Nvidia on Tuesday unveiled the Jetson Orin Nano 2, an entry-level edge AI computer aimed at robotics applications. The Santa Clara-based company said the module delivers up to 78 trillion operations per second with 8GB of memory, running on an 8-core Arm CPU.
The new device offers twice the inference performance of its predecessor, the Jetson Orin Nano Super, while operating at 40% lower power in 15-watt mode. Nvidia retained the compact form factor of earlier models, targeting space-constrained deployments.
The system supports large language models and vision-language models optimized for edge inference, including Nvidia Cosmos, Nvidia Nemotron, Gemma 4 and Qwen 3. More than 3 million developers have built on Nvidia’s robotics stack, the company said.
Wing, an Alphabet subsidiary focused on drone delivery, plans to evaluate the Jetson Orin Nano 2 for its fleet. Dinuka Abeywardena, head of perception at Wing, said the hardware could enable faster, more energy-efficient drones capable of real-world understanding. Matic Robotics, a home cleaning robotics firm, will adopt the module for its platform, according to Navneet Dalal, cofounder and CEO of Matic.
Cognex and Doosan Bobcat are among other companies exploring the system for their applications. Nvidia said the Jetson Orin Nano 2 module and developer kit are slated for release in the first half of 2027.
