BlackBerry Limited (BB) Earnings Call Transcript & Summary

Great. I'll kick off. I mean, a little early, but might just kick off the next section. I think -- with BlackBerry. I think corporate access was a little surprised when I signed up BlackBerry to attend. I think most people are just not aware of how much auto exposure the company has. But if you go into cars today, QNX, which is the BlackBerry system is the core operating system in the vehicle. And the company has gone through a lot in the last year with a new CEO, I guess, 1.5 years ago. They re-segmented the company. They've divested some noncore businesses and instituted $150 million restructuring program. So they've been quite busy. I think the business now is more focused actually on the QNX auto system. So it is kind of an interesting sort of auto play that people don't really fully appreciate. So for that reason, it's actually probably good we'll kick it off with a presentation kind of walking through the business, and then we'll open up to Q&A after. So maybe you guys want to...

Okay. Good morning. Delighted to be here today. My name is Tim Foote, I'm the CFO of BlackBerry. And I'm joined here by Mattias Eriksson, who's the President of our QNX division. Just to kick off, safe harbor statement. We are actually in our quiet period, and we report in a couple of weeks' time. So we won't be talking about the current quarter that we're just about to report on. As Colin says, there's been a lot of change at BlackBerry. And ultimately, it's probably a good place to start, and I'll try to keep this brief, just an overview of what we are now. So we are a software company, just for the avoidance of doubt. And it's -- we are divided into two primary divisions. So the first is our QNX division. It's very focused around automotive and some adjacent verticals, some industrial verticals, which Mattias is going to talk to as soon as I'm out of the way. It's around about half of our business, and it's been growing very nicely. And I'll come back to the financials a little bit later in the presentation. The other side of our business, also software, really points to the legacy, the BlackBerry legacy. Obviously, the phones or the keyboards were cool, but the security that ran through those phones is another compelling reason why people used BlackBerry phone. So today, we're focused on secure communications. It's obviously hardware agnostic. We have 3 key products in there, unified endpoint management, controlling data flow to endpoints securely. We've got our AtHoc critical events management, which is like a highly secure advanced version of AMBER Alerts. And then finally, we have our Secusmart business, which is encrypted voice and data, really military-grade encrypted voice and data. This business, Colin mentioned a divestiture of a business called Cylance, which was an endpoint security product. I'll come back to that a little bit more. But ultimately, following that divestiture, the Secure Communications business is a very solid, stable source of EBITDA and cash flow for this business. I'll briefly mention we also have a Licensing business. Being in this business for around about 40 years, BlackBerry generated some 40,000 patents. It's one of the biggest portfolios in the world. And just over a year ago, we sold the vast majority of the noncore patents to a company called Malikie. And as they start to monetize those assets, there's potential for BlackBerry to generate some profit share anywhere between 0 and $700 million. So we see this as like a call option to the story. Obviously, it's kind of not the primary focus, and we won't spend any more time on it, but it is kind of an upside that other companies may not have. We're very diverse in terms of our geographies. You can see here, revenue is kind of spread between North America, EMEA and Asia Pacific, and that's very much the case also for our QNX business. So I'm going to hand over to Mattias, who's going to run through a part of the story that everyone I speak to is very excited. We're excited about, too, which is the QNX business.

