Cars Are Taking Off for the Clouds

A car that allows the driver to press a button, sprout wings, and take off from the road to the clouds is unlikely to ever become a practical reality. What is real now and growing is the car’s dependence on the cloud. In recent weeks, there have been numerous announcements from automakers and suppliers about the new relationships being developed with cloud computing providers. 

Back in the 1990s, companies such as Oracle, 3Com, and Sun tried to make the case that the future of computing would be all about having cheap information appliances in the home. These appliances would connect to powerful servers in a data center somewhere that handled most of the actual workload. They weren’t entirely wrong, but their timing was certainly off. 

In the 1990s, most homes were still on dialup if they were connected at all. The first connected vehicles, such as GM’s OnStar, came out with first generation analog cellular devices. The connectivity was simply not adequate to make a device like the 3Com Audrey useful, and OnStar’s original functionality was limited. As of early 2021, most new vehicles come out of the factory with embedded 4G LTE data modems, and 5G will start rolling out over the next 2 years. 

Why Do We Need So Much Bandwidth from the Car and Computational Power in the Cloud?

Most modern vehicles already have anywhere from 50 to 100 computers onboard controlling everything from the powertrain to the system that makes power windows retract if a child’s hand is in the way. Connectivity can’t be guaranteed, so a lot of functionality will always need to be controlled locally, especially anything that is safety critical such as driver assistance or automation. 

However, that doesn’t mean that those systems can’t be enhanced with help from the cloud. Increasingly, systems are taking advantage of map data to determine where a hands-free system can be activated, to decide when speed needs to be reduced to safely handle a coming curve, or to plan the most efficient route. Keeping map data fresh and accurate is crucial to those functions.

An EV driver in need of a battery top-up needs to know not only where there are chargers but also which chargers are actually functional and available. In cold weather, that information can also be used to trigger preconditioning of the battery so that it can accept a faster charge, reducing the time required to wait at the charger. All of this requires real-time connections. 

Vehicles aren’t just ingesting data from the cloud—they are also uploading it. As highly automated vehicles begin to arrive, they need even more detailed HD maps. Some of these maps, such as Mobileye’s Road Experience Management system, are already being built by collecting data from vehicle cameras and sending it to Mobileye servers. 

The in-vehicle user experience can also be dramatically improved with help from the cloud. Voice recognition has been in the car and annoying drivers for more than 2 decades. But new systems that process natural language voice requests in the cloud are now helping to keep the driver’s hands on the wheel more of the time. 

All of these capabilities and more are why GM and VW have partnered with Microsoft Azure, why Ford is working with the Google Cloud platform, and why many automakers including Stellantis are incorporating Amazon Alexa voice services to support new generation vehicles. In the coming years, almost every new vehicle will likely be connected to the cloud, whether drivers realize it or not.