In October, Nvidia became the first public company to be worth $5 trillion — then promptly lost hundreds of billions of dollars in market value amid concerns about a bust of a potential artificial intelligence bubble.
As analysts and even founders continue to warn that AI may be more speculative than substantive, car companies may beg to differ. This month, Nvidia is inching back, valued at around $4.55 trillion, bolstered in no small part by confidence in the automotive sector.
Nvidia has collaborated with many carmakers, including Toyota, Volvo, BYD, Li Auto, Lucid, NIO, Rivian and General Motors, to develop AI-powered autonomous-driving and advanced driver-assistance systems. That’s part of its estimated $5 billion automotive business in 2025, according to Bloomberg Intelligence. On Jan. 5, the company announced details about its new and unique MB.Drive Assist Pro for Mercedes-Benz, the next generation of AI-based driving.
But what is it like to experience a personal car being driven completely by AI? I tried it in December in San Francisco riding with Joe Taylor, a senior systems engineer at Mercedes, in the automaker’s new $47,250 Mercedes-Benz CLA. It has the most advanced AI driving system you can buy in the U.S. in a private vehicle to date.
My takeaway? An hour in a car driven by AI makes everything else feel, well, dumb. Once consumers feel safe relinquishing control of their daily commute, I predict they will eagerly adopt AI-driven cars, which will become status symbols among those in urban centers.
The trial consisted of two portions. In the first part, I drove the car using a less-advanced system, MB.Drive Assist, which uses a combination of highly intelligent algorithms and end-to-end AI models to help the car drive itself on highways and city streets. It is already available across the entire Mercedes lineup, but it still needs occasional human input like a hand on the steering wheel. The service costs $1,950, a onetime fee.
I engaged it by pressing a button on the left side of the steering wheel, which let the car creep forward into traffic, eventually steering itself through complex situations like a multilane left turn onto a main road. It avoided unpredictable drivers in intersections, navigated the generalized chaos of roundabouts and adjusted its speed around construction sites, all without my hands on the steering wheel.
In fact, I stayed hands-off for chunks of time, occasionally being prompted to touch the steering rack to let the car know that I was still engaged. (Sensors in the steering wheel detect your hands there; flashing icons and sound alerts ensue if you go too long without touching it.) Mostly I kept my hands on my knees, taking in the view as the car crested steep hills and held in place while women with strollers crossed the street — a surreal pause that kept me a little anxious. I hovered my foot over the brake pedal, just in case.
At one point, the CLA didn’t slow down and move over fast enough for my taste when approaching a double-parked cargo truck outside a cafe; I quickly braked and steered left. Other times, it appeared to simply turn off the system when it didn’t like the environment, like entering a valet lane at a hotel. A dashboard icon alerted me when it disengaged; the car would roll to a halt at that point unless and until I started operating it again.
MB.Drive Assist left me thinking I could easily become too disconnected from the act of driving, when in reality I still needed to be present and alert. Driving with the more expensive, more advanced and not-yet-available MB.Drive Assist Pro inspired more confidence, since the car is consistently in charge when the program is running. It’s already operating in China and will be available in the U.S. by the end of the year. A three-year subscription costs $3,950.
MB.Drive Assist Pro is extreme SAE-Level 2 driving, which means it doesn’t require hands on the steering wheel at any time. The car will drive itself completely from a starting point to a final destination, even though the driver must still remain attentive (eyes on the road) because they are still legally responsible for safe-vehicle operation. This was the closest I felt to the Waymo rides I took last fall — but far different, too, because this was a normal-looking car made for anyone to buy, not an awkward-looking fleet vehicle.
The leap forward here is that the driving being done by the car can be collaborative if and when I want to jump in. Drive Assist Pro weaves in any input on the steering wheel, if I decide to intervene, without canceling out the whole program. That means I can make minor adjustments from the driver’s seat if I want and it’s not a big deal. Mercedes calls it “our philosophy of human-machine collaboration.”
Ali Khan, the director of product marketing at Nvidia, described the car as “AI-defined,” a vehicle informed, enhanced and enabled by AI to its very core. The humanistic element of this level of AI is essential, he says, so that “the car sees everything and understands what it sees.”
MB.Drive Assist Pro uses 10 cameras, five radar sensors and 12 ultrasonic sensors that provide raw data to a supercomputer that makes sense of those massive data streams. It employs Nvidia’s AI end-to-end stack for core driving tasks, plus a parallel classical safety stack — built on Nvidia’s Halos safety system — that adds built-in redundancies, fail-safe checks and other safety guardrails. (“End-to-end stack” means the whole layered system was developed over an entire life cycle of building, deploying and training the AI, from initial data collection to final integration into the car. Halos ensures the vehicle operates within defined safety parameters.) All of which is to say: The more you drive it, the better it gets, since the AI is constantly learning from the data it gathers on every drive.
I felt comfortable inside the cabin because the car moved with total authority. It anticipated the road, avoided potholes and deep rain puddles, and it never hesitated in sticky traffic. It expertly swam in San Francisco’s sea of jaywalkers and zippy Waymos; it sailed through roundabouts and accommodated errantly parked delivery vans without a hitch. Most impressive was its ability to know the nuances of good driving, like when it rolled through yellow lights and crept forward a smidge to evade other traffic, correctly reading street conditions and behaving accordingly.
The highlight? It knew when to turn right on a red stoplight and when to remain stationary until the light turned green if right-on-red was signposted as forbidden at that particular intersection.
We haven’t quite reached the AI utopia Sam Altman foresees, even in the world of cars. The new voice-activated virtual assistant — which is powered by Google Cloud’s Automotive AI Agent — ignored my repeated requests that the car stop reading news headlines during my drive. It also flubbed questions about whether it was in fact ChatGPT, and didn’t respond when I asked it to reduce its own speaking volume to a conversational level. It never did figure out how to navigate me back to the 1 Hotel San Francisco, either, instead trying to offer me options for “one hotel in the city.” I ended up using my smartphone.
Safety is paramount — the public perception of which is as critical as the reality that AI-driven vehicles are in far fewer lethal accidents than those driven by humans. For all its advances, MB.Drive Assist Pro works only on city streets, not highways, for now.
Still, AI driving is authentically here, able to intuit, anticipate and adapt with elegance and ease. The jump from rote algorithm to humanlike comprehension has me believing that for many people, learning to trust an AI-defined car will be a welcome relief from the doldrums of daily commuting.
I’ll still want the option of driving my own car, of course, but I won’t miss sitting in gridlock traffic, hands glued to the steering wheel while my eyes glaze over. A machine that can handle that unpleasant task while I read or answer emails? Yes, please.
