NVIDIA Jetson Orin Nano – Embedded AI Module (4 GB / 8 GB)
Key features
Compact module in Jetson Nano / Orin NX format, easy to integrate into compact embedded systems
NVIDIA Ampere GPU with Tensor Cores for running edge AI models
6-core Arm Cortex-A78AE CPU up to 1.5 GHz
Up to 67 TOPS of AI performance depending on the version
LPDDR5 memory with high bandwidth for intensive workloads
NVMe storage support via PCIe
Compatible with the NVIDIA software ecosystem: JetPack, Isaac, Metropolis, Holoscan, TAO Toolkit, Omniverse Replicator
Pin-compatible with Jetson Orin NX for scalable designs
Applications
The NVIDIA Jetson Orin Nano is intended for applications requiring advanced AI computing capabilities in a small footprint. It is particularly suitable for:
autonomous robotics
smart cameras and embedded vision
video analytics and retail analytics
automated inspection and industrial QA
drones and compact mobile systems
edge generative AI embedded systems
Jetson Orin Nano version comparison
The Jetson Orin Nano family is available in two complementary versions. The 4 GB version is intended for more memory- and budget-efficient integrations, while the 8 GB version is better suited to heavier AI workloads, concurrent pipelines, and more demanding models.
Module
GPU
CPU
AI Performance
Memory
Power Consumption
Jetson Orin Nano 4 GB
Ampere – 512 CUDA + 16 Tensor Cores
6-core Arm Cortex-A78AE
34 TOPS
4 GB LPDDR5
7 – 25 W
Jetson Orin Nano 8 GB
Ampere – 1024 CUDA + 32 Tensor Cores
6-core Arm Cortex-A78AE
67 TOPS
8 GB LPDDR5
7 – 25 W
A high-performance platform for Edge AI
The Jetson Orin Nano is a particularly attractive solution for integrating embedded AI into compact products. Thanks to the NVIDIA JetPack software stack and its associated ecosystem, it enables faster development of robotics, computer vision, video analytics, and advanced sensor processing applications.
Technical specifications of the NVIDIA Jetson Orin Nano board
GPU architecture: NVIDIA Ampere
GPU:
4 GB version: 512 CUDA cores + 16 Tensor Cores
8 GB version: 1024 CUDA cores + 32 Tensor Cores
CPU: 6-core Arm Cortex-A78AE v8.2 64-bit
CPU cache: 1.5 MB L2 + 4 MB L3
Max CPU frequency: up to 1.5 GHz
Max GPU frequency: up to 625 MHz
AI performance:
4 GB version: up to 34 TOPS
8 GB version: up to 67 TOPS
Memory:
4 GB version: 4 GB LPDDR5, 64-bit bus, 34 GB/s
8 GB version: 8 GB LPDDR5, 128-bit bus, 68 GB/s
Storage: external NVMe storage support
Cameras:
up to 4 cameras (or 8 via virtual channels)
8 MIPI CSI-2 lanes
D-PHY 2.1 up to 20 Gbps
Video encoding: 1080p30, supported by 1 to 2 CPU cores
Video decoding:
1 x 4K60 (H.265)
2 x 4K30 (H.265)
5 x 1080p60 (H.265)
11 x 1080p30 (H.265)
PCI Express: 1 x4 + 3 x1, PCIe Gen3, Root Port and Endpoint