Rivian has acquired Third Dimension, a UK artificial intelligence startup whose software converts images and videos into lifelike 3D worlds. The Rivian Third Dimension acquisition is intended to sharpen the simulation tools that Rivian uses to train and validate automated driving systems. Financial terms were not disclosed.
What Third Dimension built
Third Dimension was founded in 2024 and had raised $7 million in seed funding from investors including Abstract Ventures, MVP, Soma Capital and Solari Capital. The company employs 15 people, most of them based in the UK. Its technology began in video game development, but its founders saw potential in defense and autonomous vehicle applications.
Under Rivian, that work will narrow to automated driving. The entire Third Dimension team is joining Rivian’s autonomy and AI group, which is led by James Philbin. Tolga Kart, Third Dimension’s co-founder and CEO, said the team was eager to concentrate on a single problem, adding that running a small company for many different customers is difficult. Piotr Sokólski is the startup’s other co-founder.
Why simulation matters to Rivian
Self-driving systems depend on simulation to train and test software before it reaches public roads. Traditional simulators rely on rendering or game engines and synthetic data, which can leave a gap between the simulated world and the physical one. Third Dimension’s approach uses AI to reproduce the physical world more closely, according to the companies’ description of the technology.
Philbin said the goal is to make real scenarios simulatable. In his account, the simulated events are drawn from real-world situations rather than game-style synthetic scenes, and the system can alter camera movement, the positions of other agents, lighting and weather. That flexibility lets Rivian test complex driving conditions without waiting for rare edge cases to occur naturally on the road.
Rivian also gathers driving data from customer vehicles whose owners have enabled camera sharing. The data is aggregated and anonymized, and Rivian receives more than 400,000 sequences uploaded daily from its customer fleets. Simulation complements that stream by letting engineers replay and vary situations that would be impractical to stage in real life.
The validation question
Philbin framed simulation as a way to judge whether a system is ready to ship. For that to work, engineers must be able to trust the sim-to-real gap, the difference between simulated conditions and actual driving. Rivian evaluated many providers before selecting Third Dimension, Philbin said.
The acquisition also supports Rivian’s plan to expand its AI hub in London, a step Philbin linked to the deal. The startup’s UK base fits that effort.
Software roadmap and the Uber robotaxi
Rivian already offers Universal Hands-Free driving software for highways, though drivers must still monitor the road. The company plans an end-of-year release that adds hands-free operation on surface streets as well as highways, with driver supervision still required.
Further out, Rivian aims to field fully autonomous R2 SUVs as robotaxis on the Uber app in 2028. Simulation is central to that timeline because it allows Rivian to train and validate its models against a wide range of conditions before deployment.
The deal’s price and other terms remain undisclosed, as do the exact release date for the end-of-year software and the details of Rivian’s full autonomy system. Rivian has made clear, however, that it is not entering the gaming business; the 3D world technology is being absorbed into its autonomy and AI team.
