Like fellow semiconductor giants, Intel and Qualcomm, Nvidia joined the party at CES 2022 where it showcased the latest generation of its Drive Hyperion 8 automotive stack. It is a suite of hardware solutions that will power self-driving cars and trucks. Hyperion 8 includes the latest Drive Orin system on chip, 12 surround cameras, nine radars, 12 ultrasonic sensors, one front-facing LiDAR, and three interior cameras. Of course, the Orin chip crunches the data from all these sensors, radars, cameras, and LiDAR in real-time on-device.
Nvidia says the platform has been designed in such a way that there are multiple layers of redundancies baked in ensuring that a self-driving car can transport its passengers safely. Volvo and its subsidiary Polestar have announced that they will be integrating the Drive Hyperion 8 stack in their vehicles. Later today, Volvo along with LiDAR maker Luminar which whose hardware Nvidia integrates into Drive Hyperion 8 will share more information about what they are doing together.
The mapping system will embedded in Nvidia Drive
Nvidia also has dealt with Chinese players like Nio, Xpeng, Li Auto, R auto, and IM motors - all of whom have adopted Drive Hyperion 8. In fact, considering Nvidia's hardcore chops in autonomous driving it even counts robotaxi startups like Amazon-owned Zoox, DiDi Chuxing, GM-backed Cruise, and trucking services like Navistar and Plus as its customers. Autonomous trucking company TuSimple has also built its latest platform on the Drive Orin autonomous chipset.
Nvidia also showcased Drive Concierge, an always-on, digital assistant for drivers which is part of the software that the world's most valuable semiconductor maker has been developing. Its Omniverse Avatar created an agent which leverages Nvidia's advances in speech recognition, computer vision, natural language processing, using recommendation engines and simulation technologies. It basically acts like Siri or the Google Assistant for the car and Nvidia even showed a demo of the restaurant reservation being done by the assistant using just speech.
