I’m sitting in the passenger seat of the latest BMW i5 while, in the driver’s seat, Qualcomm’s group general manager of automotive, Nakul Duggal, is fiddling with the air conditioning on the giant display stretched out across the car’s dashboard.
This is the 2025 i5, the latest EV model in the luxury carmaker’s lineup. It’s here to show off Qualcomm’s vision for how its chip technology can help make smart cars be more than robots on wheels.
“There are probably 500 sensors in this car that measure all types of things,” Duggal tells me in our exclusive chat in the vehicle. Cars, he says, will be integrated into your digital life, “an extension of your home.”
The in-car cameras have many uses — including selfies.
This car doesn’t have all of the AI tech that was announced for automotive at this year’s Qualcomm Snapdragon Summit in Hawaii — if it did, it might be able to see that Duggal and I are both struggling with the Maui heat and then work out that we need all the cool air it can throw at us. The small camera he points out to me positioned above our heads could perhaps spot that we’re visibly perspiring, and the sensors inside and out of the vehicle surely would be able to tell that the internal temperature is undesirable for two people who have just stepped out of the midday sun.
At the summit last week, Qualcomm announced major upgrades to two of its automotive platforms, the Snapdragon Ride Elite (for more sophisticated assistive driving) and Snapdragon Cockpit Elite (for digital experiences geared to drivers and passengers). Over the next year, automakers including Mercedes-Benz and Chinese EV maker Li Auto will start to use Qualcomm’s latest chip developments to build out their next generation of AI-powered driving and in-car experiences.
These could be purely functional — spotting a safety issue with your car, identifying what maintenance it needs and booking it in to be fixed by finding a spare spot in your calendar. But they could also be whimsical. During the summit, Li Auto demoed a feature that would allow you to capture a picture of the scene you’re driving through and have it turned into artwork in the style of Van Gogh.
As Qualcomm CEO Cristiano Amon said during his summit keynote, it’s hard to think of an automaker the company isn’t partnering with these days. It goes to show that car manufacturers are betting big on these AI-powered conveniences being an essential factor in people’s buying decisions.
Road tripping with AI
Duggal described the result of the Li Auto demo as “beautiful.” When I ask him what AI features he most hopes automakers will take advantage of when they start to use the Snapdragon Ride Elite and Cockpit Elite platforms, he points to creative experiences such as this that bridge the worlds inside and outside the car.
“People have a lot of memories when they go on trips, when they are on a vacation, when they spend time with their family, and it’s very difficult to capture those things,” he says. The car could easily ask you whether you want to capture special moments on the road and record them for posterity, he says.
AI can help bridge the gap between the car and the world around it.
The brilliance of generative AI, Duggal says, is that it’s “finally allowed the machine to speak to the human.”
Until the artificial intelligence embedded in your car gets to know you, there will be a certain amount of it asking you questions by seeing if you want to do things on repeat — playing the same playlist every time you commute or knowing that you like to stop for caffeine on your way to work and always routing you via your favorite coffee shop, for example. “The customer will be included in the conversation,” Duggal says. “It’s kind of like learning like, hey, every time something happens and I like it, I would like to reinforce that behavior.”
But speaking to you doesn’t just mean adding to the clamor in your head by pestering you about your preferences, it means anticipating what you want and doing it before you even need to ask. “When AI is effective is when it’s just operating in the background and you’re not even aware of it,” says CCS Insight CEO Geoff Blaber. If AI works as Qualcomm promises, it wouldn’t add to our mental load, but carry it for us.
As much as it sounds like there’s plenty AI can do for the car, it also seems that there’s plenty the car can do for AI.
“The car is going to make AI much more relevant, much faster,” Duggal says. “That environment is very rich, it’s very noisy, but it’s also very powerful in terms of what you can learn,” he adds. The sheer amount of inputs in a vehicle and data it gathers will allow new use cases to spring up.
Duggal gives the example of a car noticing that a teenager, who has only recently learned to drive, is repeatedly distracted when reversing out of the garage. The car could take note and inform the teen’s parents. Yes, it’s snitching, but it’s the kind of potentially life-saving snitching that parents would love.
Screen real estate is taking up more and more space in cars.
The wheels of change
AI is changing cars so fast that it’s basically “overnight,” Duggal says.
Not my car. It’s a 2008 Toyota that came to me via my parents and my brother, and it’s not about to benefit from AI. It’s from a different era of cars, one in which software-driven experiences, never mind AI, didn’t factor into carmakers’ decisions at all. That’s just not the case anymore — it’s evident in the giant dashboard display that stretches out in front of me in the BMW i5 that while I’ve been driving about in my old beat-up banger, a shift has occurred. It is, as Duggal points out, “not incremental.”
The speed at which this is all happening is something that came up frequently at the summit — especially as it involves automakers, which aren’t known for their sharp pace of change. “They have all moved towards software and being software-led companies pretty quickly for what is a slow-moving market,” Blaber says.
A couple of years ago during the global chip shortage sparked by the pandemic, car companies had to halt production and sales of new vehicles. It’s something that never would’ve happened in five or ten years before, when automakers weren’t reliant on silicon supply chains.
I’ve always viewed cars from a utilitarian perspective — safely getting me from A to B is about all I ask of them. Being in the i5 shows gives me another perspective. As I sit and have a game of Uno with the screen on the dash in between meetings and take a selfie with one of the internal cameras, I see the potential for fun.
BMW i5 at rest, packing AI.
Qualcomm’s vision is to take this to the next level. Zoned audio could allow each passenger (and the driver) to have a personalized experience based on their own preferences and what they’re doing on their journey. The in-car cameras will recognize people, so will know if they’ve traveled in the vehicle before and remember their preferences, Duggal says. This is what I think he means when he says the car will be an extension of your home. There will be a level of comfort added when the car can anticipate your preferences and be ready to welcome you back on board.
“They’re presenting a vision,” says Blaber of Qualcomm’s big auto AI push. “Whether or not we get there, the reality is that the car is becoming heavily software defined.”
As fast as this is happening and as exciting as the vision is, it doesn’t feel entirely accessible. It seems like a luxury experience — one that will delight the i5 owners of this world within the next 18 months or so, but will remain a pipe dream for secondhand Toyota owners such as me. I point this out to Nakul, but he believes that AI will infiltrate cars of all prices and won’t be a feature just for those who can afford a luxury vehicle.
“There isn’t really a correlation with people who will appreciate it and the cost involved,” he says. Qualcomm has developed different solutions for different price tags, he adds. Each platform has enough power to support artificial intelligence. Plus the company signs long contracts with carmakers so the platform will continue to be supported — maybe even long enough for people like me to take advantage.
Each of Qualcomm’s auto partners, whether it’s Li Auto, Mercedes, Rivian or General Motors, are all very different and will use artificial intelligence in their own ways. “The possibilities with the car are just so many,” Duggal says. Everyone should be able to experience them, he adds.
