BlackBerry Limited (BB) Earnings Call Transcript & Summary

Up here, we have with us BlackBerry. And I think what's going to be really interesting about this, I know you guys aren't sort of the typical sort of like auto, auto run, but I think what's interesting is QNX is actually very central to auto and the software-defined vehicle. So very pleased to have with us Tim Foote, the company's CFO; Grant Courville, in product Strategy; and Justin Moon in product engineering. So team is going to go through a series of slides, and then we'll open up for questions and then great.

Sounds great. Hello, everyone. Thank you for joining today. So yes, to Dan's point, like BlackBerry, and auto conference, that kind of doesn't really make a lot of sense. But BlackBerry has been under a big transformation over recent years. And we find ourselves today as a software company and particularly for the context of today, we're a very dominant position in foundational software in the car, in particular. I'll just quickly click through the safe harbor statement. So around about half of our revenue comes from QNX. So QNX, predominantly automotive. We're powering 255 million-plus cars around the world. We're dominant in domains such as the digital cockpit, ADAS and other things like body and chassis. And what we're seeing here is that we've got a strong moat around our business, where we're focused on safety critical very reliable, highly certified product, but also a combination with high performance as well, which leaves us in a very strong position to capitalize on some of the secular trends that you're seeing happening in software in the car. We have a very strong backlog, which unlike a lot of software companies, gives us a long line of sight into revenue coming into the future. Our backlog for estimated future royalties currently sits at $865 million, and that's been growing at a CAGR of north of 20% over the last few years, which actually exceeds our revenue growth over the same period, which means that over time, we expect that to start to come together and we see an inflection point in revenue. It's a very profitable business. Last quarter, we achieved a Rule of 40 and we're very excited about where this business can go from here. So to talk to you more about what the product actually does and its applications and why we're really excited about the future. We have Grant Courville here, who will walk us through some of the applications and the trends that are relevant.

So hi, I'm Grant Courville. I'm responsible for product and strategy at QNX, and thank you all for coming. Really appreciate it. I'm truly excited to share what we're doing and some of the trends that we're seeing, and Tim touched on a number of them. So if you think about what we do, we've been in business for decades. And what we've provided since day 1 is foundational software. So foundational software for advanced edge compute. We're going to talk about what we're doing today. If we think about foundational software, the easiest way to probably visualize that is think about the hardware that you might have, so the CPU and whatnot. Then there's the foundational software. Some people will call it the operating system or virtualization, whatnot, but it's the layer between the hardware and the applications and the higher-level middleware. That's the way to think about it. So you won't see our software necessarily interacting with the end user. However, will be that core software that has to work all the time reliably. And that's what we've been focused on for, quite frankly, decades. In terms of advanced edge compute, what we're all seeing in the industry now is more compute power, more intelligence at the edge, more software at the edge, more criticality at the edge. And when I say edge, it's devices that have a certain level of criticality, whether that's in a vehicle or as you can see below, whether it's in the energy sector, transportation, medical, think of surgical robotics and whatnot, think of patient monitoring, whether it's smart agriculture, heavy machinery and whatnot, all of those markets share a common thread, which is a degree of criticality and the need for more compute and intelligence at the edge. So again, we provide that layer of software. Tim mentioned it earlier in terms of Rule of 40, you've probably seen the statistic. We're well over 255 million vehicles on the road, and that's something we're very proud of, and that number has been growing over time. About 5 years ago, we invested in our latest product, QNX 8.0. It was released a couple of years ago. And the big focus there was all about performance, scalability and real time and real time, so be able to provide levels of determinism, make sure things happen when they're supposed to and all the time. And that was the thing that we needed to do to secure our software for the future, for the needs of the future as the compute power will grow. And we're very proud to be able to say we can now scale with the number of cores in a processor. So as you have more processing, more cores, our operating system will scale linearly with that. That's quite the accomplishment without sacrificing the architecture we have. Our architecture is absolutely the most secure architecture on the planet. It's called the microkernel architecture. You can see some of the stats below with respect to OEMs and number of EVs. So we're seeing tremendous success. Obviously, we're seeing growth in our revenue as well. So this is maybe a picture you've seen before, but it is a fundamental trend we're seeing in automotive. If you think about vehicles tended to be dozens and hundreds of systems. They call them ECUs, but systems in the car. And that traditionally was how these cars were designed and built and whatnot. And now there's a trend to start to consolidate, bring some of those together. So this is where you're bringing more compute into a single system in the car. And the easiest system to point to would be the dash of your car. It's called a digital cockpit, for instance. And that's where I'm bringing my instrument cluster and my infotainment system and maybe other functionality into one, system and you're in the dash of your car. Again, you won't see that as the user of your car or as a consumer. But behind that dash instead of having 2, 3 or more systems now, you've got 1. So the level of criticality has gone up. So that's the first place you're seeing it in the car. And this is where you start to move over from that distributed system, of which many of those systems we could not run on. They're smaller microcontrollers, if you like. Now they're moving to more high-performance compute systems. And the minute you do that, then that opens the door to our software and to the value that we provide from our software in those systems. So digital cockpit is the first shift we've seen. We're just starting to see that now in safety systems. And you've heard of ADAS or advanced driver assist systems. So safety systems are going that way. And ultimately, you'll reach what they call central compute. The time frames for this evolution, they're up for debate, but the evolution is clear, and it is happening. It's already happening. So as the compute level goes up, the complexity goes up. As the complexity goes up, we provide more value. We say complexity is our friend. And building on that, our operating system is safety certified multiple functional safety standards as well as security certified, incredibly important in any connected device, let alone automotive systems.