Great. Thank you, Tim. Good morning. Thank you for spending the time with us this morning. My name is Mattias, and I run the QNX business. And let me start with the key points here. I have 10 minutes, 5 slides for you basically to frame the Q&A and the discussion afterwards. I want to tell you a little bit about who are we and what do we do. I want to talk about the portfolio. I want to talk about the customers and partners we have. And last but not least, I obviously want to talk a little bit about the investment thesis, why are we investing in QNX. So I'll go fast, and hopefully, we can come back to many of the topics because there's a lot to unpack here. So if someone twisted our arms and said, summarize in one sentence, what is QNX? Well, QNX is a leading foundational software provider for what we call advanced edge compute. A lot of buzzwords. Let's start with foundational software. What does it mean that we're a foundational software provider? Well, we do not do end-to-end stacks. We do not touch our customers' customers. We do B2B only. We typically sit underneath all the great work that other companies are doing. And in general, for advanced compute stacks, we provide hardware-software separation and a number of other things. In automotive, it's particularly complicated in terms of how people define vehicle platforms, OSs and so forth. And let me clarify a couple of common misunderstandings upfront. We do not compete with Android in digital cockpit. Typically, every advanced Android implementation in the world, we sit underneath. It's our hypervisor that touches the silicon. If you look at the ADAS stack, we do not compete with NVIDIA and Qualcomm for their software packages for ADAS. Qualcomm Ride, NVIDIA DRIVE, we are part of that software package. So if you see a win with NVIDIA DRIVE or Qualcomm Ride, it's QNX sitting in that software package. Advanced edge compute. What do we mean by advanced edge compute? Well, in particular relevant for automotive and the changes that are going on, we sit in the stacks that have the highest requirements for performance, safety, security and reliability. That typically means an MPU. So multi-core silicon, think NVIDIA Orin 4, think Qualcomm Gen 4, Gen 5. We do not cover MCUs. So single function compute in the car, it cannot run QNX, and that is being consolidated, and that's opening up an opportunity for us. Leading, what do we mean by leading? Well, by leading, we mean that we have the great fortune of working with the best of the best from silicon to cloud across all of automotive and across the other key segments for us that are industrial automation, medical OEMs, robotics and anything that looks like a car, commercial vehicles. So one sentence, leading foundational software provider for advanced edge compute. Customers, portfolio and value proposition. So starting from the left here, what is our portfolio? We provide a software development platform for advanced edge compute. At the heart of that platform is our advanced microkernel architecture for the RTOS. Around that RTOS, there's obviously development tools, there's frameworks. There are middleware layers depending on the industry vertical we're talking about. There's a cloud solution that we have a lot of traction with, providing bit parity between the cloud software stack and the edge compute stack, very, very important for automotive going forward. We have a lot of certifications. There's a lot of -- in the industries we serve, safety certifications required, security certifications that is part of the core of what we do. We provide that for the foundational software. And last but not least, if a customer requires some extra help, we wrap this in professional services so we can help them bring that foundational piece of the software stack up. We do not do services as a business. We do it to deploy the software development platform. The next sliver here on the slide, leading customers and partners. This is focused on automotive. So let me focus there. As I said, we have the greatest customers and partners in the world. We work with everybody. Think of us as the plumber from Switzerland across all stacks at the bottom of the stack. It's about our customers. It's not about us. So most of our design wins and references, they are not public. The stats are on the right-hand side of the slide, 10 out of 10 of the leading OEMs in the world. 7 out of 7 of the top Tier 1s are our customers. 24 out of 25 of the EV companies are our customers, the top EV companies. The one missing is Tesla, as you might understand. 9 out of 10 of the leading medical OEM equipment, think surgical robotics, for example, are our customers. We have 255 million cars on the road, and we have been in this business for quite some time. What does the customer buy from us? And that's the bit parity at the bottom there. So the value proposition is pretty straightforward. What they want from us, and there's obviously differentiation on the needs, but what they want from us is, first of all, faster time to market. We help them come to market faster. That's a pretty relevant topic for OEMs at present. We lower the lifetime cost of maintaining an architecture for an OEM. And last but not least, they come to us because they want to guarantee that they have safety, security and reliability in the stack. A big portion of what we do is obviously future-proofing that. These contracts and customer relationships we have, it's a little bit like a marriage. You get into bed for 10 years plus. So trust and longevity is very, very important, and we're very proud of being allowed to work with the leading companies here. How do we frame the investment case? Well, this is how we think about it internally, and I think it's equally relevant externally. We have two components. It's the fundamentals of the segments that we serve. And these are large segments, but we are very focused on a particular sliver of these segments. The large segments, and we talk about the trends in those. And then there are our -- is the position that we have in these segments. Going from left to right. The segments that we serve, take automotive because I think all of you are focused on automotive, the transition to the software-defined vehicle is a large segment that has structural growth for the next 5 to 10 years. We're in early stages of this segment. That is an important part of the investment case. We are in growing markets. The trends in these segments, both on the commercial side and on the technical side, are literally pushing customers to our core value proposition and differentiation. The requirements for performance, safety, security and reliability are going up. It's not us trying to sell that. They are literally asking for more, and it's directly in the core value proposition of what we do. The third one, going to the right here, we are -- if you take automotive to begin with, we are the clear segment leader in automotive for foundational software. And I'll come back to some stats about it in a second. We have very unique capabilities. We have unique IP, over 1,100 software patents associated with foundational software and growing. And very, very importantly, given the marriage characteristics of these customer relationships, we have trust. We have never missed an SOP. We have never delayed an SOP. Customers know that when QNX commits, we will deliver. Last but not least, these relationships extend beyond the commercial contract. To give you an example, last year, we had the biggest product launch year in a decade for the company. Our next-generation software platform, SDP 8 was literally developed based on requirements for NVIDIA for automotive. 3.5 years of development work, NVIDIA basically came to us and said, "Look, you need to find a way to ensure that the multicore silicon that we're deploying can scale linearly also in the software stack." Our previous platform did. We worked on it for 3.5 years, and we now have a foundational software solution that scales linearly with multicore, unique in the industry. So deep relationships extending beyond the commercial aspects of the contracts. Let me try to give you an example for automotive. And if you're covering automotive, you have seen pictures like the one on the left here, 100 times the last 5 years. Just to recap for you the trend for the architecture, distributed ECUs, many of them being single function, 100-plus of them in a car. At some point, it's going to be centralized compute. We're going to have a high-performing silicon and a new architecture for the car. All the OEMs extend across the spectrum of their transition into this space, but we are in early stages. What the right-hand side of the slide is trying to do is to give you an indication of where we're at just by numbers, and there are many other indicators here. So above the bars, starting at 2023 and going to 2030, you have light vehicle production. So let's say it's roughly 90 million a year across the world, light vehicle production. The bar is color coded, and it maps to the architectures. Gray is legacy architecture. So the gray taking, let's say, 2024 as an example, 80% of all cars that shipped last year were still in the legacy architecture, meaning they haven't even reached the domain structure. There's been lots of talks, lots of plans, lots of hype, but this is the reality. The 80% last year of the cars shipped were dominantly distributed compute architecture. QNX serviceable available SAM is the light blue and the dark blue. So starting in 2024, roughly 20% of just automotive, our lead segment, just 20% of the cars are sort of available for us to deploy in. You look at 2030, and you see it has flipped. Let's see if that plays out, but the trends are clear. That's where it's going. So there's structural growth in advanced edge compute for automotive, and we are riding that wave. Obviously, our SAM is growing significantly here. You can see some of the stats from 2024 below. So if you have a consolidated digital cockpit stack, meaning you have really centralized the digital cockpit architecture, we sit in 16 out of 20 OEMs today, but it's a small portion of all cars that are shipping. For ADAS, it's similar. We sit in 15 out of 20 of the advanced centralized compute stack for ADAS. Okay. Let me summarize and let's see which ones you want to come back to in the Q&A. So starting in the top left-hand corner. So QNX is a bet on the evolution of edge compute, advanced edge compute. The interaction between cloud, everything was about the cloud the last 10 years. There's a lot of discussion of the hybrid compute structure for automotive, in particular, but also for other industries over the next 5 to 10 years. Secondly, the transition is in early stages. In our conviction, automotive is the lead edge compute segment. There is no other edge device that has more compute, more memory, more sensors, more connectivity, more requirements on the legislation and restrictions around -- the car is the tip of the spear for edge compute today, and we are winning in the car. We believe that we'll have spillover effects, and we can talk about examples for that, that we already have. Number three here, we are winning in automotive. And again, that is -- that has been the strategy for the last 4, 5 years, win in automotive and others will ask the same questions. And last but not least, within those segments, we do have unique IP, unique capabilities and unique relationships with our customers. And that is the investment case for QNX. Back to Tim.

