00:00Tell us first about the genesis of this business, because even I, as a gearhead, did not know what dominance you have over the automotive software market.
00:13Morning, and thanks for having us. Yes, so QNX has a long history of supporting the most advanced edge compute stacks in the world.
00:21We really doubled down hard on automotive several years back. The main reason being that for advanced edge compute, there's simply no other edge device that has more compute, more sensors, more connectivity, more regulatory requirements.
00:36If you want to lead in advanced edge compute, you need to win in the car. We've been very successful. You mentioned the 250 million.
00:42We are global. 10 out of 10, the top OEMs are our customers. 24 out of 25 of the top EV manufacturers are our customers.
00:50And we also have a significant footprint outside of automotive for advanced edge compute.
00:55So what is the trend? What is the AI-driven trend that you're spotting? How is that going to look in cars that we buy in 2030, 2035?
01:06Yeah, so as you mentioned, we have done a survey for embedded software developers in automotive.
01:14There's lots of feedback in that survey. AI is one component. Maybe hit that first and then come back to a couple of the other components.
01:23Essentially, what developers are telling us is that they believe AI and software development for advanced software-defined vehicles could, from their perspective, replace 30% of the workforce.
01:36Long-term. I think a simple survey like this, you shouldn't over-pivot on that number.
01:41The reality is that with regards to automotive software, there is a significant backlog for the automotive industry.
01:48And if AI can drive significant productivity improvements, I think that would deal with some of the other challenges that the automotive industry has with regards to software.
01:58Long development cycles, significant recalls, significant program delays, high complexity when it comes to integration, etc., etc.
02:04What are you doing to reduce that complexity, Matthias? Because, you know, during the supply chain crunch especially, a lot of automakers complained about the fact that these products come with thousands of chips.
02:20Does it stay that way? Do we get even more chips in cars? Or can you reduce that in some way?
02:26Yeah, so multiple components of that question. Let me start with the basics.
02:30So, simplistically, the automotive industry is on a multi-year journey from a very decentralized compute architecture to something that at some point will be wholly centralized high-performance compute.
02:43You know, there were hundreds of ECUs in the car for many years.
02:46That is being centralized into compute stacks that have much more compute capacity.
02:51Think centralized digital cockpit, centralized ADAS, etc., etc.
02:55And throughout that journey, software gets a larger and larger portion of both the value creation and value capture.
03:01What we are trying to do is to help OEMs, automakers, players in the value chain that are building these next generation stacks to accelerate their time to market, reduce cost and complexity, increase safety and security and reliability by allowing them to take away the foundational software in that stack.
03:22We're a foundational software provider. We build the bottom of those advanced software stacks.
03:29Think of us as the plumbers of advanced edge compute.
03:31And if you don't have to build all the stack, if you don't have to deal with hardware-software separation, all the plumbing, you can move faster and drive innovation in a better way for your end consumers higher up in the stack.
03:44Is there a difference, by the way, between the highest end and, you know, your run-of-the-mill $40,000 car?
03:50We're going to talk later on today with Benedetto Vigna, the CEO of Ferrari.
03:55Does he do something, you know, different for a $400,000 or $500,000 car than maybe Honda does for a $35,000 Accord?
04:04There is a massive difference, but it's not directly correlated to price.
04:10So, depending, I mean, obviously, Ferrari is a high-performance car.
04:13They do a lot of really cool stuff.
04:15Software is not their focus for digital cockpit, I would say, compared to, for example, an EV startup.
04:21But there's a significant spectrum of complexity with regards to software across different auto manufacturers.
04:27So, a modern car, we used to talk about 100 million lines of software code in a car.
04:32There are cars launching today with more than 500 million lines of code across multiple compute domains.
04:38You can imagine managing that, updating that, and making sure that the consumer has a great experience.
04:44Matias, one final question.
04:46Competition.
04:47I mean, you haven't got much.
04:49Right now, there's, like, Android has the second most significant share, I would guess, and then Linux and some open source stuff.
04:55But everybody else, there's got to be sharks in the water, right?
04:59What do you do to fend them off?
05:01Well, so, first of all, it's a common misconception that we compete with Android.
05:05Google is one of our great partners.
05:06When you have an Android stack implemented in the car, we actually sit underneath that stack.
05:11There are multiple operating systems in the car.
05:14There are multiple middleware layers.
05:16We sit underneath Android, and we have, by far, the best implementation of the interfaces to the Android software stack.
05:22In terms of competition, quite frankly, the biggest competition is the legacy architecture and the challenge of moving from that legacy architecture, decentralized compute, to the nirvana of high-performance compute.
05:35There have been, as no need to tell you, you're an automotive guy.
05:38There's been significant delays in programs.
05:40There's significant challenges.
05:42Our biggest competition is they stay too long on that legacy architecture.
Be the first to comment