So I think -- yes. I mean what we're talking about here is the cross-section from the solution space in terms of mission criticality and high-performance compute. They talked previous to us. It was all -- there was discussions around safety and the necessity of certain things. Really, what we're saying is as we increase those core counts or those -- the performance of these systems, it becomes even more clear that safety and security is of the most importance in the market. And our latest products and even our previous products actually handle that level of complexity and make it easier to build those systems moving forward.

And one other point to build on that, as these systems consolidate, that's an opportunity for us to run in more of the systems in the car, but also increase the software that we provide in the car, again, at that foundational layer. So you might hear us say more sockets, more layers, but that's really what it means, more systems in the car and more IP, intellectual property, more software in that core layer, which is, again, another opportunity for us and for our customers to build on. So to finish off, I think this might be the last slide. This is sort of showing you some numbers from industry analysts that are showing you the growth from the distributed architecture that I talked about earlier, which is a mixture of microcontrollers or small systems that we would not be able to run on to those much more, say, consolidated systems. And you could see the growth over time through 2030, and that will continue. Again, you're seeing a bit of a shift for a number of reasons, but the trend is obvious. There are many reasons for the automakers to pursue this path. And again, it's something that they rely on us to underpin with our QNX operating system, our virtualization technology and our foundational software. And about not quite a year ago, we also announced our vehicle software platform. And maybe we'll get a chance to talk a little bit about -- a little bit more about that because that plays into this trend quite well, and it's something that customers have been asking us for, which is fantastic. So Tim, I don't know if you wanted to comment at all on the backlog.

The only thing I'd like to mention is looking at the stacked bar chart here, the important thing, as Grant was mentioning, the gray portion, which is the distributed portion is the percentage of the market that we don't currently address, which roughly speaking, in 2025 is about 75%. So we're addressing 25% of the market, and we are dominant in that space. You see within 5 years, S&P Global estimate that, that 25% goes to 75%. So that's a tripling of our addressable market within the space of 5 years. So that's a very significant growth. And to be honest, this is not a growth story that's dependent on increasing vehicle production. In actual fact, even if vehicle production were to decline over that period, our addressable market would be so much bigger than it currently is today.

Okay. So late in the day, I want just to be more organic -- my coffee here. But actually, I just want to start -- and I'd like to sort of go high level, but there was something interesting on that last slide. Just go back a slide. Just help us understand on backlog because you're -- that last bar there, you're saying you're at nearly 4x your '25 revenue. Backlog means different things to different people. And I think we -- in auto, we've become a little skeptical on what backlog actually means. And what does that number mean?

