Rivian Wants to One Day Build Cars That Drive Themselves

Rivian CEO RJ Scaringe says self-driving technology could let the company build very different cars from the SUVs and trucks it sells now. He said it in an interview at TechCrunch Disrupt, reported on Oct. 19, 2026. TechCrunch
He was talking about the future. Beyond the R2 and R3, Scaringe is thinking about a self-driving car that a person could own. That opens new design options. The company has not described those designs. The basic point is that once steering controls and constant driver attention are no longer required, designers have more freedom.
Rivian is working toward that in steps. It has shipped a driver-assistance system that allows hands-off driving while the driver watches the road. It has promised software that can handle point-to-point trips without intervention, with the driver supervising, by the end of 2026. Full self-driving is planned for robotaxi versions of the R2 SUV, which will be available on Uber's ride-hail app. TechCrunch
That plan sits next to factory growth. Rivian is at the start of scaling up production of the R2 SUV. Public customer deliveries of R2 and orders for reservation holders began June 9, 2026. The company showed the full R2 lineup with versions and pricing on March 12, 2026.
The base design was made with this path in mind. Rivian announced a midsize electric-car platform for the R2, R3 and R3X in March 2024, using high pressure die castings to combine many parts into fewer pieces. At the time it said R2 and R3 would have much stronger self-driving ability. In December 2025, Scaringe laid out a plan for Rivian cars to drive themselves more over time, based on custom computer chips and lidar, a laser sensor that maps what is around the car. TechCrunch
The taxi part is already agreed. Uber and Rivian partnered to put up to 50,000 fully self-driving robotaxis on the road, announced March 19, 2026. Scaringe previewed the Disrupt appearance in August, when TechCrunch said he would explain Rivian's bets on electric cars, robots and autonomy together. He also discussed R2 demand and self-driving work with CNBC in July 2026.
The broader context here is the gap between help with driving and no driver at all. First hands-off driving while you watch, then a full trip with you supervising, then a taxi with no driver in set areas. That order lets a company test sensors, road decisions and backup safety with a person present. It also splits two separate product jobs.
In my view, the job Scaringe named is the more interesting one. A car you still supervise needs a driver's seat, controls and a clear view. A car you own that drives itself would not, in principle. It becomes more like a small living room on wheels. Doors, seat direction, storage, power and cooling for computers and sensors, and screens all change once the driving task is gone.
Worth flagging is that new freedom brings new limits. Laser sensor placement and cleaning, backup power and brakes, easy computer repair, wiring layout, and crash safety with new seating positions become the hard packaging work. Custom chips help with power and cost, but the car still must stay clean, lined up and working across ten years of private ownership. Fleets can handle that with service depots. Private owners cannot.
Looking at what this means for Rivian specifically, order matters. R2 is the volume and cost test. Supervised driving is the data and safety test. A robotaxi R2 on Uber is the daily operations test. Only after those three does a brand-new owned self-driving car make factory sense. Scaringe is describing step four while step one is still ramping. That is normal as product planning, if read as planning.
On a personal note, I have watched my own children go from car seats to driver's licenses to calling rides on phones, and each step changed what a car was for. Autonomy could change it again, from a machine built around driving to a space built around time. If Rivian gets the middle steps right, it will have earned the chance to design for that.