Okay. Thank you, Mattias. A lot covered there. I'm going to keep this very brief, so we got some time for Q&A, Colin. So you mentioned earlier, Colin, that we've been through a transformation in the last year, and that's not an understatement. So I'll summarize it in three buckets. So the first is focusing on our fundamentals, which I'll get into more, a significant change in our financial profile, profitability and cash flow. And then finally, a pivot in our direction for capital allocation. So one of the first things we did was to divide the business into two virtually autonomous business units. So the QNX division that Mattias leads now has its own team fully focused on this great opportunity that Mattias has just described. And likewise, for the Secure Communications division, we have a stand-alone team there that every day eats, breathes and sleeps secure communications. That has been really helpful in improving our focus and ultimately, our results. So you can see that QNX in terms of revenue has been growing at over 10% CAGR over the last 4, 5 years. So it's growing nicely and it's growing profitably. You can see that its margins last fiscal year were 25% adjusted EBITDA. So it's well on the way to becoming a Rule of 40 business. The backlog is more of a look to the future. So this is the royalty portion of revenue, which is the largest piece, but this is about designs that we've already secured, what's the future expectations for revenue from royalties from those designs. That's been growing even quicker. So that CAGR is 23%. And currently, it stands at 3.7x FY '25 revenue. So in addition to the focus, we've taken a lot of cost out. Colin mentioned we've taken $150 million plus out of the business in terms of reductions and the increased focus. And almost paradoxically, you take out that much cost, you're left with a better business. Both sides of the business are performing better. What you can see from FY '23, BlackBerry was burning over $260 million in cash per year. We've now flipped that on its head. And this past fiscal year, the Q3 was the first time in 3 years that we turned a positive cash flow quarter. And then the following quarter was the highest it had been for 4 years. So significant change in our financials. And then in terms of balance sheet, we see a very strong balance sheet. So a couple of years ago, we took down the amount of gross debt. We are net cash positive. We've got over $410 million in cash and investments as of the end of prior quarter. We mentioned the Cylance sale, moving away from this big bet in endpoint security. We still have another $40 million of cash and also 5 million shares in the purchaser, Arctic Wolf. So finally, in terms of capital allocation. So having bought Cylance for a big-ticket price, tough market, it was our #1 capital allocation priority. Having solved that now, we're able to refocus. We've stood back from the portfolio, and we see some really strong growth, multiyear growth opportunities for QNX. So QNX is now our #1 capital allocation priority. And even with increased investment, which we're putting in this year in a number of activities: One, to expand the TAM beyond automotive into adjacent verticals, Mattias touched on briefly and also to try and increase our share of the stack with a more fully integrated prepackaged software platform. Despite that, QNX will still be solidly profitable. So company is generating cash. It's profitable in totality. So last month, we announced a share buyback program, which gives us more optionality in terms of how we deploy our capital. It also shows confidence in our plan going forward to be able to generate cash and profit. And also finally, that we believe our shares are undervalued currently. Going forward, we have optionality also to look at tuck-in acquisitions, if that makes sense, the bar will be high. We want it to be absolutely accretive, but it's options to fast track the opportunities that we see before us. Okay. Let's move over to Q&A.