Sure. Okay. Let me have a run at that. So we look at individual design wins. So I don't know, for illustration only, Ford is going to make a new FM 50, and it's going to have an x amount of QNX content in it. Over the life cycle of that design, we're going to get revenue that comes in 3 buckets, Dan. So right at the very beginning, the customer will buy effectively a software development kit. We call it a development seat license, which enables them to build software on top of QNX effectively. Roughly speaking, and it varies from design to design, that's about 20% of the revenue we'll get from the life cycle of that design. Then over the next 2 to 3 years, whilst they're designing the vehicle, they're going to consume professional services from us. So it's going to help with things like integration, safety certification, all these kinds of things. That's another 20% of the revenue. Then the vehicle moves into production. And when it's shipped, we'll get a onetime revenue -- royalty payment from them for the amount of content -- QNX content in that vehicle. This number here represents the royalty piece only, so the 60%, and it's the estimated future royalties from all the design wins we already have in the bag. So price, it's -- obviously, it's a P times Q, price times quantity. Price is contractual. We agreed that with the customer in writing. The quantity is the estimate piece, right? So clearly, some of the larger, more established OEMs have probably a better line of sight to what their future volumes are going to be. Maybe some of the newer entrants, maybe there's a little bit more volatility in there. But what we do is we take a prudent view, we will haircut accordingly to the level of risk, and we'll estimate price times quantity. So the $865 million represents vehicles that are going to come off the production line and the amount we expect to get from revenue. And generally speaking, Dan, we're pretty close to the pin. Obviously, you get some kind of black swan events like COVID, chip shortages and stuff, which can delay timing. But generally speaking, you don't see too much volatility. And one of the beauties of QNX is we're so pervasive across the industry that we work with so many different OEMs that if one is not doing as well, it generally means another probably is. And therefore, we can win on the upside.

Okay. That's helpful. All right. So maybe just a high level, and I think you've sort of frame this. But I think what's always helpful for folks in auto is we hear a lot about software in the vehicles. Can you just maybe conceptually, and I think -- I don't know if it was this deck, but maybe it was on the deck, lay out sort of the layers in an automotive software stack, where you sit, which is more of the OS level and why that level -- that layer of the stack is sort of secure and there's really a more limited base of competition.

Sure. I'll start and then Justin...

Absolutely.

So to your point, we're that foundational layer of software. So we sit below the applications and whatnot that the automakers will develop or Tier 1s will develop. The personality of the car, the brand of the car, the personalization of the services and features in the car. Our software essentially has to work predictably all the time. And we do really tight integration with the silicon vendors. So think of the Qualcomms and NVIDIAs and others. So we have very, very close relationships with those silicon vendors by the nature of our product and obviously, the success that we're seeing. So we're that layer in between hardware and call it the applications and some of the middleware and whatnot. Case in point, if you were to look at a digital cockpit that I brought up earlier, many of them run Android today, from Google or from other suppliers. That environment is very complementary to what we do. We provide that software happens to be our virtualization software that allows, for instance, for an Android infotainment environment to run side by side with the digital instrument cluster. So we guarantee the separation between those such that neither one can affect the functionality or the safety of the other. So again, that layer below what you would see in the dash of your car being infotainment and digital instrument cluster. And Justin, you might want to add to that as well.

Sure. I mean as we -- if we look at history as to what a software stack look like in a vehicle, it was some form of operating environment, a layer of middleware that performed things like signal management and things like that. And above that, the OEMs would provide their value add to their customers in terms of application. And by application, I don't necessarily mean an app like an app on your phone, but an ADAS, it could be a perception engine or lane keep or something like that as that's more higher layer from an application perspective. What you saw in those stacks is a lot of code from activating something to actually doing something on a piece of hardware, ton of layers. Grant mentioned the complexity and the cross-section between high-performance compute and safety. It's an extremely complicated environment. In fact, one of the most complex environments from a software perspective. What we're looking at doing with -- and Grant hinted at the vehicle platform, what we're looking at doing is kind of changing the mindset from integration from an abstraction perspective to integration from -- to a performance perspective. The idea is to integrate appropriately, reimagine what does it mean to build that stack with like-minded companies, with like-minded individuals to truly get everything out that you can from the hardware that the OEMs are paying a lot of money for, right? I mean it's the single largest item on the bill of materials would be the hardware. So what we're establishing with QNX 8.0, our latest product line as well as that foundational vehicle software platform, it's a reimagining of what does it mean to build that entire foundational stack, working with very like-minded companies. It's public knowledge that we are building a vehicle platform with Vector, not a small company. And it's not about just gluing technology together. It's about fundamentally understanding what that integration should look like from a performance, a latency management as well as safety and security perspective. So what we want to do is make it less complex as opposed to more complex when we start talking about the system designs. And when we say foundation, we mean truly foundational software that needs to be in just about every ECU. So it's a solid operating environment. It's diagnostics, it's logging. It's all of those things. We don't want to compete with the value add from an OEM. We want to say, here are the things that you consider undifferentiated list. I think they're very important because that's what we do. But from a differentiation perspective, it's not a differentiator for the OEMs. The OEMs differentiate the high-value applications and features at that level. So we want to make it easier from a base layer perspective to do a more performance integration. And that's kind of where our head is at from a focus perspective moving forward. I hope that helps answer the question.

