BB:CC - BlackBerry Limited (BB) Presents at Wolfe Research Global Auto Auto Tech and Mobility Conference 2023
BlackBerry Limited (BB)
Wolfe Research Global Auto, Auto Tech, and Mobility Conference 2023
February 16, 2023 08:55 AM ET
Company Participants
John Wall - Head, QNX, Senior Vice President & Co-Head BlackBerry Technology Solutions, Products, Engineering & Operations
Tim Foote - Head, Investor Relations
Conference Call Participants
Presentation
[Abrupt Start]
…So many of us have -- are really auto analyst, but we're increasingly having to become at least armchair software analyst. So there's – there are around 80 to 100 million lines of code in a vehicle today. And that's likely to increase to over 200 million over the next few years and then grow further from there, driven by areas like ADAS, infotainment, battery management software and much more, and underpinning these applications is the real-time operating system and the clear leader in this area has been QNX. They've developed in the 1980s, QNX has been deployed to over 250 million vehicles on the road. They work with 45 automakers globally, including all the leading global OEMs and several of the leading Tier 1 suppliers. It's a $200 million business today growing at about 20% per year for the next five years.
So here to discuss QNX and how the software and vehicle architecture are poised to evolve over the next few years is John Wall, Head of BlackBerry QNX and Tim Foote, Head of Investor Relations. Thanks so much, guys. Thank you.
John Wall
Thank you.
Question-and-Answer Session
Q - Unidentified Analyst
So maybe before we get into QNX specifically, I wanted to start with the big picture. So as I mentioned, QNX works with every automaker in the world, and you've got conversations with their software teams daily. So broadly, what has been the biggest pain point for OEMs today? Where do you think they're really struggling as they're trying to become more software defined?
John Wall
Yeah. I think that's a really interesting question. I think, the OEMs that we work with, as you mentioned, we're working with just about every OEM. They recognize the importance of software in the vehicle, they know the car is going to become software defined. They know that from a monetization perspective software is going to play a big role.
I think the struggle for the OEMs has been how to build a software team within what is largely a mechanical company. And going out and hiring a bunch of software developers at a university is a really difficult way of building a software team, you need a lot of structure. You need a lot of experience.
I think, another area that's been very challenging for the OEMs is at what level should they be focusing on developing software versus what they should buy. And so I think that that's been the struggle. And I think, certain OEMs are coming to the realization that they're starting to understand a little bit more where that line is, where they should be providing value and software, things that their customers will appreciate versus maybe getting into the weeds of software where they really don't need to be there.
Unidentified Analyst
Yeah, yeah. And that -- yeah, so that maybe leads into QNX. So, just wanted to, obviously, as I mentioned, you guys have been incredibly successful in this market. Why have so many OEMs chosen to use your RTOS and Hypervisor solutions?
John Wall
Yeah. As you mentioned, we've been around for a long time. I wasn't there in the 1980s. But I've been there since the mid-90s. And I think, one of the advantages that we have and one of the reasons that we're successful is our pedigree in automotive did not start under the hood, on MCU’s, or small processors, our pedigree in automotive started in infotainment. And so from 2003 or 2004, we were -- we had a large market share of the infotainment sockets within the vehicle.
And so that really focused our operating system on high performance, running on high performance processors, rich graphics, but we really focused on that and so when we started to see a shift towards the software defined vehicle and that really started in the early 2000s and 2010s, discussions about autonomous drive, the need for high processing power within the vehicle, high speed networking. We basically took our product and we said, okay, well we're going to safety certify this product. And we're really going to focus on what's happening in the future. We could see Android was coming in infotainment. We knew our days were numbered. So we really started to focus on the rest of the vehicle, really focusing on functional safety, making sure that we had the right pedigree for our product to do, the highest levels of safety, security within the vehicle.
And I think as we shifted into that direction, we had the right product at the right time. And that is the consolidation of all these small, single function ECUs to these domain controllers on high performance compute. We had great relationships through our infotainment days with the likes of Qualcomm and NVIDIA, which is very important to have the support for the processors that are being used in the vehicle, TI, NXP, and I think we were at the right place at the right time, with the right product.
