Rivian announced on 22 October that it has acquired Third Dimension AI, a UK start-up that turns real-world images and video into controllable 3D simulation environments. The technology will be used to generate synthetic driving data, recreate rare and dangerous scenarios and train and validate its driver-assistance and autonomous driving systems. The team of about 15 people joins Rivian's Autonomy & AI organisation. Financial terms were not disclosed.

What it means: So-called world models, systems that rebuild the real world in digital form so algorithms can be tested inside it, are becoming a core tool for autonomous driving. The reason is simple: the most critical scenarios, such as a child running into the road or a truck shedding its load, are rare and too dangerous to collect on real journeys. A US carmaker buying a small European team shows strong demand for skills in 3D reconstruction and synthetic data. For Greek computer vision, graphics and simulation teams, of which there are quite a few in universities and research centres, it is a sign that specialised tools can find a buyer or partner among large manufacturers. The same technology has uses beyond cars, for example in warehouse robotics, drones or infrastructure inspection, where data on rare scenarios is also scarce. At the same time, in Europe the validation of such systems will be increasingly tied to vehicle safety rules and the AI Act, so documenting how synthetic data is produced and checked now carries real value.

Read the original on Business Wire (Rivian)

Related service: Custom software

Related stories