Our focus today is going to be really around the car. Through their QNX division, they've really become a critical enabler of software-defined vehicles, providing real-time operating systems and the hypervisor solutions. They're really in a unique position as QNX really enables the applications that OEMs or third parties create. While QNX really is the clear leader here in the vehicle, their software is deployed on over 235 million vehicles on the road today. They work with more than 40 automakers globally. 24 of the 25 top EV automakers are using QNX. Basically, all the Tier 1s also use them as well. They also offer unique solutions outside the vehicles to help developers create, test, and update applications in the cloud and then push those updates over the air. So we're going to talk a lot about that.
Here to discuss QNX, how the software and vehicle architectures are going to evolve, is John Wall, the Senior Vice President and Co-Head of BlackBerry Technology Solutions, and then Tim Foote, Head of Investor Relations. So thanks, guys, very much for being here.
Thank you, [Shreyas].
So maybe before getting into QNX specifically, I wanted to start with maybe the big picture. Automakers have been talking about software-defined vehicles for the last seven or eight years now. We're actually starting to see the rubber meet the road. There have been challenges. We can go through all of them. But a lot of them have been well publicized. You have conversations with these software teams pretty much every day. What would you say broadly is the biggest pain point that they're facing? Where are they really struggling?
Yeah, that's a great question. I mean, we have the privilege of working with just about every OEM in the world, whether it's in APAC, North America, Europe. You talked about 24 of the top 25 EVs. Well, we don't work with Tesla. That's the one that we call out. And I think the biggest pain point that we're seeing, the OEMs used to buy their ECUs from a Tier 1. They would specify the ECU. Tier 1 would build the ECU. Tier 1 would pull in all the suppliers, build the ECU, ship it to the OEM. The OEM would accept it, put it in the car, et cetera. Now the OEMs are taking a lot more control of software. And for instance, our contracts are directly with the OEMs in many, many cases. We still obviously work with the Tier 1s. They are now playing that integration role.
I think that that's been the biggest challenge for the OEMs. If you look at a lot of the delays that have occurred in programs, and there's many of them and they're very well known, whether it's two-quarter, four-quarter, six-quarter push to the right, has been at that integration layer. That is the main pain point. We are actually having conversations with the OEMs just in the last few months how they need to get out of that business. They really need somebody to provide a platform that they can move from hardware to hardware while maintaining a layer that doesn't change for their applications. Really getting stuck in the integration of having to work with 10 or 15 suppliers, bring everything together, integrate it, figure out what's missing, managing all these suppliers, that's been a very big pain point.
So I'm glad you brought that up because it seems to me that even some of the ones that are still being advertised by the OEMs, to me, still feel very complicated and still would require a lot of integration. I'm thinking about the Mercedes OS, for example, right? They were showcasing that at CES this year. Three different operating systems, yours, AUTOSAR, Linux. They've got multiple hypervisors, multiple developer toolkits. It would seem to me like that is still going to result in quite a lot of complexity even when that thing is fully developed and put into vehicles. So I know QNX can offer solutions where your hypervisor can support multiple OSes, for example. But are they sort of hearing the message that you're conveying that you really need to move to a more common development platform?
Yeah, they're not only hearing it, they're telling us. They're actually coming to see us and saying, "We want you to start managing these suppliers. We want you to create this, what I would call, agnostic vehicle OS that can underpin whether it's" and I'm not specifically calling out Mercedes-Benz, but you brought it up as an example. But they can underpin MB.OS. They're coming to the realization that what differentiates them is the applications that their customers can see, not whether they actually did the integration of the operating system or hypervisor at the base level. They're spending way too much time, spending way too much money. And so they're looking to companies like us to say, "Hey, you work with the adaptive AUTOSAR vendors. You do the proper integration.
You bring us a fully baked platform that is hardware-agnostic that will allow us to get to market quickly." Part of the reason that they're coming to us, QNX, is we're an operating system. Part of the problem of all layering these software stacks on top of each other is everybody's trying to be agnostic to the level below them so that, "Hey, we can support this. We can support that. We can work with this vendor. We can work with that vendor." What that does is it creates layers of complexity that lead to poor performance, bad stability. They look at QNX and they're like, "Well, you own the bottom layer. Here are the operating system, the hypervisor, some of the middleware.
