Joining Rivian’s Race to Autonomy | TechCrunch

The robot swerved around the cafeteria of Rivian’s Palo Alto office, where the shelves were decorated with refrigerated canned coffee, until it stopped doing so. Five minutes later, a man gently pushed him out of the way, the words ‘I’m stuck’ flashing yellow on the poor droid’s screen.

It was an inauspicious start to Rivian’s “Autonomy & AI Day,” a showcase for the company’s plans to make its vehicles self-driving. Rivian doesn’t make the cafeteria robot and isn’t responsible for its capabilities, but there was a familiar message in its weaknesses: This stuff is hard.

Hours later, as I drove a 2025 R1S SUV during my 15-minute demo of Rivian’s new self-proclaimed “Large Driving Model,” I was reminded of that message.

The EV, equipped with automated driving software, drove myself and two Rivian employees on a winding route near the company’s campus. As we floated past Tesla’s engineering office, I saw a Model S ahead of us slowly merging into the rival company’s turf. The R1S eventually noticed this too and braked hard, just before the Rivian employee almost intervened.

During my demo ride, there was one actual withdrawal. The employee in the driver’s seat took over as we drove through a one-lane stretch of road due to tree trimming. Little things in general. But it wasn’t exactly rare either; I saw several other demo rides that had withdrawals at.

The rest of the ride went well enough for software that isn’t ready to ship yet, especially considering Rivian has thrown away its old rules-based driver assistance system and adopted an end-to-end approach – which is how Tesla developed Full Self-Driving (Supervised). He stopped at traffic lights, took turns, slowed down at speed bumps, all without any programmed rules telling him to do these things.

A quiet pivot in 2021

Rivian’s old system “was all very deterministic and it was all 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.”

Scaringe said that when Rivian saw transformer-based artificial intelligence taking off in 2021, he quietly “reassembled the team, started with a clean slate and said, let’s design our self-driving platform for an AI-centric world.”

After spending “a lot of time in the basement,” Rivian launched the new core operating software on its second-generation R1 vehicles in 2024, which use Nvidia’s Orin processors.

Scaringe said his company only recently started seeing dramatic progress “when the data really started pouring in.”

Rivian is betting that it can train its Large Driving Model (LDM) on fleet data so quickly that it will allow the company to roll out ‘Universal Hands-Free’ driving in early 2026. That means Rivian owners can take their hands off the wheel on 3.5 million miles of roads in the US and Canada (as long as there are visible painted lines). In the second half of 2026, Rivian will enable point-to-point driving, or the consumer version of the demo we received on Thursday.

The ‘eyes off’ to ‘hands off’ challenge

By the end of 2026, after Rivian starts shipping its smaller, more affordable R2 SUVs, it will ditch the Nvidia chips and equip those vehicles with a new custom autonomy computer unveiled Thursday. That computer, plus 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 regaining control of the vehicle – is well beyond that and will largely depend on how quickly Rivian can train its LDM.

This rollout introduces a near-term challenge for Rivian. The new autonomy computer and lidar won’t be ready until months after the R2 goes on sale. If customers want a vehicle that can drive eyes-free (or more), they’ll have to wait. But the R2 is a crucial product for Rivian, and the company needs it to sell well — especially in the wake of declining sales of its first-generation vehicles.

“When technology evolves so quickly, there’s always going to be some degree of obsolescence, and so what we want to do here is be very direct” about what’s to come, Scaringe said. The early R2s will still get Rivian’s promised ‘point-to-point’ driving, which will be based on the new software and will be hands-off but not eyes-off.

“So [if] If you buy an R2 and buy it in the first nine months, the limitations only become more limited. I think what will happen is that some customers will say, ‘That’s very important to me, and I’m going to wait.’ And some will say, “I want the latest, greatest stuff now, and I’m going to buy the R2 now, and maybe trade it in in a year or two, and get the next version later.” Fortunately, there is so much demand backlog for R2 that we think that by being honest about this, customers can make the decision for themselves.”

“In a perfect world, everything is running at the same time, but the vehicle timeline and the autonomy platform timeline are just not perfectly aligned,” he said.

When I interviewed first Scaringe in 2018, before Rivian even showed off what its vehicles looked like, he shared a goal that still rattles around in my head. He wanted to make Rivian’s vehicles capable of driving themselves so that: “if you’re taking a walk, and you start at one point and you end at another point, the vehicle will meet you at the end of the path.”

It was the kind of heavenly promise about self-driving cars that was all the rage seven years ago, but it stuck with me, at least because it was something that fit with Rivian’s whole kind of ambitious adventure.

Scaringe told me Thursday that he still thinks it’s possible for Rivian to enable such a use case in the coming years. That certainly won’t happen until the company tests and builds its more capable R2 vehicles, which will take at least another year at best.

“We could [do that]. Not much attention has been paid to it,” he says. That could change as the company gets closer to Level 4 autonomy, as by then the company will have trained its LDM on trickier roads without guiding features like lane lines.

“Then it becomes a bit like: what is the ODD [operational design domain]? Dirt roads, off-road? Easy,” he said. Just don’t expect a Rivian to drive itself up The gate of hell in Moab.

“We are not putting any resources into autonomous rock crawling,” he said. “But as far as getting to the trailhead? Sure.”