Rivian Acquires UK Startup to Turn Customer Driving Footage into 3D Simulations

Rivian has acquired Third Dimension, a UK startup of 15 people, as part of an effort to rebuild real roads as 3D test environments for its self-driving system. TechCrunch first reported the deal. Rivian did not disclose the purchase price.

Third Dimension was founded in 2024 by a former Tesla Autopilot simulation engineer and a veteran of Wayve's neural simulator team. It described itself as a generative AI tool that builds 3D environments ready for game engines. When it raised a $7 million seed round led by Felicis in October 2024, it listed game studios, film, the US military and autonomous vehicle companies as potential customers.

The founders' backgrounds point to the autonomy angle. CEO Tolga Kart spent more than two years at Tesla working on Autopilot simulation, after five years as a senior director on Call of Duty at Activision, and later led engineering at synthetic data company Parallel Domain. Cofounder Piotr Sokólski spent four years at Wayve building photorealistic neural simulators for self-driving cars.

The entire team is joining Rivian's autonomy and AI group, which is led by James Philbin. Because most of the staff are based in the UK, the deal also expands Rivian's presence in London. "Simulation is central to how we train and validate our autonomous driving systems," Philbin told TechCrunch, adding that Third Dimension "has built impressive technology that turns real-world data into detailed 3D environments."

The approach draws on footage from owners of the R1S, R1T, R2 and Rivian's commercial vans who opt into camera data sharing. Philbin said the fleet sends back more than 400,000 driving sequences every day. Most of that material is uneventful, while the situations that most challenge autonomous systems are rare. Examples include a mattress in the fast lane, a police officer waving traffic through a red light, or a child running out between parked cars. Rather than wait months for such events to appear, engineers can recreate real scenes in 3D and alter the variables.

Rivian is not the first to pursue this. Waymo unveiled its Waymo World Model, built on Google DeepMind's Genie 3, in February to generate scenarios its vehicles have not encountered. Tesla has also shown a neural "world simulator" for its Full Self-Driving software.

Time pressure adds urgency. In March, Uber agreed to invest up to $1.25 billion in Rivian to deploy R2 robotaxis, though only $300 million is paid up front. The remainder depends on autonomy milestones, beginning with 10,000 driverless R2 vehicles on Uber's network in San Francisco and Miami from 2028.

Consumer timelines have already shifted. Rivian's point-to-point driving feature is due to reach a small group of invited Autonomy+ owners later this year, and those testers must sign a nondisclosure agreement. "Eyes-off" driving, which CEO RJ Scaringe promised for 2026 in January 2025, is now targeted for 2027. Level 4 capability is planned for 2028 on the Gen 3 hardware Rivian showed at its AI & Autonomy Day.

Electrek's assessment was that the deal looks sensible: small, targeted, and bringing in a team that has done similar work at Tesla and Wayve. It noted that simulation is a training and testing tool rather than proof of safety, and that a system can perform well on generated edge cases yet still fail on a real street. It added that rebuilding real scenes from real data is about as close to the real thing as simulation gets.