Unidentified Analyst
Okay. Makes sense. Yeah. And just a level set, when we think about the layers of software. Theirs is, you've obviously you guys are basically in that operating system, and then Hypervisor and middleware and then above that is where the OEMs or the suppliers or whoever can play in terms of application?
John Wall
Absolutely. So we are – I like to – we joke about this we're the plumbers.
Unidentified Analyst
Yeah.
John Wall
We're what's behind the walls, what's under the floor. You don't really see what we do. But it has to work, you know, for the rest of the applications to work. We don't have a strong desire to move up into the application space. We're very much foundational software, Hypervisor operating system. Middleware that helps the OEMs or Tier 1s or other Tier 2s in the market create their applications.
Unidentified Analyst
Yeah. And then so you know, maybe just double clicking on that. So, several OS and Hypervisor players, there are a lot of them in the market, right. There is Green Hills, Automotive-Grade Linux, number of people here were at Aptiv's Investor Day a couple days ago and they talked about Wind River. So -
John Wall
What would they talk about Wind River?
Unidentified Analyst
Yeah. They brought it – somehow they brought it up. So I guess, I'm trying to think about what – what differentiates these solutions? So maybe if you could put yourself in the in the shoes of an OEM when they're trying to figure out, which one to go with which KPIs Do they really care about?
John Wall
Absolutely. And I mean, it's a great question. It's one that comes up quite often. So I would say, when you're talking about Green Hills, and Wind River VxWorks, I mean, strong pedigree and safety. These are operating systems that are used have been used under the hood. They're used in avionics. They have certifications deal 178C, not an area that we play in, but I think as I mentioned earlier, we're focused on high performance. Our operating system is geared towards the next-generation, high performance compute, multi core, four cores, eight cores, 16 cores, very much performance that's similar to Linux.
I don't believe that Vx and Green Hills have that performance characteristics on the latest processors. So I think that gives us an advantage against them. If you look at Linux and AGL and AGL is a great name, Auto Grade Linux, it makes it sound like it's safety certified and but it's not. And, there are initiatives out there to safety certified Linux, BMW was a large proponent of safety certifying Linux, invest a lot of time and energy, in 2021 Frank Weber, their board member came out publicly said, we've wasted our time in the weeds. We're now going to use QNX as the foundation of our – of our products that are – that need safety certification. And that's an example of an OEM that felt they needed to own the software from the ground up and I think came to the realization okay, this is where we should buy and focus our energies at a higher level.
Unidentified Analyst
Yeah. And maybe, maybe talk a little bit about that, right. So because obviously, I'm sure you saw a few years ago, pretty much every OEM was saying they want to develop their own OS. And I think we've all learned over time that that was a very loose use the word…
John Wall
My CEO used to ask me about it all the time.
Unidentified Analyst
And it's not really what and now they they're sort of pivoting a bit. But thinking about the cost benefit, right? Like if I wanted to go and actually do what you guys are doing, the cost of doing that versus the cost of simply just using your solution?
John Wall
Yeah. I mean, let's be clear, when companies talk about developing their own OS, they're not talking about creating a QNX.
Unidentified Analyst
Yeah.
John Wall
What they're really talking about is creating a much higher level and I think the perfect analogy is Android, building more of an ecosystem for the car. When you think of the software defined car, what you're really thinking about is a handset, an ecosystem, iOS ecosystem, Android ecosystem, ecosystem for the car. So that's really what they're talking about. You have a very complex system, many, many processors ECUs. They're trying to create a layer above that that kind of unifies all the different parts of the vehicle and creates API's and interfaces to be able to talk to the car essentially. They're really not talking about writing a QNX operating system. I mean the amount of time and energy that goes into developing an operating system and getting it safety certified, I mean, it's prohibitive.
Unidentified Analyst
And I mean, as we're kind of looking ahead to the next generation of architectures, right, there's going to be more consolidation. You've already seen that now. As that happens, effectively the number of domains is not really changing. We're still going to have an infotainment layer. We're still going to have BMS and all these things. But does all of that need to be now done over using the same operating system? Because or does -- or are we going to be creating kind of a -- are we not really solving the problem? We're just solving the problem of having too many boxes?
