Autonomy, robotaxis, and Mind Robotics

Rivian’s pioneering full self-driving technology (Level 5). Its Gen 3 Autonomy Computer processes 5 billion pixels per second. Rivan’s path to autonomy now includes 10 HDR cameras, 5 radars, a RAP1 processor running AI, and proprietary LiDAR- Light Detection and Ranging laser pulses that create a real-time 3D depth map. Capable of 200 trillion operations per second, Rivian’s in-house processor and in-vehicle inference platform currently enable hands-free assisted driving.

Rivian’s self-improving data flywheel takes human driving nuances, organizes, filters, and labels the data sets, applies generative AI and reinforcement learning to synthesize driving behaviors, and pushes refined capabilities to customer vehicles via OTA updates in near real-time. Altogether, Rivian’s features make full self-driving possible.

“There’s an in-vehicle inference platform that’s doing the real-time decisioning but we’re able to capture unique events occurring as we drive- across our fleet- that are sent into a cloud and used to train our much larger offline model” -Rivian Founder RJ Scaringe.