Sure. I'll probably start with just a couple of basic questions just so everyone gets the lay of the land. I mean, one, can you just remind us who are the main competitors that you go up against? I mean, I know Aptiv bought Wind River, which is a competitor, not even that familiar with who else is in this space. Any color on your share...

It's a great question. And because the segment is small, it's a bit complicated, it's complicated to tease this out. But simplistically, you can say there is a bunch of what I would label legacy RTOS providers. And there's lots of them, Wind River, Green Hills, Mentor Graphics. There's a lot of them that traditionally, even Microsoft -- I mean, Microsoft was a big player in RTOS on the embedded side. In automotive, the lead segment for centralized compute for high-performance compute domains, we don't see any problem yet. And you saw that in some of the numbers, digital cockpit, 16 out of 20 and so forth. So in that -- in the really high end, for the hypervisor and the bottom layer, we don't see that many. You move up the stack and there are -- there's a different discussion, and we can come back to that. Then for the higher layers of the stack, if I move above the RTOS, so the middleware, there is still a lot of open source ongoing from -- for what is non-differentiating. And there have been many, many attempts at safety certifying, various types of Linux. Most of this sits within our customers and partners, so OEMs and Tier 1s doing work. If I were to summarize it over the [ last ] months, they start with great ambitions. It hasn't worked out very well, and they come back to us. So there's sort of two categories. It's the legacy players that we see primarily outside of automotive. And then it's sort of open source within customers themselves, so OEMs and Tier 1s. And then it's really a question of where you are in the stack, how far up you go.

And the OEMs and Tier 1s are trying to do their own system or is the middleware?

