Rivian acquires Third Dimension to accelerate generative AI world models for autonomous driving
The 15-person London startup converts video telemetry into physics-consistent 3D simulation environments to train robotaxis and driver-assist systems.
Technology news
Electric vehicle manufacturer Rivian announced Thursday that it has acquired Third Dimension, a London-based generative AI startup specializing in creating real-time 3D world models from video and sensor streams. The transaction absorbs Third Dimension’s 15-person engineering team directly into Rivian’s core autonomy and artificial intelligence group led by Chief Autonomy Officer James Philbin.[4]
Founded in 2024 by gaming and computer vision veterans Tolga Kart and Piotr Sokólski, Third Dimension previously raised $7 million in seed funding to build physics-accurate 3D reconstruction engines. While initially explored for video game production, the startup’s technology converts real-world dashcam and multi-camera fleet video into interactive, photorealistic 3D simulation environments that allow autonomous driving systems to test rare edge cases, severe weather conditions, and collision scenarios virtually without requiring physical test track driving.[4]
Rivian plans to deploy Third Dimension’s world modeling pipelines to accelerate development across two strategic pillars: its proprietary Universal Hands-Free advanced driver-assistance system for consumer R1 and upcoming R2/R3 vehicles, and the custom autonomous robotaxi fleet it is contracted to manufacture and deploy in partnership with Uber. The acquisition also establishes an expanded artificial intelligence research hub for Rivian in central London, complementing its existing engineering teams in Palo Alto and Irvine.[4]
Our take: Rivian’s acquisition of Third Dimension highlights how the competitive frontier in autonomous mobility has shifted from raw fleet mileage collection to generative world simulation. By synthesizing infinite, physics-compliant 3D driving environments from fleet video, Rivian can stress-test safety-critical edge cases that physical vehicle fleets rarely encounter. This generative approach gives Rivian the technical autonomy needed to compete with Tesla’s FSD and Waymo while preparing its vehicle hardware for Uber’s commercial robotaxi network.