 Hey, welcome back everybody. Jeff Frick here with theCUBE. We're at the Autotech Council, let me get all this right, innovation and motion event on mapping and navigation. And it's hosted by Western Digital. And it's all about navigation, mainly in the context of autonomous vehicles, but it's fascinating the layers of complexity and opportunity around navigation driven by autonomous vehicles. And we're really excited to have representative from Western Digital. He's Martin Booth. He's the director of marketing for Automotive for Western Digital. So Mark, I'm welcome. Yeah, thanks, Jeff. So what do you think about an event like this? I mean, you got to be in your happy place because you got all these phenomenal innovators around mapping and this complex layers and methodologies. And at the end of the day, they need you guys. Yeah, I mean, the automotive industry is changing rapidly. And it's becoming a driver for a lot of technology companies here in Silicon Valley, including Western Digital. Right. That's why I love the space so much because it is this bright shiny object literally that represents IoT, that represents industrial internet. It brings together all these different technologies, cloud and edge computing and 5G, all in this one thing that we can actually see and touch and enjoy. So what have you seen today that surprised you? Anything? So I think there's a lot of discussion on mapping and how the size of maps grow. Today in a car, you might have up to about 64 gigabytes of map data in a European model in the US. The maps are a little bit less. But as you go to advanced driver assist systems and autonomous drive, we're talking about adding HD map data and layers and that drives up the content significantly. A lot of the discussion today has been about the trade-off between do I download that from the cloud as I go along or do I store it locally in the car? Right. And of course there's cost and other trade-offs that you have to make in those considerations. There's no single answer to that question. I think the bottom line though is a lot of storage is gonna be needed locally. There's gonna be a lot of edge processing locally. And yes, stuff will be sent to the cloud and from the cloud to the car, but they're both important and they're not necessarily mutually exclusive. Yeah, we were joking a little bit before we turned on the cameras about bigger and bigger SD cards for our cameras and our phones and everything else. And you touched on, but the application in the car space, the automotive space is very different than kind of a classic, something I buy at B&H. I wonder if you can speak to some of the specific considerations you have to think about when you're designing memory for the automotive application. Yeah, there's very different types of storage memory available out there from low-cost consumer products to very high reliability, high performance products designed specifically for the automotive industry. In a car, the temperature ranges that you're dealing with, the longevity of the product in the field, the reliability considerations are very different from a mobile phone, for example. And as we heard over and over, then there's all these multi-layers within the applications. You got all these different applications. So even within the car itself, the variation between the demands that you're going to need based on the application devices, what layer of the map it's working on are going to be highly variable. Yeah, there's over a hundred computers in a typical car today already. And a few of those already use Nantype memory for map data and navigation and infotainment purposes. But as we move forward over the next five years, the compute power you're putting in your car is going to grow exponentially. To do things like autonomous drive, you need to compute a more powerful than you've got in your pocket or in your desktop PC today, actually more powerful than an average server that you've got today. And so that's going to require a very different type of memory solution. That's something that Western Digital has been focused on for the last few years and is optimizing products specifically for this type of application. And it's not just one system in the car. The system's going to be distributed. Each of them is going to have their own specific and unique requirements. Right. Although it's not as big as you think. I interviewed somebody from NVIDIA at the International Auto Show in Detroit earlier this year, and the little autonomous card pack thing is a lot smaller than I thought. You know, it can hold it like this. Yeah, the autonomous computer itself has to be shrunk down into a form factor that can actually fit in a car and not take up too much space in the cabin, for example. But it's not just the autonomous computer. It's the infotainment system. It's the digital cluster. There's a vehicle-to-vehicle communications module. There's a telematics gateway. There's a drive recorder, potentially. There's dash cameras that might be recording. So there's a lot of different applications and areas in the car where they're going to be storage requirements. Wow. And then the other, of course, big thing, which you touched on briefly, but to call out more specifically is this concept of edge computing, working in conjunction with the cloud computing. So some stuff you need to process locally. Some stuff you can send back to the cloud. Some stuff you want to suck down from the cloud, whether it's updates to the maps or new updates to algorithms that are based on a aggregated, bigger data set than simply coming off your vehicle. Yeah, I mean, with all the sensors, you're going to get added to the cars over the next few years. You've got potentially half a dozen cameras. You've got radars or lightars. You've got other types of sensor data coming in. That type of data is coming in at you at gigabytes a second. And that has to be processed locally. You don't have enough bandwidth to go to the cloud with all of that data. So that data gets processed locally. You're going to make a lot of decisions based on local database that you're comparing it against. You can't necessarily rely on connectivity in your car. At your house, you're always connected. But if you're driving around even here in California, you don't have to go too far north of San Francisco and there's no cell phone signal. So you need a local copy of the data. You need to do the edge processing. And then, yes, there will be exchange to the cloud when you have connectivity or when the connectivity is cheap enough. But it'll be a fraction of the data that goes backwards and forwards versus what the car is actually collecting. Right. There's a funny question in one of the panels. Somebody asked, like, look ahead 10 or 20 years, which is an awfully dangerous thing to do where we live. But just looking ahead in the next couple of years, what are some of your kind of product priorities as you see this market evolving so rapidly? What are some of your kind of top-of-mind priorities that you and your team are working on over the next year, 18 months? Yeah, so I think in the semiconductor industry in general, but certainly in the memory industry, it's always about reducing cost, improving performance, increasing density. And so there's a lot of new technologies coming in the roadmap in the next couple of years that will do all of those things. And as storage becomes faster and less expensive, more of it gets used. And the applications in the car are gonna require that as we go forward. And there's gonna be a trade-off between how much do I put locally versus what do I rely on being connected to? No slowdown in the consumption of storage. I think you're in a good spot. Yeah, you can never have enough. All right, Martin. Well, thanks for taking a few minutes out of your day and congratulations on helping out with a great event. Yeah, thank you. All right, he's Martin Booth, I'm Jeff Frick. You're watching theCUBE. Thanks for watching. We'll be right back.