John Wall
Yeah. So we are definitely solving the problem of having too many boxes in the car. And that's the consolidation. Obviously, I would love them to -- I would love the OEM to say, hey, we're going to use QNX across everything. We're not at that stage. There's still really if you look at what the demand is within the vehicle, it's Android for infotainment, which we handle in a digital cockpit with our hypervisor. We actually work very closely with Google to make sure that the interfaces to the hypervisor are defined with the idea being that you can update the Android within the vehicle very easily. And you have a mix, you have Linux for things that don't require a safety certification like telematics, although even telematics with e-call now is having starting to have a requirement for safety. And then you have an RTOS, RTOS for safety. And that's kind of what the mixture looks like today.
Unidentified Analyst
Yeah. Okay. Interesting. And then yeah, I mean, maybe if you guys maybe can help kind of frame the revenue model of QNX. So I think 60% of the revenues come from one time royalties per vehicle, 40% from engineering, and maybe you can help unpack that? And then also, are you seeing sort of a move away from that royalty to sort of recurring revenue model?
Tim Foote
Yeah. Okay. I'll answer that. So if we take a look at over time, and in last couple of years has been somewhat turbulent for QNX and for the auto industry over time, typically, when we get designed into a vehicle, so let's say, I don't know, Toyota choose QNX in, you have an initial two to three-year period, before the vehicle starts to be produced. It's being designed or the software has been designed and the physical features and stuff.
In that first period, as you rightly say, we get approximately 40% of our revenue that comes from two main sources. The first is what we call dev seats, developer seats, which is effectively a software developer kit for the OEMs or the Tier 1s to actually develop the software on QNX, and that's roughly 20%. And then we get a further 20% from professional services where our people will help with things such as safety certification, and the like.
Then it moves into production. And every time a vehicle gets shipped, we'll get a one time royalty. And that is kind of crazy recurring because the production then will be you know, fairly constant for a period of, say five years. So that's the total -- sorry, Tim…
Tim Foote
Well, I'd say, it's important to note that it's a royalty per instance of the operating system. So if there's multiple modules within the vehicle, that's using QNX and they're all individual.
Unidentified Analyst
Yes. Okay.
John Wall
But right now, Shreyas [ph], we're in a position where production is obviously slightly subdued. So the royalties side of the equation is slightly down. But we're producing double-digit year-over-year revenue growth driven by new designs. So we've won some very big designs recently, VW, BMW, these are names we've given publicly there are not a lot more that we can't give publicly. But these designs are much bigger than they were before.
So as John was saying, designs five years ago, there's probably only one spot in the car with enough compute power to run a high performance real-time operating system like QNX and that was infotainment.
Now, as we've seen this consolidation of compute power into these higher powered MCUs, we're actually seeing much -- many more sockets as we call them, for QNX to operate. And where we see those ingredients come together, we're winning a very high proportion of those sockets. So designs now could have three, four, five instances of QNX, whereas in the past, there was only one. So right now we're winning big designs, which is translating into that early stage revenue, and that's driving the year-over-year growth at this point.
And the good news, that's effectively a leading indicator, because these big designs within a couple of years, start to enter production. And as production starts to normalize as well, that's a potentially another tailwind a couple of years down the line.
Unidentified Analyst
That's helpful. And maybe so if I use kind of that example, right, so let's say, I'm just making, let's say, VW has five domains in the car.
John Wall
Yes.
Unidentified Analyst
And three of them are using QNX and two are using somebody else. So for each one of those, if -- for the software developers that are writing the applications for these domains, they each will have to use the three that use QNX will have to use the QNX developer kit, the two that are using somebody else will have to use someone else and then there has to be someone that's integrating all that. Is that how it basically works?
John Wall
Yeah, basically, I mean, you're like as an example, you could have an ADAS team at a digital cockpit.
Unidentified Analyst
Yeah.
John Wall
And, you know, they would all -- they would have their software development team, and they would have, you know, X amount of people working on the QNX part of it. They would have to buy dev seats to support that. The integration could either be done by the OEM or it could be done by, you know, bringing in a Tier 1 or it can be done with it with partners, but they're responsible for doing the integration and a lot of the services that we provide professional services is in helping with that integration.
