Rivian Buys AI Startup to Simulate Self-Driving Edge Cases

Rivian is betting on simulation to speed up its self-driving development. Rivian announced it has acquired Third Dimension AI, a startup that uses AI to turn real-world driving data into detailed 3D environments that can be edited and replayed at scale.

Financial terms weren’t disclosed, but Third Dimension’s roughly 15-person team, most of whom are based in the U.K., will join Rivian’s Autonomy research and engineering teams.

What Third Dimension AI Does

Third Dimension was founded in 2024 by Tolga Kart, who previously worked at Tesla and EA, Piotr Sokolski, and Özgun Pelvan. Its platform reconstructs real-world scenes from sensor data, then lets engineers recreate tricky driving situations and change things like lighting, weather, and other road users. That means Rivian’s autonomy models can train on rare edge cases without waiting for them to happen on real roads.

“Simulation is central to how we train and validate our autonomous driving systems,” said James Philbin, Rivian’s SVP of Autonomy and AI. He added that the technology will help both Rivian’s consumer and robotaxi plans, including the R2 robotaxis it’s building for Uber that are due in 2028.

Rivian Wave

In a post on X, Kart said he and Sokolski started the company believing “the future of simulation would be grounded in reality.” He added that the acquisition was a competitive process, and that one reason they chose Rivian is that they’ll “get to put what we’ve built directly in the loop with real vehicles, real-world data and an autonomy stack.”

A Different Approach From Tesla

Rivian still relies on real-world data. It collects more than 400,000 driving sequences a day from customer vehicles with camera sharing enabled. Third Dimension’s tech lets Rivian get more out of that data by turning it into scenarios it can tweak and rerun.

Tesla, meanwhile, leans more on the size of its fleet. It feeds billions of miles of real-world driving footage into its training models and finds edge cases in actual footage from vehicles around the world. As FSD improves, more people use it, which gives Tesla even more data to train on. That creates a data flywheel that keeps feeding itself, with less reliance on computer simulation.

Point to Point Is Coming Soon

The Third Dimension acquisition comes shortly after Rivian announced an early-access program for its Point to Point hands-free driving feature. Select R2 and Gen 2 R1 owners with Autonomy+ will start getting it in December in California and Florida. Point to Point is Rivian’s answer to Tesla’s FSD (Supervised), letting the vehicle autonomously follow your navigation route through turns, highway transitions, and city streets.

After that, Rivian plans to start rolling out Level 3 autonomy next year, beginning with eyes-off highway driving. Third Dimension’s simulation tech will likely play a quiet role in validating Point to Point and Rivian’s other autonomy features before they reach every Autonomy+ subscriber in 2027 and beyond.