And the customer base. So I think we've known that there's -- you're talking about being embedded in ECUs. There is the OEM itself, but then you also have Tier 1s, many Tier 1s that are responsible for the ECUs. How do you look at the customer base?

Good question. And that has, as we've all seen, I think, has undergone some changes over the last few years where traditionally, an OEM would outsource the development of that ECU of that system to a Tier 1, hardware, software, the complete environment. What we're seeing now, obviously, is the OEMs are now -- they see more value and opportunity in taking on the software. So they're starting to take ownership of that software. And as Justin mentioned, they're looking for a platform to build on. So traditionally, our -- from a supplier perspective, we would be supplying to the Tier 1. We'd also have incredibly strong relationships with the OEMs, but the OEMs weren't taking on the level of software development that they're taking on today. What has now changed is the OEMs are taking on much more, I'll say, ownership of the software, ownership of the architecture, ownership of the systems that go into the vehicle. And the Tier 1s are now being asked in some cases to say, only provide the software, sometimes only provide the hardware and the integration is falling to the OEMs. There's also the traditional model as well where the Tier 1s are taking on the entire hardware and software as well. So it's gone through a change. And from a QNX perspective, we continue to work incredibly closely with the Tier 1s and the OEMs. It really depends on the vehicle and specifically the program and the system that we're in, whether it's a digital cockpit or I didn't talk earlier about smart cameras, but the number of systems in that vehicle is growing, smart cameras, smart sensors and whatnot. So again, some of those are built by the Tier 1, I'll say, in the more traditional fashion and others, it's a combination of, say, the OEM doing the software, Tier 1 maybe doing the hardware, another Tier 1 doing the software integration and whatnot. So the good news is we've been in this business for 28 years, Justin. 28 years in automotive. And so we work with all of them, again, at that foundational level and allow each of them to provide their value add and differentiate.

So I mean the value statement in my -- and you can keep me honest on this one. The value statement when we're looking at a Tier 1 versus an OEM is virtually identical. We're providing an undifferentiating software from their perspective. There are higher value things that the Tier 1s want to provide to the OEMs. The OEMs want to provide higher value things to the vehicle owners. So again, to Grant's point, we work very closely with the entire ecosystem. Tim likes to say that we're the Switzerland of working with the tier -- the multiple tiers, whether it's 1, 2 or directly with the OEMs. I don't know if there's anything...

And the OEM customer base, because you've seen your list, your legacy OEMs, same with Rivian, right? And I think we've seen efforts really vary by OEM, right? Like a company like Rivian is really taking much more ownership of the underlying -- so how do the efforts vary across those automakers?

Good question. And you're right. You said it perfectly. It does vary across the automakers. Some automakers are looking to develop the complete stack, as Justin mentioned earlier. Others have tried to do that, and we're, I'll just say, unsuccessful at doing that. And that's where it's been interesting because many of them have come to us and said, listen, we would like you -- we build on your operating system and software today. We would like you to build a more complete software foundation. And Justin talked about that earlier, the foundational vehicle software platform we announced earlier. So it really does vary by OEM. The key is we -- again, we provide that core software that's nondifferentiating. And I think many of the OEMs are realizing, yes, it truly is nondifferentiating. It's absolutely critical. It must work. And they'd like to use it across multiple ECUs. You can't scale and really smartly invest, I'll say, if you're always, say, redesigning and starting over. And that's been one of the challenges and almost a little bit from a mindset historically in automotive is to redevelop start over almost every time. And you won't achieve that scale, you won't achieve that abstraction, if you like, between the hardware and the software and the vehicle and truly be able to differentiate and provide value if you're always restarting and rebuilding the software stack from the ground up. So today, it really does vary. To your point, you mentioned a few OEMs. There's others, like I say, that take a different view where they absolutely want to build on what we have today and what we're building in the future for a vehicle software platform. So it's been very interesting and very much in transition. Like I said, some have felt the pain. Others are still attempting to build that complete stack. We're happily supplying and working with, quite frankly, all of them.

The dividing line isn't new entrants versus companies that have been in the industry for a long time. There are many very established OEMs that are very capable of building out that entire stack as well. It isn't just the new entrants that are trying to do things.

