in academia and elsewhere, which are starting to show promise. So rather than dedicating silicon today from our side, we want to ride the wave in terms of where this evolution goes and really have the expertise to take advantage of whatever off-the-shelf silicon there exists out there.

And I mean the upgrade cycle presumably could accelerate, which would be the big benefit.

That's right. That's right.

Right. Well, let's leave it there. Thank you so much, everyone, for coming. And thanks again to Mahesh for flying over from Boston at short notice. So I appreciate it. Thank you.

Thanks for having me.