You could take this and really integrate this tightly into the platform, provide the best level of performance, and on top of that, safety certify the entire stack so we're not dealing with five vendors that are giving us safety manuals for their pieces. We're giving them one safety manual. They can start building our applications.
Okay. That makes sense. Maybe just also to level set here, so how do we think about your market share and maybe the kind of the way the business operates primarily today on sort of a royalty basis? I mean, I think on the Q3 call, you had mentioned that 44% of QNX revenues came from royalties that quarter. If I'd used some rough math, I said maybe 50% market share and the number of vehicles produced, that was maybe something like $2 a vehicle. I don't know if that's the right way to think about it, but any thoughts?
Yeah, it's a great question. And it's one that comes up an awful lot. So what is your ASP per vehicle? And we deliberately shy away from giving that simply because it's commercially sensitive, as you can imagine. If you're talking to a large OEM, they know your average price. They're going to want to drive that down, right? That's just the way this thing works. What we have said is that the expansion of the ASP comes in two ways, really, two main ways. So the first is what we call a horizontal expansion. So you go back five years, you're running QNX. The only place that could justify running QNX from a compute power complexity standpoint would be an infotainment. Now you see the Digital Cockpit, the combination of the instrument cluster and infotainment. You've got ADAS. You've got gateways. You've got engine control.
That's the expansion, to have QNX running in these high-powered compute domains. And then the other aspect is a vertical expansion where, to John's point, we're not only just selling the core operating system, we're looking to sell extra layers as well. So middleware. So you're at CES, Shreyas, and we launched QNX Sound, which I'm sure we'll come to. There's IVY. There's all these other things. So we're trying to offer more, to John's point, to solve some of those sort of low-level, undifferentiated problems for OEMs.
Yeah. And so the important point, too, is as the vehicle is moving towards fewer ECUs, but each one being more powerful, that's not a negative for you. That's actually increasing content.
Increasing content. It actually goes beyond even some of those ECUs you would think about. It's smart radar. Smart radar. We run in smart radar systems, smart LiDAR cameras, camera systems that are becoming much more sophisticated, doing a lot more processing and cooking of the data before it actually goes to the centralized ADAS controller. So there's a lot of opportunity as the car gets more sophisticated, as this processing power increases, more opportunity for us.
Okay. That's good to know. And then maybe just think, so just to your point, as you're moving vertically here, it seems like one of the challenges outside the vehicle has really been around developing, testing, updating applications. Some of that seems to be around just the vast web of interdependencies that you have with different applications where one will feed off of multiple bins and libraries that a second one will. So one update will require revalidation of a lot of lines of code. I think you're offering a solution here with cloud enhancement that you talked a little bit about at CES. Stellantis is actually one that was a very big win that you mentioned. Maybe talk a little bit about that. It sounds like you're kind of creating essentially a digital twin of what you can do in the vehicle.
It can be in the cloud, same OS, same hypervisor.
Exactly. Yeah. I mean, anybody who's been involved in a large project for the car and we had the luxury of QNX being owned by Harman for six years and myself specifically being the interface to Harman, I spent a lot of time in Germany getting yelled at and working through these programs. And the biggest challenge was actually maintaining versions across different engineering sites, especially as you start to have engineering sites in India, maybe engineering sites in China, and trying to maintain that everybody has the same version of the tools, the same version of all the binaries that go into that system. Very, very challenging. Cloud is going to be a game changer for that. You're going to get your tools from the cloud. You're going to get your repository of software from the cloud.
When we talk about a digital twin or a bit parity, what's running in the cloud is QNX. It's the exact same version of QNX that's running in the cloud that's running on the ECU. It's going to be a game changer. I think it's going to be a very big part of our business going forward for development and then even perhaps beyond development.
Yeah. And so the ability to basically mirror what's in the vehicle and in the cloud, so you don't need to make any upgrades to the OS or anything like that. Because, for example, we've heard some other developers talking about you really need to be moving towards more like a containerized RTOS, for example, because otherwise it's not really easy to update and handle this issue with dependencies.
