Today at the Nvidia GTC AI Conference in San Jose, CA, automotive mega supplier Magna International announced a program to integrate the Nvidia Drive AGX platform within the company’s next generation of advanced technology solutions. The collaboration will help Magna enable cutting-edge AI (artificial intelligence) and varying levels of autonomous driving (AD) and interior cabin applications.

The Drive AGX Thor system-on-a-chip (SoC) runs the safety-certified DriveOS operating system and is built on the Blackwell GPU architecture, consolidating increased functionality to improve efficiency, speed, and scalability. Using the accelerated compute performance and scalability afforded by the next-generation SoC, Magna will develop and test enhanced vehicle safety and comfort technologies such as the latest advancements in SAE Level 2+ through Level 4 active safety solutions.

“Combining Nvidia accelerated compute and AI capabilities with Magna’s extensive automotive expertise and innovation, we aim to explore new standards for next-generation software-defined vehicle intelligence and autonomy,” said Steven Jenkins, Vice President of Technology Strategy at Magna Electronics. “Our collaboration allows us to develop market applications for AI-powered solutions that could redefine the driving experience and address the evolving demands of the automotive industry.”

For a pilot program with Nvidia, Magna will leverage its expertise in developing, integrating, and launching advanced features such as adaptive cruise control, driver and occupant monitoring, Highway and Urban Navigate-on-Autopilot, and interior cabin AI companion functionalities. The initiative focuses on solving the complexity of computing availability with performance for integrating ADAS (advanced driver assistance system), AD, and interior cabin AI features to provide scalable, flexible, and customizable system solutions to meet market needs and regulatory requirements.

“As the automotive industry transitions to safer, more intelligent vehicles with autonomous driving capabilities, our collaboration with Magna is the latest in our endeavors to bring our safety-certified in-vehicle accelerated compute and AI to the transportation industry,” said Ali Kani, Vice President of Automotive at Nvidia. “By combining core technologies and Magna’s integration expertise, we aim to shape the future of mobility.”

Thor delivers up to 1000 TOPS (trillion operations per second) of AI compute power, featuring 8-bit floating point support optimized for transformer models, large language models, and generative AI workloads.

Magna plans to unveil a working demonstration platform in Q4 2025, tapping Nvidia’s DriveOS for development, integration, validation, and production workflows.

The latest development with Nvidia highlights a core focus for Magna on active safety. According to the company, addressing the challenge of detecting and classifying potential threats in these demanding environments is crucial for the next generation of ADAS.

The key to overcoming these limitations may lie in the fusion of thermal imaging and Doppler-type imaging radar. Futurride covered Magna’s latest innovations in thermal imaging for AEB here.

By combining the heat-signature-detecting advantages of thermal technology with imaging radar’s ability to determine an object’s location, speed, and direction, the advanced system can provide quicker, more reliable environmental analysis under adverse conditions. The next-generation technology could augment or even replace conventional detection systems as well as complex lidar-based systems.

In addition to multiple sensing modalities, AI-driven machine learning strengthened by the new Nvidia collaboration could help to process sensor data earlier through early fusion and more accurately track pedestrians, animals, and other objects.