You know, as an OS company, we're kind of the lowest, lowest man on the totem pole when there's a problem and it trickles down. When I turn around, there's nobody to help me. You know, I -- we have to solve the problems and we work with the chip vendors. So that's a lot of what we do is we help with that integration. We're not the integrator..
Unidentified Analyst
Yeah.
John Wall
But we help with the integration.
Unidentified Analyst
And then you know, thinking about the -- as these OEMs, obviously want to recruit --update these domains frequently. Is there an opportunity to you know, every update to generate some kind of additional revenue streams?
John Wall
Depending if they're adding new features or if they're adding something that's a license like ours either way our product is kind of composed as we have the operating system, we have the hypervisor. That's what people know us for, but we have a lot of middleware components as well. And we have a lot of safety certified middleware components like file systems, crypto, acoustics, communication. So if they're adding new features that triggers when one of these pieces going into an upgrade then there is an uplift.
Unidentified Analyst
So -- yeah, because I was maybe trying to do the math just to maybe frame like you know, if I -- if I said like, if I took the 200 million of revenue and said 60% came from royalties, I assume like maybe 50% -- 60% of vehicles were using QNX, maybe flow lower than that. But it's like a low single-digit per car kind of numbers. So like, again, thinking about that versus trying to go and do it yourself doesn't really make a lot of sense.
Tim Foote
Yeah, well, I admire the attempt or the math that everyone tries to do it and ask me about it. So in terms of market share, if we look at the vehicles that are being produced today, they are at different stages in their lifecycle, right? So you've got some coming to the end of their production run. Maybe they've been in production for five years. The design phase is an additional three years before, it could be a 8 years -- 10 years old design still being produced today, paying us royalties where if you think about a car 10 years ago, there's probably only one socket that was available, whereas some of the newer ones are multiple sockets.
So to try and say that we've got, you know, 50% -- 60% share of the average. As we mentioned, the conditions weren't right for multiple instances of QNX multiple years ago. So it's kind of it's a difficult one to do. I'd say you're a little bit low, but directionally speaking we don't actually give this number because it kind of affects us commercially, if we tell you how much on average people pay. Directionally is going up and to the right and that's to John's point, as there are now multiple instances of QNX.
John Wall
Yeah, like, let's be clear that the move to domain controllers is new.
Unidentified Analyst
Yeah.
John Wall
Okay. I mean, there are not a lot of cars out there today that have a lot of this good solid domain controllers. There are some there are cars today shipping where there are four instances of QNX.
Unidentified Analyst
Yeah.
John Wall
But it's not common. Its..
Unidentified Analyst
And how do you see that playing out? Because I mean, we're -- we see a lot of things like cockpit domain controllers coming in. But, you know, I think the eventual state is maybe two or three big central domain controllers, or how do you see that?
John Wall
Yes, I mean, there's domain controllers but then there is the need for high performance compute. And there's a difference in my view. I think, right now when people started talking about software defined vehicle and that that is a very popular term in the last year. It's – I don't think it was that popular. When I tried to make a presentation in Germany about software defined cars four years ago, they looked at me like I was crazy and I wasn't able to do it.
The big move right now is centralized compute. That's, you know, you'll see Qualcomm and NVIDIA have come out with chips that really talk about how they can bring a lot of different things together isolate. But if you started looking at the architecture of the car of the future, you also start to see, what they call zonal controllers, which is handling the mechatronic rim of the car. These will be high performance compute. They might not be the highest performance compute but they will be MP use that require at QNX type operating system.
You're starting to see things like smart sensors, smart LiDAR, smart radar, smart cameras, where a lot of processing is done on the edge. Again, running an operating system like QNX, which we have a number of design wins in this area and you know, as the car is collecting more data, more sensors within the vehicle, there's going to be more edge processing of that data before it goes to a centralized compute, or to an ADAS club. So we believe there is – we're just scratching the surface, we believe there's still a lot of opportunity
Unidentified Company Representative
I want you to then maybe touch on IVY, which is your edge to cloud, in-vehicle data platform, you've co developed this with Amazon. It sounds like a big opportunity here. I mean, I think you've mentioned before, like an $800 million serviceable market just by 2025, so just three years from now. And this is a fairly new product. So maybe talk a little bit about what the platform can enable, you know, and the importance of having a partner like AWS, we don't normally think about them in the context of automotive well, and maybe the business model as well.
