BlackBerry Limited (NYSE:BB) Barclays Global Automotive and Mobility Tech Conference Call November 30, 2022 2:00 PM ET
Company Participants
Tim Foote – Investor Relations
Mattias Eriksson – President of IoT Business Unit
Conference Call Participants
Dan Levy – Barclays
Dan Levy
Great. Thank you, everyone, for joining. I’m Dan Levy. I lead Autos Research Coverage at Barclays and glad to have you as we continue the Barclays Global Automotive and Mobility Tech Conference. We’re touching on now the opportunity in automotive software and we have with us BlackBerry that should give us really interesting insights with us is Mattias Eriksson, who’s the President of BlackBerry IoT as well as Tim Foote, who leads Investor Relations effort. So Mattias and Tim are going to run through a series of slides and then we will go through a series of questions, fireside chat style.
Anyone who wants to ask questions, please submit the questions via the viewing pane. There is an option there and I can ask those questions anonymously or you email me directly dan.levy@barclays.com, d-a-n.l-e-v-y-@barclays.com and I’ll ask those question-and-answer. So Mattias and Tim, thank you so much.
Tim Foote
Perfect, Dan.
Mattias Eriksson
Great. Thanks, Dan, and let’s bring up the slides. Welcome to the BlackBerry IoT session. I was asked to provide just a couple of basic slides here to frame the real discussion. So I’ll go through this fairly quickly. Hopefully, most of you are aware of the things that I will cover here but just for context. We have a Safe Harbor statement. And just as a recap for everybody here.
BlackBerry IoT is an investment in the rapidly evolving intelligent B2B stack. Simplistically, at the top level, the business has secular trends driving the growth of these segments, essentially more devices, B2B devices at the edge. Each of these devices are becoming more intelligent, more compute, connectivity, sensors, memory, etcetera. And as they become more intelligent, they almost, by definition, end up being software-defined and that’s where we come in. Our core offering in the IoT BU is our QNX advanced microkernel real-time operating system that typically sits at the bottom of these advanced edge B2B devices in between the SoC and the rest of the layers of that sophisticated stack. I think not just cars but robots, high-end medical equipment, industrial automation, etcetera, etcetera. We are very focused on automotive. And we are focused on automotive for one simple reason.
We believe — and there’s plenty of evidence for it but there’s no more advanced IoT edge compute device in the modern car today. There’s more compute, more connectivity, more sensors, more memory, more cloud interaction in the car than in any other device. And by winning and satisfying customers in the automotive industry, we believe we get a head start on the rest of the segments that are, quite frankly, moving in the same direction when it comes to advanced software stacks at the edge.
Quickly or I should say briefly, you are familiar with our footprint in automotive. We’ve been serving automotive for quite some time, originally as an integrated vertical IVY stack, all the way from the chipset to the HMI, that changed 5, 6 years ago, we pivoted and we are now deploying for all compute domains as foundational software. 250 million vehicles, 45 automakers, 10 out of 10 of the largest auto OEMs users, 7 out of 7 of the largest Tier 1s, 24 out of 25 of the EV start-ups that matter, etcetera, etcetera. We have a strong footprint in automotive and is getting stronger. What has happened in the last 12 months after we set this BU up, we have had tremendous progress and the new structure has allowed us to be very focused in terms of execution and investment, etcetera, etcetera. On the left-hand side, I have a couple of bullet points for you. The first 2 quarters of this fiscal year, we grew 23%.
You should put that in context obviously, a difficult macro environment. We have guided long-term for the next 5 years, roughly 20% CAGR organic growth. So we feel good about the first couple of quarters. Obviously, you know this is a high operating leverage business. So the margins are high. Maybe some of you are not aware of what is arguably the most important leading indicator for this business, the design win. So you know how it works — we work with various types of OEMs. We get a design win. And then as that is being ready for production, we get development seats and services and so forth. And then there’s a long tail of royalty coming as the production ramps — so wins today, they are royalty revenue for us, ’24, ’25, ’26 and beyond. In the first half of fiscal ’23, the first 2 quarters, we had more dollar design wins in those 6 months than we have ever had for any previous full year period very significant momentum in spite of the macro turbulence.
