NVIDIA unveils Jetson NanoX: sub-10W edge AI board for real-time vision and robotics

Tech · 3 min read

The Jetson NanoX delivers a new Ampere-lite GPU and an energy-optimized ARM core pairing designed to run common vision stacks, transformer-based perception models, and small LLMs under a 10W thermal envelope. NVIDIA positioned the board between the Jetson Nano and Xavier lines for cost-sensitive edge deployments.

Alongside the hardware, NVIDIA is shipping an updated JetPack with a containerized runtime optimized for quantized vision models and real-time ROS integrations. The company emphasized improved driver stability and pre-built modules for obstacle avoidance, multi-camera stitching, and pose estimation.

Early partners include several robotics startups and an autonomous inventory robot maker. Jetson NanoX will be available to developers next month with volume pricing for OEMs later in the quarter.