So I think that comment comes from the fact that the way workloads are managed in the cloud is containers, Docker containers. That is the language that the super scalers speak about. What we're talking about is actually QNX running in the cloud, middleware running on top of QNX that to a very high degree will allow you to do some of the development. For instance, a digital cockpit. We have graphics. We have audio. We have input. We have our VirtIO interface. So we're able to run Android as the infotainment guest. We're able to do a digital instrument cluster. It's not 100%. I mean, you can't do all the development in the cloud. There are hardware dependencies over time. But more and more we're working with companies that are building emulations of that hardware in the cloud. So you can do more and more in the cloud.
The container thing is a completely different topic. QNX does have containers now. We have built containers because we believe that's how workloads may be moved from the cloud to the car in the future. We've stuck to being OCI compliant. All the tooling for Docker and everything would work with our containers. The difference is our containers are safety certified. We'll be safety certified. We'll be able to manage safety-critical workloads.
Okay. And so far that you would be the only one that has the safety certified container. Is that right?
I don't know that I will make that claim. I'm not sure. But I would certainly have not seen it elsewhere.
Yeah. Okay. And then maybe just also another point that you've talked about recently is the compute requirements in the vehicle are obviously going up. And so you need to sort of future-proof a vehicle because the amount of compute required today or tomorrow is going to be a lot more than today. So you did push an important update, the SDP-8. Maybe just talk a little bit about why that's such an important?
Yeah, absolutely. I mean, if you were to look at where QNX was competing five, six years ago for safety-based systems, we were competing with Wind River, Green Hills. And our advantage at that time is that our operating system had been built with SMP in mind, so multi-core. But even then we started to see that once you got to four or eight cores, we didn't scale the way Linux scaled. So instead of looking back on Green Hills and Wind River and some of the other RTOSes, we set our sights on Linux, that we want to be able to scale and have the performance that Linux has while maintaining functional safety, ISO 26262 certification, and be able to keep our security pedigree like ISO 21434, etc.
We did a lot of surgery, probably the biggest update to our operating system in the last 20 years where we completely rewrote the kernel to actually allow the high-runner kernel cases to run on a per-core case. Now we have now scalability up to 64, 128 cores where we have almost a one-to-one scalability. This goes to future-proofing our product for years to come.
Just for context, I mean, today, I think most of the vehicles would have some like four.
Yeah, four. In some cases, eight. But if you look at next-generation NVIDIA parts, next-generation Qualcomm parts, they're north of 16 cores.
Yeah. Okay. Maybe also we could speak about the opportunity that you're seeing in some of the other middleware. The QNX Sound, as Tim mentioned, was one of the announcements also at CES. It appears that it offers the OEM really the opportunity to basically provide software-defined audio experiences without having the actual hardware, the heavy and expensive hardware that you would need to do that. I'm just curious how the OEMs have, first of all, responded to this. And is there any way you could frame the market opportunity?
Yeah. So QNX has always had an acoustics portfolio, so acoustic echo cancellation for hands-free calling. This is something that when we were part of Harman and we spun out of Harman, there was this group called Wavemaker that's located in Vancouver that spun out with us. And so very high pedigree of acoustics knowledge. And so in-car communication, acoustic echo cancellation, noise suppression, engine sound enhancement, etc. But about five years ago, we started thinking, why is it we can't bring more things into the car such as Dolby Atmos or other kind of audio functions that are done today on a DSP in a branded amplifier? And I think we were early. You have to hit these markets at the right time when the OEMs are looking to make a change.
I think with a lot of the talk about SDV, especially for me, the term SDV has become much more prominent in the last two years. Not that the activities have really changed. It's just the word has picked up a lot more steam. I think it's actually brought a recognition, hey, software-defined or sorry, software-defined audio is a thing. And the benefits are really cost reduction and being able to provide a similar experience to what you get in the standard audio. The head units have enough power now to be able to run some of these algorithms. There's DSPs that can be leveraged on a lot of the silicon. And I think the reaction that we've seen from the OEMs has been very, very positive.
I think whether we win all this business going forward, I think the opportunity is going to be there very strongly over the next five to 10 years.