And that really comes from this secular trend of the evolution of the software-defined car and all the work that is going on with regards to software for the OEM. So we reported out the backlog, the royalty backlog, I should say, Q1 fiscal ’23, 560 million. And one of the things that we’re going to talk quite a lot about at CES is we spent an enormous effort last year with many customers and partner outlining a revised 5-year technology roadmap. And I know it sounds a little bit weird to talk about 5-year road maps but that is the length in which you need to plan in order to be at the cutting edge here. And we’re going to talk quite a lot about the progress with that at CES this year and I hope to see some of you in January in Las Vegas. Just to give you a couple of examples. We are — as assessing the slide here, we have validated the road map.
We have made a lot of progress over the last 12 months on things that have been in the cooker for quite some time. We have the next-generation kernel coming up. The microkernel doesn’t change that often. In the last 35 years, I think we’ve changed it 3 times. This is a re-architecting of the kernel to cater for linearity beyond 8 cores. So as you scale the number of cores and you get more compute capacity, the performance requirements are very different. And we’ve been working on this for roughly 3 years now. We have been in early access with key partners for a while. We believe we will be GA this in the second half of next year and we have some demos at CES. If you’re following us, you have probably seen a couple of the announcements this week at reinvent [ph].
So one of the things that we have been working on for a while and we continue to work on for the next 12, 18 months is the clarification of QNX, extremely important in many different contexts and we can come back to that during the — Q&A. That was announced formally this week. And then obviously, IVY has been iterating the POCs, as we have announced previously and we are making tremendous progress on that road map and we hope to be able to have a GA announcement around IVY for next year at CES. Let me stop there. A little bit of context before the Q&A.
Dan Levy
Great, I want to start because I think we know that the trend of software in automotive is just becoming more and more pronounced. You’re sitting right on top of that trend. You’re very involved. So maybe you could just help us with a bit of a visualization?
Mattias Eriksson
Sure.
Dan Levy
Walk us through what a typical vehicle software stack looks like maybe all the way from the compute, the hardware all the way up to the application layer. What are each of the layers and then where do you sit? What are your solutions in each of those layers?
Mattias Eriksson
Yes, great question. And let me start at the top level and then we can come back and do clarifications. So simplistically, at the top level, what is happening is obviously all OEMs and they are all at various stages of evolution in this decade-long transformation that is ongoing on the software side. All OEMs are stepwise moving from the decentralized ECU architecture that we’re all familiar with for many years which became untenable. 100, 150 ECUs all across the car, single function, maybe a couple of functions, towards what — most people have started calling domain compute. And when I say domain compute, I mean things like there’s a digital cockpit domain that has a compute stack. There is an ADAS domain that has a compute stack. There is for body and chassis, a compute stack evolving.
There is — for drive train, electrification or hybrid or whatever it is that you’re doing, there’s a sophisticated compute stack with a layered software implementation that is involved. The gateway is becoming more compute-intensive. And ultimately, at some point, this will land in high-performance compute, whatever you mean by high-performance compute. But it’s a process that runs over many years here and all the OEMs are at different stages of maturity. In the decentralized ECU architecture, there was really no need for something like QNX. So that implementation for single function, we had basically no opportunity. In all of these domains and the most important ones, the most sophisticated ones today is the digital cockpit and the ADAS compute domain. We are winning. We are sitting at the bottom of that stack close to the SoC and we provide the foundational software layer upon which they’ll then layer various types of middleware, they lay the application layer. They might have what they call a CarOS sitting on top of us. And we are basically supporting all the OEMs and their partners in this evolution. There are many misconceptions about who we compete with and who we partner with.
And so, let me just take the most obvious ones upfront here. So first of all, we do not compete with Google. For those of you who — have been following us for many years, you know that when we had that integrated IVY stack obviously, we were competing with them because we went all the way up into the HMI layer. We realized many, many years ago that, that was an untenable situation. So we pivoted that vertical stack to the foundational software for all compute domains. Typically, when you see an Android implementation for digital cockpit, we sit underneath with our hypervisor and we typically also have an advanced virtualization framework, the Virtuo [ph] implementation on top of that hypervisor to help the peripheral sharing and so forth. We do not compete with the chipset partners. So if you take the ADAS stack as an example, there’s lots of press in general about the progress of NVIDIA for DRIVE and Qualcomm for ride. And when you dig underneath that, you realize that every NVIDIA DRIVE implementation has our operating system after that implementation. Every Qualcomm Ride implementation has us as part of that implementation. So we work very closely, obviously, given where we sit in the stack with the chipset partners.