John Wall
I'll let Tim talk about the business model. I'll talk about what propelled this collaboration. So you heard me earlier talk about the car companies trying to create an OS for the vehicle, something akin to Android and iOS. What we see or what we saw and part of the reason that we did this collaboration with Amazon is that yes, that's what they're trying to do. But we're in the Wild West back when BlackBerry, symbiosis, where there was no common ecosystem. And none of the cars are built the same, none of the cars at the hardware level look the same with I'm looking at a brand like BMW, I'm looking at a brand from Mercedes.
So the idea with IVY was really to, okay, the cars are not common. So let's create a layer, a middleware layer that will abstract the differences of the vehicle and create a common API. And we'll also do some edge processing to be able to provide insights of the vehicle so that we're not just sending up raw data. We're doing some machine learning, some processing, gaining some insights at the low level.
The collaboration with Amazon was, well, who can drive an ecosystem? You know, a big company like Amazon can drive an ecosystem. And so the idea was if you could get enough traction with this product across different OEMs, you now made it much more attractive to ecosystem developers, whether it's applications within the vehicle or applications in the cloud, to come and be innovative and create some interesting applications for the vehicle. That you know would help the OEM monetize the data from the vehicle. That was really the impetus for it.
Today, I think car companies are doing – a lot of them are doing it themselves. But I think they're missing that big picture of if you have more commonality, you'll attract much more of an ecosystem.
Unidentified Company Representative
Yeah.
John Wall
If I'm able to write an application that works with Ford, works with GM, works with BMW, and do it once as an as an app eco developer or an app developer in an ecosystem, that's much more attractive.
Unidentified Analyst
I see. So yes, so the benefit is, from an OEMs perspective is that if this can be kind of broadened out, right, not just GM having one solution and Ford…
John Wall
So the idea is this gains momentum as it goes across carmakers.
Unidentified Company Representative
Yeah. So it's like a network effect that you just end up with a much broader range of apps and services that you can -- and just take away the heavy lifting from the OEM of having to deal with the issue of gathering data in the car and get it into a usable form. Effectively, like we were saying earlier, focus on the differential experiential type things, forget about the plumbing give that…
John Wall
Yes, this is plumbing. This is plumbing. And how does it work in terms of the business model? For example, if I'm an automaker and I'm already using AWS, does this just sort of like an add on to that or do I need to -- it's a sort of a separate. So before you get into that, it's very important to note, it's cloud agnostic and it's OS agnostic. It's not tied to QNX.
Unidentified Company Representative
Yeah. So you can plug this into any cloud you want, because obviously OEMs have a big aversion to vendor lock in and we want them to be able to use this with any of any supplier of their choice.
In terms of the business model, effectively what we're supplying is the platform to enable this. The data belongs to the OEM. We have no interest in controlling the data. They own it. They can choose what they do with it. But they would pay us for using it and the difference between IV and QNX is that QNX is a one-time royalty at the point the vehicle was shipped. For IV, you would continue to use IV for the lifetime of the vehicle.
So now we're into a recurring revenue stream. And right now because the product is new, we don't want to limit our market opportunity by being too prescriptive with the model. But it has to be recurring. We're saying you can either do a pay by usage, model, or if you prefer and you're going to be a high user or you just want more security, you just pay a flat fee. But you pay it every year. And that's how we -- and that could be really transformational for us, because clearly it's -- we love the QNX model, but as soon as you move into that recurring model, and it started -- that's a potential for a hockey stick.
John Wall
Yeah. And this is not just for OEMs, because I think you have an agreement with Bosch.
Unidentified Company Representative
Absolutely. I mean, this is if a Tier 1 wants to integrate this into their platform and add value to their platform, absolutely.
John Wall
We think we have a few more minutes here.
Unidentified Company Representative
Yes.
Unidentified Analyst
[indiscernible] why wouldn't that be a better world?
John Wall
I think it would be a better world for sure. I think if -- from a development perspective and having everybody doing the same thing and being able to share across I think, would definitely be better world. You know, will we end up with that at some point in the future maybe? I don't think it's going to be Android, because Android is not a safety certified operating system. It's very much purpose built for the handset and infotainment type functions.
