LAS VEGAS — Automakers and Tier 1 suppliers introduced big news at CES this week, with technologies surrounding new vehicle concepts, automation and everything related to data.
But global tech companies such as Amazon, Intel, Sony and Microsoft also made big appearances in vehicle technology and the future of mobility this week at the show.
Tara Prakriya, general manager for Microsoft Azure IoT mobility, which includes connected vehicles and Azure Maps, spoke to Automotive News staff reporter Alexa St. John at CES about the company's work on transportation technologies, artificial intelligence and translating vehicle technology to consumers.
Questions and answers have been edited for clarity.
Q: Explain the concept behind Microsoft Azure IoT mobility.
A: Azure is organized like you would expect a hyperscale cloud, that is aspiring to be everyone's computer, to be. There are many specific areas, including storage, networking and all those pieces. We have a very large investment in IoT as well as in AI. The reason why Azure AI and IoT are really interesting is because in many cases they represent new workloads, new opportunities for customers to take advantage of cloud. . . .
It tends to be some of the more cutting-edge type of use cases. So when we look at the IoT space, we can look at just the basic requirements of connecting various devices, whether they're blenders or refrigerators or whatever they are, to the cloud. But for our customers to be successful and, at our core, we're a B2B — our mission statement is about empowering others to be successful — it became very clear that just that basic stuff wasn't enough and we really had to build more, what we would call platform as a service on top of that, that stitches together many different Azure capabilities.
Q: Where does Azure Maps come in for all automotive technologies?
When you think of this as a mobility problem and not just a connected-vehicle problem, [with] a mobility problem, you've got to think about all the aspects of routing. You've got to think about multi-modal transportation options. I have to think about a whole bunch of geospatial analytics. Azure Maps for us is not just an aggregation of a geospatial data, but it's also geospatial algorithms. We think that when you start combining routing for, say, not just taking into account traffic, but taking into account, is there parking at the destination we're looking for? Are the trains running on time? Do you maybe drive to the train station, then take the train and then drive, or whatever that happens to be? When you want to consider all those different pieces, you now need real-time algorithms to handle that as well. It isn't just about having data sets available to you.
Q: Explain to consumers how significant AI is to new vehicle technologies.
If I take something like weather, if you think about your IBI screen or the infotainment experience, a lot of folks have weather in their experience. Now, certainly having a temperature gauge of seeing what's outside, it's been there for a long time. But if you think about features that are in cars now, where at a certain temperature outside, you might just get a warning signal it's going to be icy or, whatever the regular warning signals are. So, see that connection of, OK, I'm just in my car, I'm getting that temperature sensing and I'm going to go at least alert the user for something.
We can go a lot further than that, both in the car itself, and we have what we would call an intelligent edge — really that ability to combine intelligence across different systems together. And if you then combine that with the intelligence that you can get from the cloud, not just what is the temperature outside, but what is the temperature going to be along your route, you might actually want to change your routing based on avoiding the icy bridges or whatever that happens to be.
You might also want to change whether or not you're going go into sport mode or you're going to go into traction control based on that. It's not just a warning light of when something is cold, but an actual change in the behavior of the vehicle as a result of that. And then combined with the prediction of what the weather is going to be on the route as you go along, and that prediction is only better if it's connected to real-time updates of that prediction, not the weather report from five days ago . . . That's where it becomes much more interesting.
Q: Do consumers see the cost of this cloud technology in the cost of the vehicle?
I think those business models are still getting figured out. I think this is an industry that's going through the changes of trying to figure out how to bring this value to consumers in a way that is profitable enough and yet, that the cost isn't prohibitive to taking advantage of the new functionality. If you price it too high and no one uses it, what was the point?
Obviously, you can't run in the red line for too long and really be sustainable. I think we don't know. I don't think we know how much of it is going to be in the fixed price of the car, how much of that's going to be in the variable price of services that get charged, how much that might be upsell for additional functionality that maybe the car OEM doesn't believe is core to the functionality but is nice to have.
Experimentation is important because the current model for cars is, you decide something and then you're stuck with it for the lifetime of that vehicle. That ability to be a little bit more dynamic in terms of not just functionality, but potentially even in terms of what the cost to the consumer or how that cost is really delivered means that we can take some risk.