And then the one that is probably most confusing in general because it varies so much by OEM, it’s this notion of the CarOS. And in general, our OS is part of those Oss. So if I take the last publicly available announcement from us, you take CARIAD and Volkswagen and VW.OS. QNX is part of that OS. So you have sort of a nested your multiple OS is in the car and they are nested and that’s part of what is a little bit confusing.
Dan Levy
Great.
Mattias Eriksson
Missed off there, I might not have covered all the questions.
Dan Levy
You know, that’s a helpful visualization. So you have the RTOS and you also have the hypervisor piece as well?
Mattias Eriksson
And middleware on top of that and frameworks depending on what the partner wants to implement, yes.
Dan Levy
Right, so where are we in this transition, right? So you’re saying there’s no need for QNX in a decentralized ECU structure but is it fair to say that we’re still very early into this software transition?
Mattias Eriksson
Yes, very early is unfair because people have been working on it for quite some time. But there is definitely an inflection point — there have been an inflection point in, I would say, the last 2 years. I mean, money is pouring into this from the OEM and the technology partners because everybody realizes that the future is going to be software-defined and it’s going to play out at least over the next 10 years. So, if I look at the design wins that we have for launches ’24, ’25, they still have not approached full HPC implementation so all the OEMs are working really hard on this. It’s a very difficult problem to solve. From my perspective, it’s probably the most interesting software problem across any industry at this point in time, how to develop a sophisticated software-defined vehicle. So this is going to continue for quite some time and it varies significantly across different OEMs, where they’re at.
Dan Levy
Okay. Let’s talk about the involvement of the OEM, because I think we know on a simplistic level, there is a push by — everyone says okay, the legacy automakers want to copy Tesla. Tesla controls a lot of their, own software. They do a lot in-house? They do their own middleware, etcetera. Give us a sense why this is not feasible. And maybe you can talk a bit to what has happened at VW with CARIAD that VW tried to bring a lot in-house and it’s been choppy?
Mattias Eriksson
Yes, so I mean, if you take a step back, I wouldn’t say that it’s impossible but it’s a very, very difficult problem. I mean being the CEO of an OEM today is an incredibly difficult job. You are hit from all angles by a number of different really big megatrends and solving all of them simultaneously is very, very difficult. Why is the software transition particularly difficult? Well, it’s particularly difficult because software is one of the most nonlinear functions there is if you look at a company. You can have one software engineer or you can have 100 software engineer, that one guy can make — have a higher output than the 100 people that you hire. So it’s not just throw people at the problem and it’s fixed. So I do think it’s fair to say that most OEMs have a strong interest in controlling the software stack as this transition goes through.
And that is a major shift, not just on the technology side but also on the commercial side. And for us, for example, there is more and more OEMs requiring direct commercial agreements with us. And I think that’s good because I think that makes it easier for us to help them and support them and so forth. That said, on this transition and this is actually at the heart of our value proposition. If you make a strategic call that somewhere along this 10-year journey I am going to own the entire software stack. And I’m going to do that because a big chunk of the value creation and value capture for the car is going to be software-defined. As I go through that journey, how do — where do I focus my resources, my precious software resources. And our part of our proposition is you should not focus on the plumbing of the software stack.
You should take your precious software resources and focus them higher up in the stack where you drive direct consumer benefits, direct partner benefits, things that customers can see and touch and so forth and let us worry about the underwear and how to interface with the chipsets and so forth and that resonates. It resonates because it reduces complexity for them quite significantly. It improves time to market. It lowers their overall cost. It just shrinks the problem to something more manageable and that is a big trend for us. We see a lot of people that maybe initially had said, I am going to do all of this myself saying, why am, I actually trying. It is hard and slow and it’s not necessarily differentiating for me. So let me focus on where it matters. You had a sub question I think I missed — probably be up [ph]. Sorry — yes, I asked a lot yes.
Dan Levy
No, that’s great. So is it fair to say then the — you focus on the plumbing, the automakers focus on the application layer or what generally touches the customer. Is it fair to say that with the exception of Tesla to maybe a little bit of myth-busting here? Even though OEMs want to own the software stack, aside from Tesla, there really won’t be many others or any others that will want to own or do their own middleware, ADAS, hypervisor, etcetera?