But I mean, over time, you could see a homogeneous platform that is OneOS.
Unidentified Analyst
I mean, I can't help but think that Android, Google, Alphabet will do whatever it takes to get there.
John Wall
Yes. I'm not debating whether they could do it if they wanted to do it. I'm saying it wouldn't be Android. They develop something else.
Unidentified Analyst
Thanks. Help me understand what the process for safety certification as you keep putting that out there as this barrier, how long does it take for them to get over that? And give us a lot more information there. Thank you.
John Wall
Yes. So, safety certification is part of the development process. It's how you develop your product. It's how you create your design docs. It's how you do risk and hazard analysis. It's how you do code coverage. It's how you do testing of the product. so, there is a lot of work that needs to be done upfront before you actually start typing on the computer your code.
So, it's -- it can take any -- on a new product for us, it can take anywhere -- it can add a year and a half to the development process that we already have in place. The real secret with safety once you achieve functional safety is that you have to be able to analyze any changes you make to the product going forward.
So, if you have a bug, you have to be able to look at that component and say is that a safety impacting bug? And that means doing an analysis of exactly what the fix is for that bug? How that bug will be addressed? And whether it introduces a problem with your safety case?
So, you have to be very intimate with the components that you're analyzing, which means since QNX, develops our operating system, the developers are all at QNX whether it's file systems, the kernel, process manager, other parts of our product, we are able to do that impact analysis.
We're also able to be very targeted when we make fixes, that we're only fixing a particular problem and we're not fixing or adding other pieces to it that might impact the safety certification of the product. So, one of the challenges with open source software is you have less control over what's getting put into the product.
So, for instance, if I'm using Linux, and I have a very, very specific bug, I may have to grab a patch that incorporates 40, 50, 60, 100 different fixes that I don't want. And so it's very hard to manage that. And with Linux, you have very fast development, and you don't have the collection of all those artifacts to be able to do the safety certification. I mean, it's a complex process.
Unidentified Analyst
I would take one more here.
Unidentified Analyst
Thanks. Just want to go back to the Wind River comparison for a little bit, are you guys generally after the same sockets as you define them in terms of the -- like how you define that that opportunity?
And then just going back to what you said you kind of said, you see an advantage in in compute -- performance compute for your product versus them. We heard earlier this week that that the management team thinks very highly of their own product and says like they think they're years ahead of everyone else. I'm just trying to understand like, what is the case they're making there? And why -- is this a threat to you guys over the coming years?
John Wall
Yes, so I think we take all competition seriously and we don't discount them, but I'm going to be honest with you when we're in competing for awards, we don't see them much and we don't lose to them. We know -- I'm going by the feed, I don't have VxWorks on my desk at home. I mean, they don't give it to me. But we're going by the feedback that we're getting from our customers, why they're picking QNX. And you know, we have a new product coming out at the end of this year, a new version of our operating system that we believe will be the fastest operating system on the planet, as far as scaling with the next generation processors cloud compute and the highest level of determinism and that you know, that's really what defines in our toss. People will often ask, what's a real time operating system? A real time operating system is being very deterministic and how you react to events. You react the same way every time, and obviously fast, not slow, but there's no variability. If I say, I can react to something within a millisecond or within a microsecond, I can do it consistently no matter what's happening on the system. But they believe they’re years ahead of us. I've talked to people adaptive, they believe where they’re years behind.
Tim Foote
And if we just look at the relative footprints in auto, I mean, we're in over to 115 million vehicles and I believe 30 million for active. So, just in terms of that focus and the level of interaction with the OEMs and understanding what their needs are, we're kind of very tuned in to what they are.
John Wall
And let's be clear, we still work with Aptiv, not to visit Tier 1. We have designed wins with Aptiv. So, I guess we're frenemies.
Tim Foote
All right. I think, we're going to have to wrap it up there. But thanks so much, guys.
John Wall
Thank you.
Tim Foote
Appreciate it.
For further details see:
BlackBerry Limited (BB) Presents at Wolfe Research Global Auto, Auto Tech, and Mobility Conference 2023