Maybe you could just flip back a couple of slides. I want to sort of unpack more of these trends here. Go forward one. Okay. So -- and then I think you also had an uptake slide saying you're not playing distributed, but these too centralized. Okay. So we've heard of all of this before, right? Like -- and this has been every automaker that we look at says they want to have a network architecture. They want to have ECU consolidation going to domain controllers to centralized compute to zone controllers. But at the same time, we know that like Ford literally scrapped, they had FNV4. They scrapped that. Others have tried to do these fully network architecture and they've failed. There's questions around what happens of VW, Rivian, does that succeed or not? Maybe just zoom out, where are we on the uptake of the software-defined vehicle? And why do you think this hasn't maybe played out as advertised, is it related to EVs? Was all of this linked to EVs? Maybe just help paint picture or am I wrong? Is it actually playing out?

How much time do you have? Anyways, it's comp...

That is why we had you. There's a lot of interesting things here.

No. I mean, I'll say a little bit of optimism, I'll say, but the trend is clear. It's absolutely moving -- if we look at this slide from left to right, it is absolutely moving to more consolidation and ultimately to high-performance compute centered around a few domains, and you'll hear zonal controllers and whatnot. And you're absolutely right. This slide, we could have shown it, others could have shown this many years ago. I think the only thing that's changed really is the velocity at which this is happening. And to your point, it's incredibly complex to do this. You're bringing together disparate systems that we're communicating over a network and now you're bringing them together onto one system on chip, one system in the car and now you need to make use of, say, virtualization, for instance. And there's the technology piece of it, I'll say. Then there's the operational piece of it within -- if you think of the OEMs and whatnot, I would have a digital instrument cluster team and I have an infotainment team, let's just say, well, now I have to bring them together to work on a consolidated system. You also have the introduction of system-on-chip or an SOC is the term you'll hear and the number of cores are increasing. So again, in many of these systems in the past, it'd be 4 core, well, now you're starting to see 8 core and beyond. So you have a lot more opportunity to do the integration. Where the challenge has occurred, and I think was the big realization, I'll say, is the challenge and complexity involved around the integration. That's where we've seen so many problems, so many challenges. It's just the integration from a software perspective in these systems. And again, we work really closely with our customers and varying levels of success. And as I mentioned earlier, the first real tangible example of consolidation has been that digital cockpit. We've won a number of programs already related to ADAS, where they're taking lane keep assist and lane departure warning and automatic emergency braking and many of the safety features that we all enjoy in our vehicles, and they're bringing those together into a consolidated system as well. So we won those programs. You'll see them in vehicles in a few years from a consolidation perspective. Again, from a consumer, you won't know whether it's consolidated or not because, again, it's varied in the system. I don't know Justin, do you want to talk more?

Where we are on the software definition journey. We all know there's a ton of software in vehicles today. So from a software definition perspective, we are well down the path. Pick whatever definition you want for SDV. What we're looking at now, and I've mentioned it kind of previously, original integrations were all around abstraction and choice. Any one company has done a lot of things that they can -- and just about everything they can do in the industry today. Really where we need to go to next is how do we partner appropriately to solve the actual industry challenges. In the coming years, I truly and fundamentally believe that partnerships are the next superpower. How do we actually work together as an ecosystem to really satisfy and solve the industry challenges that exist from a software complexity perspective. And that's where our foundational vehicle software platform, part of the value of that is that starting point for commercial companies to come together to ultimately solve industry challenges, like-minded companies, culture is a big thing, making sure that we're all focused on the same industry challenges, solving them in the appropriate way, not trying to outcompete one another and all of those types of things, but really truly understand what the industry challenges are. And I get that there are plenty of ecosystems that have occurred and plenty of consortia have been tried in the industry today. There's lots of room for lots of them. But the idea here is truly having commercially backed companies focused on solving those industry challenges at the most fundamental level gives us that starting point to solve the much more value-add things moving forward. So where are we on the software definition journey? Very far along as far as I'm concerned. But at the end of the day, it's going to come down to how do we do appropriate integration from a performance and safety management perspective with like-minded companies or partnerships...

If I can add, Justin mentioned ecosystem, and that's really important. We need to build out more talent around the systems that are going in vehicles. And I'll say around embedded systems in general. And what we've done just in the last year, we've announced a program called QNX Everywhere, where we're making our software available for free for noncommercial use. So the research institutions, the post-secondary institutions and whatnot, they can now get access to our software and start to build out embedded systems, critical embedded systems and build that talent of developers that can then feed into automotive and general embedded and the market segments I mentioned earlier, where they're ultimately going to be building their systems on QNX, but we need to help build that talent base so that early on, they can gain the experience with QNX and that the talent will feed into the ecosystem. The ecosystem will help accelerate what you see here in terms of the evolution of vehicle architectures.