Mattias Eriksson
Well, so as with all interesting problems, the devil is in the details. So I think you can control the software stack without writing all the code. And in our particular example, one you control the software stack or RPs. You sign the contract directly with us you put pressure on us you iterate our road map and give us a seat at the table. So and you do that at multiple layers in the various stacks. Just because an OEM has made the strategic decision to own the software stack, it doesn’t mean that there won’t be a mix of commercial software pieces, open source pieces, some code they write themselves, some code comes from someone else. There are different ways of controlling the stack. It doesn’t necessarily mean that you need to write all the code. And Tesla is such an unfair example because there’s such an outlier. I mean it is obviously truly amazing what they managed to achieve. There are very few Tesla’s in the world.
Dan Levy
Let’s talk about the competitive environment. Google is not a competitor, the chip players are not competitors but we have seen, for instance, Wind River, they’re focused on RTOS and Greenhill as well. So maybe you can give us a sense of the competitive environment. Is it a market of several key players? And how does Aptiv’s acquisition of Wind River change the environment?
Mattias Eriksson
Yes, so let me talk about it broadly and then you can follow-up with specifics. So, the picture that we have been making for the last 7 years supply so there’s sort of 2 camps of competition. We had the traditional embedded software vendors, Wind River and Greenhill fits into that category. They are [indiscernible] not so much motor quite frankly we don’t run up against them that much anymore but our side. I mean, Wind River – is tremendous sort of break the car and the growth we have for digital cockpit and ADAS, the 2 most sophisticated stacks is a good indicator of how that is playing out. Then you have the other camp which is you go back, again, 5 years or so, it was almost a religion among everything is going to become open — please help us fix this. So I think those are sort of the 2 categories of competition.
And then, again, as you get into specific industries and deals, there’s obviously much more to say. And then I think you had a question around the Aptiv acquisition. So Aptiv been a partner and customer for us. There’s no doubt in my mind that this will affect our future business with Aptiv. We still have business with them but obviously, that is going to be affected. It’s not material for us in the overall business so from that perspective, not a big deal. I think Wind River has shifted focus over the last several years on the TPG. I think they have spent a lot of time on studio. They spend a lot of time on their cloud offering, maybe invested a little bit less than the performance in the embedded software. I don’t have the details, so I can’t really comment on the details here. And I think it’s a very important area but it’s also an area that is very, very difficult to manage.
And the reason it’s a very difficult to manage area is because you have the cloud majors and hyperscalers like AWS and Google and Microsoft and so forth. That already provides a portion of that offering in the cloud. I think that is a challenge for them and we’ll see how that plays out. In addition, I think if I look at the traditional value chain of the automotive space, where you have a very powerful, strong sophisticated sourcing machine relying on Tier 1s and Tier 2s. I think now if more and more of the car is becoming software defined, I think one of the things the OEM really wants is they want to separate hardware and software.
So providing an integrated solution, adding foundational software and development tools and so forth to a hardware business and then expecting the OEM and say, yes, I’m going to take all the strategic stuff from one vendor. I think there’s an issue — but the future will tell.
Dan Levy
And as far as the competitive environment, just to go back a little bit, another announcement we saw was GM forming a partnership with Red Hat and that’s the help on in vehicle operating systems. Is that competition or is still — there’s still place for QNX in that stack?
Mattias Eriksson
Yes, so that will fall in that second category. So first of all, I don’t have any insider information around what exactly they are doing and so forth. A lot of the public statements are, quite frankly, a lot focused on tooling and so forth. And I think that makes sense. Obviously, that would not be a problem for us. I think at the end of the day, someone like Red Hat and IBM can obviously take an initial resource approach, branch the open source code line and safety certify and so forth. I think they have publicly stated they want to go to ASLB, ASLB is not option. You need to be at ASLD. They have yet to certify it. Let’s see how it plays out. I come back to the fundamental paradox.
So as you drive into the highest requirements for safety and security, in order to get there, you need to branch the software. And if you’re branching, the software, you lose the benefits of open source to a large extent. And I think that is the fundamental question. What we are focused on is we believe there is value from taking advantage of all the benefits that open source has, as you build the entire stack and then taking advantage of the bits and pieces in that stack that require specific capabilities, specific IP and rely on commercial software vendors for it.