If I -- it's a good question. So if you go back -- if I take a step back and you just think about software-defined vehicle for the last 5 years. I would say 5 years ago, it wasn't as hyped as it has been the last couple of years. But 5 years ago, most OEMs realized that this was a very, very important topic for them, and it's going to run for a long time. They had very limited software capabilities, if you think about the full stack 5 years ago. And many of them pivoted very, very hard into software. They hired thousands and thousands of engineers. And many of them made the statement that we are going to build and control 90% of the software stack because value is migrating to software. So we need to figure this out. The last couple of years, there have been some spectacular failures because although software is where the value is going over time, you can control the software stack without writing all the code. And by the way, software engineering is one of the most nonlinear functions there is. You can have one guy that is more valuable than 100 guys. So putting, let's say, 5,000 software engineers together to build all the software in the car might not be the right approach. So the pendulum swung all the way 5 years ago to we're going to do it all, Tier 1s are not going to do anything about hardware, and then it started swinging back. So in the last couple of years, we have seen a lot of the open source -- and the way you structure that internally for an OEM, say, 5 years ago, you say, "Oh, I'm going to build it all on open source. I'm going to take various versions of open source code, and I'm going to build the entire stack myself." There are big problems with doing that. There are technical problems with doing that. But more importantly, I think the question is why? Why do you want to build the entire stack on open source where there's a problem for the foundation that has already been solved. It's called QNX. Why don't you take your precious resources, software engineers that are so difficult to hire and have them focus on things that your consumer, when they buy the car, actually can see. I had it on the first slide, this notion of undifferentiated heavy lifting that we use is an AWS term. The reality is that although this segment of software that we do is very, very important, it's undifferentiated for the person buying the car. It either works or it doesn't work. So if the OEM spends 200 software engineers full time trying to go all the way down to the silicon, even if they're successful, they can't charge for it. They will not be able to differentiate on it. So the argument we are having, the discussion we're having is always, guys, I know it's an organizational question that has to be resolved. But why are you wasting your time when you have so many other software issues to solve, trying to provide hardware-software separation, making sure that everything works at the bottom, safety, security and all that stuff. We have solved that problem, just work with us.

So I mean, you make a good argument for not doing outsourcing or whatever the OEM -- in-sourcing it. What about -- what is the differentiator between you and your competitors? Like what makes one system better than the other?

Yes. So if you take then -- and as I said, for the high-performing compute in automotive, there are very few competitors. So if you think about the hypervisor, we are dominant for high-performance compute in automotive. What is differentiated...

Level -- like advanced data...

Yes, advanced data, all the way down to touching the silicon, same for digital cockpit, the most advanced stack. So you have centralized everything that is digital in the car, including clusters and whatever cameras you have in the cabin and the infotainment and all that stuff. If you have built one stack for that and not everybody is there yet, we are completely dominant as the foundational piece of software there. But back to your question, what is differentiating? So the value proposition, again, faster time to market, lower lifetime cost, safety, security and reliability. The differentiation sits underneath those components. So if you're deploying a modern stack today, SDP 8, the software platform that we released last year, we have the highest performance. We have the highest safety. We have the highest security and it's the most reliable platform. And you go deep into the technical details to prove that. But that's where the differentiation is.

And in your presentation, you mentioned scaling linearly. Can you explain what that means? I mean, with the more power, there's no lag or anything that...

Yes. So we are a hard real-time deterministic operating system, and that's tied to the microkernel architecture. So typically, you have for many of the real-time application, a threshold of 30 microseconds for response. It's hard -- you put your gear in reverse. The backup camera needs to show up right away. The sensors need to start working. It's a hard requirement. And there are many, many other cases, as you can imagine. The scaling linearly with core is tied to -- if you centralize compute, you move from what traditionally was maybe a 1 core system, 2 core systems for less functionality to 4 cores, 8 cores, 16 cores, et cetera, et cetera. So the silicon becomes more powerful and you take all those 100 single function compute modules that you have in the car and you put them on the same silicon. Then the real-time requirements and how you distribute compute and memory, and all becomes very, very, very critical. And what our new platform SDP 8 does is that if you take a 4-core silicon and you move to an 8-core silicon, the compute actually goes up 2x. That was not the case historically. Let's be very, very clear. It was not the case. And we solved that problem. We took almost 4 years of R&D. And as I mentioned, it was based on NVIDIA obviously scaling very heavily in compute and wanting to make sure the software stack would scale with them. So what it means is when you add -- when you double the number of cores, what you actually get from the silicon in the software stack is also double the capability.

Got it. And in your slides, you talked about that with the distributed model, you don't have any content, or you have very little -- I was surprised there's 0 content on the distributed or is it just a lot bigger on the zone and the...

The way to think about that and without getting too far into the history here, but if you go back a few years, the vast majority of compute modules in a car was boxes from Tier 1s. And they were built on a category of silicon called MCU, microcontroller units. And they typically had a single function or maybe 2 functions on them. And what you ended up doing from all the Tier 1s, which is one of the reasons why we're in such a mess in the automotive industry today, is everybody had their own approach. So they used -- maybe they use some form of AUTOSAR. So there's a run time on these MCUs. It is dedicated, hard-coded to that particular silicon, and it provides this function. The problem with that is that it's not future-proof. So when you change the architecture, you start over again. You have to hard code and write a lot of software specifically for that compute model. We have -- when you run something like QNX that has a high level of sophistication, you need enough memory, enough compute on the silicon itself for it to make sense. If you're deploying a single function or 2 functions on a piece of silicon, you don't need sophisticated swapping between threads and all that stuff. So QNX can simply not run on the low-end compute. And that's why the vast -- the gray bar of that, it wasn't even a market for us. We -- QNX was not intended to run on those. They are typically all covered by some form of AUTOSAR today.