Yeah. What's maybe just like a way to frame if we are talking about savings for the automaker in like the $100, $200, $300, like that kind of range?
Yeah.
Okay. Yeah. So that's pretty significant.
Yeah. I mean, you either bring value or you save money.
Right. Yeah. And I wanted to also just turn to the business. We have about maybe 10 minutes left here. So earlier this week, BlackBerry, you guys provided some updates on the steps you're taking to separate the IoT division from cybersecurity and kind of streamline your costs at the same time. First of all, what's the impetus to separate the divisions? And then what's kind of the progress you've made so far on that? And how should we think about the company once you're complete?
Yeah. Great question. So if we look at BlackBerry today, I think, well, I'm very excited about the opportunity that John and his team on the IoT side of the business have. I mean, propelled by all these secular trends, it's an exciting time. On the other side, we have a cyber business which equally has a very exciting market opportunity. I mean, cyber risks are not going away. But we got two things that we kind of need to address here. So the first one is if you're well, probably the people in this room, you're focusing quite heavily on auto, you might not necessarily want to be focused on enterprise software, right? And then vice versa. So we can see some opportunities from having more clarity in terms of the pure-play opportunity from separating these two businesses. The second is our cost structure.
So we've been in a pretty heavy investment phase, particularly on the cyber side. So when we acquired a company called Cylance, which is a leading endpoint security company, we had a couple of product gaps. So we've had to work really hard to close. Now, we've done a lot of that heavy lifting now. So we feel we can start to pull back. So to give you a data point, we've been running R&D at around about 30% of our revenue, which is clearly quite high. Secondly, our G&A, we've got quite a heavy G&A. It's running around about 20% of revenue when the benchmark might be nearer to 10 or 12. So we're looking to do both things at the same time, Treyas. And we're working pretty quickly. So you mentioned the update that we gave. We only announced that we were doing this three months ago.
In those three months, we've done a lot. Last quarter, we announced that we're taking $50 million out of the business. That's predominantly on this cyber side. This quarter, which ends this month, in two weeks' time, we're taking out another $55 million, again, predominantly cyber. We're starting to address some of the G&A opportunities as well. We said we're not stopping there. We've got a further target of another $45 million. The implication of all of this is clearly an improved bottom line, but also an improved cash flow profile. We've said that while we expect to see some improvement, in the meantime, we would become cash flow positive from Q4 of this coming fiscal. All of this comes around to say, OK, well, how do we see this business going forward?
We've not been super prescriptive about what that looks like yet. Clearly, these are two good opportunities, exciting opportunities. If we've got two businesses that are clearly IoT is already working really well. It's got a lot of momentum. We can try to mirror something on the cyber side. That opens up a lot of optionality. What we're saying is nothing is off the table at this point. We'll take a look at all options.
Okay. That makes sense. And maybe on the IoT division, I mean, like you said, right, it's doing really well. I mean, you're on track to reach some like $214 million of revenue this year. That's basically 35% of the total revenue of the company. This is a business that's doing 80%+ gross margin. But you've got a $640 million backlog. And by 2026, you've got so it sounds like you have some good visibility out to maybe $340-$370 of revenue, which is what you've talked about by 2026, so 20% annual growth. I'm curious how first of all, how confident are you in that backlog and the ability to convert that? Maybe just giving us a background of how this business works. Is it similar to the rest of auto where you book something two to three years in advance, and then it kind of launches?
How do we think about some of these newer opportunities, whether it's in cloud, IV, how that will start to add to the business?
Yeah. So great questions. So first point to mention, the 20% CAGR, we're just coming back. We're going to take a new look at that. So we're going to probably update the three-year outlook at our upcoming analyst day. However, none of the opportunity has gone away. So I just want to reinforce that. So how this business works is essentially right. You've got a design. We get selected to go into a vehicle or range of vehicles. And typically, over time, you get around about 40% of the total revenue in the period from the award to the period that it starts production. So we kind of call that the design phase. That's a combination of an SDK, which we call development seats, which are around about 20%, another 20% generally coming from services. So you mentioned $640 million of backlog. That relates purely to the royalty piece.
