NVIDIA Launches DRIVE AGX Thor Developer Kit for Automotive Developers Worldwide

Time：2025-09-10
View volume：23


image.png


The complexity of intelligent vehicle systems is increasing day by day, with the integration of vision-language-action (VLA) models for inference, generative AI, and advanced sensor technologies. This trend places higher demands on developers, who are in urgent need of powerful, efficient development tools that meet automotive-grade safety standards.


The NVIDIA DRIVE AGX Thor Developer Kit is now open for pre-order and will ship in September. This kit provides developers and researchers worldwide with an advanced platform to accelerate the design, testing, and deployment of intelligent vehicle and smart mobility solutions.


Built on the NVIDIA Blackwell architecture, the next-generation Arm Neoverse V3AE CPU, and the NVIDIA DriveOS 7 software stack, this developer kit is specifically designed for vision-language-action (VLA) models for inference, making it an ideal platform for automotive development. Its abundant I/O interfaces support sensors such as surround-view cameras, radars, and lidars, and are compatible with mainstream in-vehicle interfaces including GbE/10GbE and PCI-Express. Additionally, DRIVE AGX Thor complies with the automotive functional safety integrity level certification (ISO 26262) and cybersecurity certification (ISO 21434).


The Growing DRIVE AGX Thor Ecosystem


NVIDIA's ADAS (Advanced Driver Assistance Systems) partners — DeepRoute.ai, Nuro, WeRide, and Zoox — are building their ADAS software platforms based on DRIVE AGX Thor. The mass-production systems for DRIVE AGX Thor are supplied by tier-1 suppliers, including Continental, Desay SV, 

Lenovo, Magna, and Quanta.


In addition, DRIVE AGX Thor is gaining support from a growing number of leading companies in sensor and embedded technologies, such as AdaCore, Lauterbach, OMNIVISION, QNX, and Vector.


ADAS Safety from Cloud to Vehicle End


DRIVE AGX Thor and DriveOS are specifically designed to achieve automotive-grade functional safety and cybersecurity, and are key components of NVIDIA Halos. The NVIDIA Halos comprehensive safety system combines NVIDIA's automotive hardware, software security solutions with cutting-edge AI research in the field of ADAS (Advanced Driver Assistance Systems) safety.