Seeding an ecosystem now is easy. You get some like-minded people together and they figure out a problem. What we want to try to do is really see the future, the far future. We want to make sure that we're building a legacy ecosystem here. And I don't mean legacy old, I mean a tangible ecosystem that can thrive for the foreseeable future. So making sure that we're working with those top research institutions and universities, not necessarily just in North America, but globally to ensure that, that next generation of automotive software engineer is available, knows everything there is to know about QNX and our technologies and the integrations there. And I think that's just as important as the tech.

Can you help us differentiate 2 areas of why your customer base is or is not this set of partners? A, Tesla. What is Tesla doing that sits so unique, which as I understand, Tesla is not a customer? And B, where are you in China, where I think people sort of cast China as the leader in -- the global leader in uptake of software-defined vehicle, just a very different set of dynamics and maybe not having the same set of challenges that we've seen in the West.

Good question. I mean I'll talk about China, maybe you want to talk about Tesla. From a China perspective, we've seen success in China. We're in digital cockpits and safety systems in China. We recently won a program for smart camera in China as well. To your point, they want to move fast. They are moving fast from a software development perspective. Where things are changing is you're seeing China obviously export more vehicles today globally. And when that starts to occur, then all of a sudden, some of the safety certifications that we have and specifically in automotive, there's something called ISO 26262. It's a functional safety standard in automotive. That's something that we've certified our products to that no matter where you're developing your vehicle, including China, that's something you want to adhere to not at a software level, but also at a system level. So all of a sudden, that comes into play. And now more recently, there's a security standard that you have to adhere to as well called WP.29. And we've precertified our operating system and our software to one of the ISO 21434. I won't go into too many numbers, but these standards and regulations come into play now the minute you want to export. So we're winning business in China for the domestic market. And then the companies that are looking to export that's when they leverage our technology, our safety certifications, our security certifications, our services that we talked about earlier to allow them to build on, quite frankly, a proven and trusted software platform that they can make use of globally. Do you want to talk about Tesla...

I mean, Tesla was considered the North Star from a software definition perspective a while ago. And I had mentioned previously, it's not just new entrants that are actually pushing the envelope in terms of software. It is even the very much entrenched industry big players are competing almost at a level playing field when it comes to a software perspective. So let's not discount them.

This is actually what they tell me when I -- maybe when I -- maybe I know we're like near time here, but I want to maybe squeeze in one more. And it's just a question of SOC providers. So you talked about NVIDIA, Qualcomm strategic partners. How does SOC provider play a role at all, if at all, in sort of your uptake in vehicles?

I mean it's the biggest item on the bill of materials. So a lot of the decisions actually start with the SOC. And we -- because we have such close relationships out of the box, QNX is the default choice. So NVIDIA DriveOS is actually built on QNX those like the default position. Qualcomm, RideOS, similar. So with these people like Qualcomm is dominant in the digital cockpit and leads with QNX. So it's a really important channel for us. Anything you want to add to that?

I mean there are very significant silicon companies that aren't Qualcomm and NVIDIA as well that we work very closely with. I mean, NXP is a very large company, and they have a very large footprint in the industry. TI, another, very, very deep ties with all of them. At the end of the day, we're an operating system. We abstracted the hardware. So we have to have those deep ties. We have to make sure that our software is as performant as it possibly can so that everybody north of us have all of the CPU cycles they need to accomplish all of the features of the system generally. So extremely important from a pull-through perspective.

With many silicon vendors, and that's part of the work that we take on, that really tight integration and that road map alignment. So that when their new silicon appears, our software is there, ready to go. And what we're doing as well in working with the silicon vendors, but also the cloud vendors. In other words -- yes, so what we've adopted is a cloud-first approach to what we do because our customers want to, again, accelerate development. So they want to start development with our software even in the absence of silicon or in the absence of hardware. So we put our software on Azure, on AWS, it's used in private cloud and whatnot. So again -- and working with the silicon vendors who are providing that parity in the cloud as well. So they can start developing there, then move it on to the "real hardware" and again, get that head start in development. So silicon vendors through all everything we're saying are absolutely critical to what we do.

Okay. Great. We'll leave it there. Thank you.

Thanks. Appreciate it.

Thanks a lot. Appreciate it.

All right. There's -- if folks want to continue, we can, beverages down there.