So that's that final 60. And that is on a one-time, when the vehicle is shipped, basis. Now, that backlog, we think, is our key performance indicator, right? Because this is not a quarterly business. There will be fluctuations from quarter- to- quarter. But if we look out into the future, if we keep adding to that backlog, then we're in a safe place. So last year, on a kind of an annualized basis, it grew around about 20%. We're going to give an update to that 640 number at earnings, which will be in just over a month's time. We've already indicated that that's growing at record levels. So clearly, bookings are very strong. In the near term, there's been some headwinds, which I don't know if we have time, but John might mention, around some delays to designs going into the build phase, I guess.
That's really the struggles that John's outlined that software for automation.
Which is also the opportunity.
Yes, indeed. It's a long-term benefit for us.
Yes. This is going to what? This is going back to what we were talking about earlier.
Platform. Exactly. But it does create some near-term challenges because if they're not developing the software, they're not consuming the services, and they're not drawing down the software dependencies.
Yeah. We got a couple of minutes here if there's any questions. Yep.
Talk about just top line there. Do we see top line connect grow in the wake of making some of these cost cuts? Because you've sunk a lot of that cap engineering. Or should we expect more moderate declining growth?
So we're going to come back to the full-year revenue outlook. However, we did reiterate that, so the key performance indicator in that side of the business is probably ARR. And that's been declining for a while. And there's some headwinds that we probably don't have time to get into now, but one of which was those software or product gaps that we had at acquisition. We've said that for the first time in a while now, we expect ARR to stabilize quarter-over-quarter, sequentially, should be flat. So we see that those headwinds, we're through the worst of it. But I'm not going to give you specific numbers around about this coming fiscal year until earnings. But where we're taking cost out is as follows. So R&D on the cyber side has been running too hot for the reasons I mentioned.
We're not going to stop investing in product. We're just going to invest at more sustainable industry standard levels. Secondly, we're taking out G&A. Neither R&D nor G&A, getting those to right levels, we expect to have a significant impact on our top line. So if we were taking it all out of sales and marketing, maybe that would be different. But we're not.
All right. Yeah. We have just a couple of minutes left. Maybe we can just touch on IVY. So it seems like you're continuing to make some good progress there. How do we think about, I guess, two things? I have questions there. Number one is just thinking about the interest that you're seeing from the automakers to really use this development platform. Is that separate from the cloud-enhanced offerings that you're talking about with Stellantis? Is that.
It is separate.
Those are separate, right?
It is separate. I think the IVY being a middleware component that provides an abstraction to normalize data from the vehicle signals. I think the biggest competitor that we had is the OEMs themselves because this is something they believe they can build. So this goes back to what I'm talking about, the platform. The opportunity now to have IVY integrated in this platform and just have it be part of the data management of this platform, I believe that the OEMs will be much more interested in realizing the ability to monetize data than building this platform. And again, I go back to the OEMs wanting to own software, believing that they need to own the software soup to nuts. And now coming to the realization that time to market, features that delight their customers is way more important than saying, hey, I developed the underlying non-differentiating software.
We are seeing more interest in IVY. We have POCs now going on with some larger OEMs. Tie that in with the platform piece that I was talking about earlier. I think there's a very good opportunity going forward. Now, having said that, that will go into the cloud. That will be able to be leveraged in the cloud. The initiatives are really two different initiatives.
Okay. So I guess if we're here two years from now, three years from now, do you think these automakers and if we maybe think about Tesla or maybe someone like that as kind of the benchmark, are we still talking about such a wide gap in software development between these two kind of companies? Or can QNX, maybe even some others, really help bridge the gap, assuming, of course, the OEMs are willing to compromise and focus?
Yeah. I think it really comes down to how much cooperation will the OEMs if they all come to the realization that this is non-differentiating software, we should be making our mark on the industry by the applications we provide to our customers, the quality of our L3 self-driving systems, whether minus fail operational can go to 120 km/hr versus one that switches off after 60 km. That's what they should be focused on, not the underlying foundation. I'm already seeing movement where I'm seeing OEMs cooperating that I never thought I would see. So I believe if the industry comes together, starts standardizing on some things, you will see less of a gap.
Great. We'll leave it there. Thanks so much, you guys.
Thank you, too.