So you're on 255 million vehicles that you said. But the global market is not addressable to you, a lot of...

A lot of it is not new...

Quite high...

Last year, the official number, I think we added 25 million cars to it. So it fits pretty much exactly to that graph. There are some gray areas, as you can imagine. MCU is not a defined concept. So it varies a little bit. But we also have a long history. So the 255 million is cumulative. It's cumulative of our entire history. So last year, I think the public number was 25 million, but we added 25 million cars.

Got it. A lot of headlines with you guys in Qualcomm. I mean how integrated are the two businesses? I mean, you mentioned NVIDIA, too. So is the -- are they like a very strong strategic partner?

Qualcomm is a very strategic partner for us. We do a lot of work with Qualcomm. Qualcomm is very sophisticated in terms of software. They have a lot of QNX software engineers. Qualcomm's default for their sophisticated silicon is the foundational software is QNX. And I mentioned, if you take something like Qualcomm Ride, that software package actually by default includes QNX. So we do a lot of work with Qualcomm, it's a fabulous silicon company, as you know, we do a lot of work with them.

And the trend toward EV and autonomy, it sounds like that would be a big tailwind for you guys. Is there -- does EV actually drive a lot of this? I'm kind of assuming because I always hear more compute in an EV.

Yes. So great question. And let me chop it up into pieces. So simplistically, the drivetrain does not matter. I mean, if you think about how an ICE engine works, it's basically based on MCU. So there are single function compute driving everything that you need for the engine. So that was never a market for us. A sophisticated modern EV stack is typically sitting or starting to sit, not quite yet, but starting to sit on a more sophisticated silicon and a more sophisticated software stack. So we're going from in ICE, we could get no compute, no instance with QNX. In EV, we can't. So it's actually adding to our TAM. And as I mentioned, we have 24 out of 25 EV customers -- EV players are our customers. That said, there is a tweak underneath that statement. And the tweak is many OEMs in trying to deal with all the trends that they have, be it autonomy or electrification or software-defined vehicles, when they make that transition, there are a lot of moving pieces. And many of them, in my mistake -- in my opinion, made a little bit of a mistake, they bundled the EV platforms with next-generation software architectures. They left the bulk of the existing business, which is sitting on ICE, at least for the legacy player, they said, "Well, we're going to use EV as the template for re-architecting also the future of ICE." As you've all seen, EVs have a little bit of a headwind in the last 12, 24 months. So what that ends up doing is many of their sort of tip-of-the-spear initiatives for the software architecture is sitting on a platform that is not ramping. And that is not great because regardless of what you think about how quickly the transition will go, you should do that for ICE too. You should go down the improved compute architecture for ICE too.

Maybe before I wrap it up, what about non-auto? Because you mentioned -- you kind of talked about it holistically. What are the opportunities outside of auto for QNX?

So today, if I look at last year, and I think we even made it public, 75%, 80% of our business is automotive, and that's deliberate. We believe it's the most advanced. We believe it will spill over to other industries. That said, I've said many times, I believe long term, the other industries are a bigger opportunity than automotive for us. And the reason for that is obvious, there are more other advanced B2B edge devices than there are cars in the world. And if I look at these segments that we have focused on that have similar characteristics, so high requirements for performance, they're moving to a multi-core architecture, higher safety, security requirements, high reliability requirements, long life cycle management and so forth. Industrial automation, medical OEM equipment, anything that looks like a car, so commercial vehicles, things that don't have a number plate, but sort of looks like a car. Those are segments where we're already present and they're starting to move. We're actually very surprised when we launched SDP 8 last year. SDP 8 was developed, as I said, on requirements from NVIDIA for automotive. But we had almost equal interest outside of automotive when we launched that product last year. So I think long term for us, and we -- as Tim hinted that, we're making a major investment in it this year. Long term, I think it's going to be bigger than automotive for us.

Okay. Great. I think we're actually a little past time. So thank you very much for joining us today.

Thank you, Colin. Thank you for having us. Appreciate everyone joining.

Thank you.