We feel we have pretty good traction at present with that approach. And so far, we’re going to continue down that path. Maybe you can talk a bit about some of your award wins and what’s driving this or what’s the path to grow from these. We know that you had an award with BMW. You’re working with CARIAD. I mean you have — I think you showed unless you’re working with everyone in the industry but maybe talk about some of the recent award wins and sort of what the growth path — broaden the offer or — what the state is to get the scope of your offer [ph]?
Mattias Eriksson
Yes, so obviously, as you saw in the stats, obviously, we work with a lot of players. And at the heart of our business model is we are a component ingredient provider and it’s not about us, it’s about the customer that we are in that stack. It’s not necessarily public knowledge and so forth. And quite frankly, a lot of our contracts state that we cannot without their written permission, tell you which deals we’re in and so forth.
What I think you can do is if you look at the overall market and you look at the progress that we are announcing, if you just take the statement I made initially here, more design wins in the first 6 months than we’ve ever had for any previous year, you realize that we are in more stacks than we are able to announce. The growth path for us is very simple. It’s — if you focus on automotive, it’s the statement around more sockets and more layers. So as the compute domains become more sophisticated. There is room for QNX. There’s a market opening up. We believe we have a strong proposition there. Depending on the strategy, we get just the basic foundational RTOS or hypervisor or we get some middleware and frameworks.
And then, I think over the next several years, we will see a third one emerging which is if I have trusted QNX with my 2 main compute domains, there is benefit of using QNX in other commute domains, even if I can find a solution that is similar because everything QNX does is needed for that particular compute need. Because at the end of the day software developers is a scarce resource and I want to reduce the friction and make it as easy as focus on the sensor data and what they do with stuff from the digital cockpit domain. This across domains it is always agnostic, it is hardware agnostic. So it’s a very clean solution and it does not make claim to the date that we are not billing data leg, the data is the OEMs which is obviously very important to them. So we are not trying to monetize the data.
What we are trying to do which I think also resonates is that if they use that sort of cloud edge platform they can instead of again building plumbing for sensor data processing. They can focus on deploying the damn models. So they can have all their engineers focus on how do I deploy a machine learning model to get the use case. How do I deploy one use case. They can start working on that as opposed to building foundation to work on what’s going to generate for the customer right.
I think that generates. It’s a complex solution because the architecture is evolving and many people have already done some work and so it is difficult to abandon work that you have done to go from something new. But we feel great about the progress. You might have seen some of the showcases posh connected where we showcased with Nestle [ph] and actually car.
At CES we’ll again do more work on this and actually an important footnote is as we put QNX in the cloud, we are also putting IVY in the cloud which allows to accelerate the ability to test and write implementation for these synthetic sensors and machine learning models. So we feel good about.
Dan Levy
Right, let’s grab with because I know we’re close to be at time here. For those who aren’t familiar maybe you could walk us through the [indiscernible] per vehicle fee walk us through the revenue model to understand QNX?
Mattias Eriksson
So obviously there are some ins and outs here and things are changing little bit. But simplistically for the core business the QNX software development platform we have three sources of revenue. As you start building sophisticated software stack you need the development tools and you buy license for making use of the tools that we provide to later on making use of QNX. So we get typically upfront payments for that usage. Some customers when they do that they also want a little bit services could be on security, safety security, consultations, help particular implementation or whatever that continues to building as they make the decision. So, when they have made the decision to okay it’s going to be QNX then typically buy even more development seats from us and they start ramping the actual integration and that’s when our service organization kicks in and we help in the area that they want to help and so forth. But the service piece is not a standalone business for us it’s an enabler to get the software platform.
And then to your point when they have grown through all the work to get ready for production and production kicks in we get a fee for every car that is produced. And it’s important it’s not every car sold it’s every car produced and that is helpful in the current macro environment and that typically runs for many, many years. And throughout those many, many years what also happens and will become more important in the future is they maintain the development tools that they have. They require some extra service. And we believe that at some point will be a growing business because people are going to manage the car throughout the life cycle and they’re going to make changes to the software and they’re going to need additions and so forth. But 3 sources, development seats, services and then royalties. That’s how the business flows.
Dan Levy
Okay. We’re slightly over time. But I want to thank you Mattias and Tim very insightful, very interesting, especially as the theme of automotive software is becoming much more relevant and much more [indiscernible]. Let’s close the session [ph].
Question-and-Answer Session
End of Q&A
Be the first to comment