A robot swerved through the cafeteria of Rivian’s Palo Alto office, its shelves adorned with chilled canned coffees, until it suddenly stopped. Five minutes later, a man carefully pushed it out of the way, the words “I’m stuck” flashing yellow on its screen. It was an inauspicious start to Rivian’s “Autonomy & AI Day,” a showcase for the company’s plans to make its vehicles drive themselves. Rivian does not make that cafeteria robot, but there was a familiar message in its failure: this technology is hard.
Hours later, during a 15-minute demo of Rivian’s new “Large Driving Model” in a 2025 R1S SUV, I was reminded of that message. The electric vehicle, equipped with the automated-driving software, carried myself and two Rivian employees on a switchback route near the company’s campus. As we glided past Tesla’s engineering office, I noticed a Model S in front of us slow to turn into the rival company’s lot. The R1S eventually noticed this too, braking hard just before the Rivian employee nearly intervened.
There was one actual disengagement during my demo. The employee in the driver’s seat took over as we passed through a one-lane section of road blocked by tree-trimming equipment. It was minor overall, but not exactly rare; I observed multiple other demo rides that also had disengagements.
The rest of the drive went well enough for software that is not ready to be shipped, especially considering Rivian threw out its old rules-based driver assistance system and adopted an end-to-end approach, similar to how Tesla developed its system. It stopped at stoplights, handled turns, and slowed for speed bumps all without programmed rules telling it to do so.
Rivian’s old system was very deterministic and structured, according to CEO RJ Scaringe. Everything the vehicle did was the result of a prescribed control strategy written by humans. He explained that when Rivian saw transformer-based artificial intelligence taking off in 2021, he quietly reconstituted the team and started from a clean sheet to design a self-driving platform for an AI-centric world.
After extensive development, Rivian launched the new ground-up 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 started to see dramatic progress once the data began pouring in.
Rivian is betting it can train its Large Driving Model on fleet data so quickly that it will roll out “Universal Hands-Free” later this month. This will allow Rivian owners to take their hands off the wheel on 3.5 million miles of roads in the U.S. and Canada, so long as there are visible painted lines. In the second half of 2026, Rivian plans to enable “point-to-point” driving, which is the consumer version of the demo ride.
By the end of 2026, after Rivian begins shipping its smaller, more affordable R2 SUVs, it will replace the Nvidia chips and outfit those vehicles with a new custom autonomy computer unveiled at the event. That computer, plus a lidar sensor, is intended to eventually allow drivers to take their hands and eyes off the road. True autonomy, where a driver does not have to worry about retaking control, lies well beyond that and will depend on how fast Rivian can train its model.
This rollout introduces a near-term challenge. The new autonomy computer and lidar will not be ready until months after the R2 goes on sale. Customers who want a vehicle capable of eyes-off driving will have to wait. The R2 is a crucial product for Rivian, especially following declining sales of its first-generation vehicles.
Scaringe acknowledged that when technology moves this fast, there is always some level of obsolescence. He stated the company wants to be direct about what is coming. The early R2 models will still get the promised point-to-point hands-off driving, but not eyes-off capability. He believes some customers will wait for the more advanced features, while others will buy the R2 immediately and potentially trade it in later. He expressed confidence that there is sufficient demand backlog for the R2 that customers can make their own choice.
When interviewed in 2018, before Rivian had even revealed its vehicles, Scaringe shared a goal that has endured. He wanted to make Rivian’s vehicles so capable that if you went for a hike starting at one point and finishing at another, the vehicle could meet you at the end of the trail. It was a pie-in-the-sky promise common seven years ago, but one that felt true to Rivian’s brand of aspirational adventure.
Scaringe said Thursday he still believes it is possible for Rivian to enable a use case like that in the next few years. It will not happen until the company builds and tests its more-capable R2 vehicles, which is at least a year away in a best-case scenario. He said that while it has not been a huge focus, it could change as the company approaches higher levels of autonomy, training its model on trickier roads without clear lane lines.
He suggested that handling dirt roads or off-road trails to a trailhead could eventually be feasible. However, he was clear that the company is not putting any resources into autonomous rock crawling, such as navigating a place like Hell’s Gate in Moab.