NVIDIA has released the Jetson Orin Nano Super Developer Kit, a compact generative AI supercomputer. The device, which measures small enough to fit in one's hand, provides increased performance for generative AI capabilities.
The Jetson Orin Nano Super delivers up to 1.7 times the generative AI inference performance of its predecessor, with a 70% increase in INT8 performance to 67 TOPS and a 50% increase in memory bandwidth to 102GB/s. It features an NVIDIA Ampere architecture GPU with tensor cores, a 6-core Arm CPU, and supports up to four cameras with higher resolution and frame rates.
The platform is compatible with NVIDIA’s AI software ecosystem, including tools such as NVIDIA Isaac for robotics, NVIDIA Metropolis for vision AI, and NVIDIA TAO Toolkit for model fine-tuning. It also integrates with NVIDIA Omniverse Replicator for synthetic data generation and benefits from an active developer community and partner ecosystem for additional support.
The Jetson Orin Nanoconsistes of 8GB system-on-module (SoM) and a reference carrier board, suitable for prototyping edge AI applications. Additionally, a software update will also boost generative AI performance for existing Jetson Orin Nano Developer Kit owners, as well as for the Jetson Orin NX and Orin Nano series of systems on modules, by up to 1.7 times. This update is available through the JetPack SDK.
This announcement has drawn significant attention from the tech community. Graham Cooke, founder of Amorphic, shared on his X account:
This means AI can run without connecting to the cloud. And it's about to spark the biggest tech battle of our time.
While Marios Karatzias highlighted the implications for robotics, stating:
We are entering the era of robots, so NVIDIA ‘cooked’ a processor suitable for #robotics.
The Jetson Orin Nano Super Developer Kit makes generative AI at the edge more accessible, unlocking new possibilities for robotics, vision AI, and multimodal applications. It is now available through NVIDIA authorized distributors worldwide.
