The robot wandered Rivian’s Palo Alto office cafeteria, with shelves lined with iced coffees—until it didn’t. Five minutes later, a man carefully pushed it out of everyone’s way, the words “I’m stuck” flashing yellow on the poor droid’s screen.
It was an inauspicious start to Rivian’s “Autonomy & Artificial Intelligence Day,” a showcase for the company’s plans to make its vehicles capable of driving themselves. Rivian doesn’t make the cafeteria robot and isn’t responsible for its abilities, but there was a familiar message in its weaknesses: this thing is hard.
Hours later, as I drove a 2025 R1S SUV during the 15-minute demo of Rivian’s new self-proclaimed “Large Driving Model,” I remembered that message.
The EV equipped with the self-driving software led me and two Rivian employees on a commuting route near the company’s campus. As we glided past Tesla’s engineering office, I noticed a Model S in front of us that was slowly turning into a rival company’s lot. The R1S eventually noticed this too, braking hard just before the Rivian employee intervened.
During the demo drive, there was a real breakout. The employee in the driver’s seat took over as we passed a section of the road with one lane due to tree cutting. Little things overall. But it wasn’t exactly rare. I spotted a lot other test runs which he had releases very.
The rest of the unit did pretty well for software that isn’t ready to ship, especially when you consider that Rivian ditched its old rules-based driver assistance system and adopted an end-to-end approach — which is how Tesla developed Full Self-Driving (Supervised). It stopped at stoplights, handled turns, slowed down for speed bumps, all without programmed rules telling it to do these things.
A quiet axis in 2021
Rivian’s old system “was very deterministic and it was very structured,” CEO RJ Scaringe said in an interview Thursday. “Everything the vehicle did was the result of a prescribed control strategy written by humans.”
Techcrunch event
San Francisco
|
13-15 October 2026
Scaringe said that when Rivian saw transformer-based AI taking off in 2021, it quietly “restructured the team and started with a clean slate and said, let’s design our autonomous platform for an AI-centric world.”
After spending “a lot of time in the basement”, Rivian launched its new advanced driving software in 2024 on its second-generation R1 vehicles, which use Nvidia’s Orin processors.
Scaringe said it was only recently that his company began to see dramatic progress “once the data really started to break down.”
Rivian is betting that it can train its Large Driving Model (LDM) on fleet data so quickly that it will allow the company to develop what it calls “Universal Hands-Free” driving by early 2026. That means Rivian owners will be able to take their hands off the wheel on 3.5 million miles of roads in the US and Canada (as long as the tracks are visible). In the back half of 2026, Rivian will enable point-to-point driving, or the consumer version of the demo we received on Thursday.
The “eyes on” challenge on “hands down”.
By the end of 2026, after Rivian began shipping its smaller, more affordable R2 SUVs, it will abandon Nvidia chips and equip those vehicles with a new custom range computer unveiled Thursday. This computer, along with a lidar sensor, will eventually allow drivers to take their hands and eyes off the road. True autonomy – where a driver doesn’t have to worry about taking control of the vehicle again – lies well beyond that, and will largely depend on how quickly Rivian can train its LDM.
This disposition introduces a near-term challenge for Rivian. The new range computer and lidar won’t be ready until months after R2 launches. If customers want a vehicle that can handle driving without eyes (or more), they’ll have to wait. But the R2 is a critical product for Rivian, and the company needs it to sell well — especially after declining sales of its first-generation vehicles.
“When technology is moving as fast as it is, there’s always going to be some level of obsolescence, and so what we want to do here is be really upfront” about what’s coming, Scaringe said. Early R2s will still have Rivian’s promised point-to-point ride, which will be based on the new software and will be seamless, but not on the eyes.
“So [if] you buy an R2 and you buy it for the first nine months, it’s just going to be more limited,” he said. “I think what will happen is some customers will say ‘this is very important to me and I will wait.’ And some people will say “I want the newest, best stuff now, and I’ll get R2 now, and maybe I’ll trade it in a year or two and get the next version later. Fortunately, there is so much backlog of demand for the R2 that we believe that, with that in mind, customers can make the decision for themselves.”
“In a perfect world, everything is simultaneous, but the timing of the vehicle and the timing of the autonomy platform just aren’t perfectly aligned,” he said.
When I first interview Scaringe in 2018, before Rivian even showed what their vehicles looked like, shared a goal that’s still rattling around in my head. He wanted to make Rivian’s vehicles so capable of driving themselves that: “if you go on a hike and you start at one point and end at another point, you have the vehicle meet you at the end of the trail.”
It was the kind of promise for self-driving cars that was all the rage seven years ago, but it stuck with me at least because it felt true throughout Rivian’s ambitious adventure.
Scaringe told me Thursday that he still thinks it’s possible for Rivian to enable such a use case in the next few years. It certainly won’t happen until the company tests and builds its more capable R2 vehicles, which is at least a year away at best.
“We could [do that]. It wasn’t a huge focus,” he said. That could change as the company approaches Level 4 autonomy, however, since by then the company will have trained its LDM on more difficult roads without guiding features such as lane lines.
“Then, it becomes a bit like, what is ODD [operational design domain]? Dirt roads, off road? Easy,” he said. Just don’t expect a Rivian to drive itself The Gate of Hell in Moab.
“We don’t have the resources for autonomous rock detection,” he said. “But as far as getting to the trailhead? Sure.”
