MediaTek and Nvidia Join Forces to Develop Chips for Self-Driving Vehicles with Advanced Graphics Capabilities

MediaTek will collaborate with Nvidia to produce chips for self-driving vehicles. The company is aiming to develop various versions of the Dimensity System on a Chip (SoC) specifically tailored for the automotive industry. These chips will feature an integrated Graphics Processing Unit (iGPU) capable of managing multiple displays within cars.

MediaTek has introduced four new variants of the Dimensity Auto Cockpit SoCs including the CV-1, CM-1, CY-1, and the CX-1. Each variant offers increased power and is designed for different automotive applications. These SoCs are primarily designed to oversee the internal systems of the vehicle, but MediaTek claims that they can also help in controlling cameras and sensors for assisted driving functionalities. 

The collaboration with Nvidia allows these chips to work seamlessly with Nvidia's Drive OS and support the company's TensorTR SDK and Nsight analysis tool. The SoCs are built on ARM's V9-A architecture and feature an integrated GPU from Nvidia, enabling the management of infotainment displays and other systems running on Nvidia's Drive OS. Additionally, the SoCs can handle speech recognition, integrate artificial intelligence into voice commands, and manage multiple internal displays simultaneously. MediaTek states that the chips support the operation of multiple virtual machines at the same time, allowing users to run Linux-based systems and Android Auto concurrently.

According to MediaTek, the built-in image processor is capable of real-time processing of High Dynamic Range (HDR) images. The SoC also includes a 5G modem, Bluetooth, Wi-Fi 7, and support for various satellite navigation and communication protocols.