Skip to main content

Worldwide Shipping · Secure Payments

AI Summary

Refurbished NVIDIA Jetson AGX Orin 32GB module delivering up to 200 TOPS of AI performance in a compact system-on-module. It combines an 8-core Arm CPU, Ampere GPU, and 32GB RAM for advanced robotics and edge AI development.

Key Highlights

  • Up to 200 TOPS AI performance with 15W–40W configurable power
  • 8-core Arm Cortex-A78AE CPU at 2.2 GHz
  • NVIDIA Ampere GPU with 1792 CUDA cores and 56 Tensor Cores
  • 32GB RAM and 64GB eMMC 5.1 storage
  • Supports up to 6 CSI cameras and 10GbE networking
  • Compact 100mm x 87mm module with 699-pin Molex connector

Ideal For

  • Robotics developers building autonomous machines
  • AI engineers prototyping edge inference solutions
  • Computer vision specialists working with multi-camera setups
  • Researchers needing high-performance embedded AI
  • Product teams developing smart vision systems

Buying Tip

Verify that your carrier board supports the 699-pin Molex Mirror Mezz connector and can supply the required 15W–40W power before purchasing.

Elecvora expert review

Our verdict

The refurbished Jetson AGX Orin 32GB module offers exceptional AI performance for its size, making it a cost-effective choice for advanced robotics and edge AI projects. It is well-supported by NVIDIA's software ecosystem, though buyers must ensure carrier board compatibility.

AI-assisted analysis based on the manufacturer's specifications. Not a hands-on lab test.

Pros

  • High AI performance up to 200 TOPS
  • Flexible power configuration from 15W to 40W
  • Rich I/O including 10GbE and PCIe Gen4
  • Compact form factor

Cons

  • Requires a compatible carrier board
  • Refurbished condition may have cosmetic signs of use

Best for

AI developers and robotics engineers needing high compute in embedded systems

Not ideal for

Hobbyists without carrier board or thermal management experience

Specifications

RAM32GB
Processor Type8-core Arm® Cortex®-A78AE v8.2
GPUNVIDIA Ampere architecture with 1792 NVIDIA CUDA® cores and 56 Tensor Cores
CPU8-core Arm® Cortex®-A78AE v8.2 64-bit CPU 2MB L2 + 4MB L3
CPU Frequency2.2 GHz
Storage Capacity64GB eMMC 5.1
Accelerometer Type2x NVDLA v2.0
CSI CameraUp to 6 cameras (16 via virtual channels*) 16 MIPI CSI-2 lanes D-PHY 2.1 (up to 40Gbps) | C-PHY 2.0 (up to 164Gbps)
Video encoder1x 4K60 | 3x 4K30 | 6x 1080p60 | 12x 1080p30 (H.265) H.264, AV1
Video decoder1x8K30 | 2x4K60 | 4x4K30 | 9x1080p60 | 18x1080p30(H.265) H.264, VP9, AV1
UPHYUp to 2 x8, 1 x4, 2 x1 PCIe Gen4 (Root Port and Endpoint) 3x USB 3.2
Networking1x GbE, 1x 10GbE
Display1x 8K60 multi-mode DP 1.4a (+MST)/eDP 1.4a/HDMI 2.1
Power Rating15W - 40W
Size (mm)100mm x 87mm 699-pin Molex Mirror Mezz Connector Integrated Thermal Transfer Plate

Frequently Asked Questions

What does refurbished mean for this product?
This module has been professionally refurbished and was previously used for testing purposes only by the Elecvora team. The product seal was opened for testing, but it remains in proper working condition.
Do I need a carrier board to use this module?
Yes, this is a system-on-module and requires a compatible carrier board with a 699-pin Molex Mirror Mezz connector to operate.
What power supply is required?
The module operates on a power supply configurable between 15W and 40W. Ensure your carrier board and power source can deliver the necessary wattage.
What software is supported?
The module supports the NVIDIA AI software stack, including Isaac for robotics, DeepStream for vision AI, and Riva for conversational AI, along with standard Linux distributions.
How many cameras can I connect?
The module supports up to 6 cameras directly via 16 MIPI CSI-2 lanes, and up to 16 cameras via virtual channels.
Is there a warranty?
No warranty information is provided for this refurbished product. Please contact the seller for specific terms.

Overview

The NVIDIA Jetson AGX Orin 32GB Module is a powerful system-on-module that brings data-center-class AI performance to embedded and edge devices. This refurbished unit has been professionally tested by the Elecvora team and remains in proper working condition, with only the seal opened for testing. It delivers up to 200 TOPS of AI compute, making it suitable for demanding robotics, autonomous machines, and vision AI applications.

Key Features

  • High AI throughput: Up to 200 TOPS for complex neural networks.
  • Power flexibility: Configurable between 15W and 40W to balance performance and efficiency.
  • Advanced processing: 8-core Arm Cortex-A78AE CPU at 2.2 GHz and Ampere GPU with 1792 CUDA cores and 56 Tensor Cores.
  • Ample memory: 32GB RAM and 64GB eMMC 5.1 storage for large models and datasets.
  • Rich connectivity: PCIe Gen4, USB 3.2, 1x GbE, 1x 10GbE, and support for up to 6 CSI cameras.
  • Video capabilities: Encode up to 4K60 and decode up to 8K30 with H.265, H.264, AV1, and VP9 support.

Specifications Explained

The 32GB of RAM allows you to run multiple AI models and large datasets simultaneously, which is essential for tasks like multi-sensor fusion and real-time video analytics. The 8-core Arm CPU handles general-purpose computing, while the Ampere GPU accelerates parallel workloads. With 1792 CUDA cores and 56 Tensor Cores, the GPU is well-suited for deep learning inference. The module's 64GB eMMC provides fast onboard storage for the operating system and applications. Networking includes both 1GbE and 10GbE for high-bandwidth data transfer. The 699-pin Molex Mirror Mezz connector ensures a secure interface to a compatible carrier board.

Compatibility and Requirements

This module requires a compatible carrier board with a 699-pin Molex Mirror Mezz connector. It operates on a power supply configurable from 15W to 40W. The module supports the NVIDIA AI software stack, including Isaac for robotics, DeepStream for vision AI, and Riva for conversational AI. Ensure your carrier board and thermal solution are designed for the Jetson AGX Orin module.

Typical Uses

Developers and engineers use this module for autonomous robots, drones, intelligent video analytics, medical imaging, and industrial automation. It is also ideal for prototyping and research in edge AI, thanks to its high performance and flexible I/O.

Quick specs

RAM
32GB
Processor Type
8-core Arm® Cortex®-A78AE v8.2
GPU
NVIDIA Ampere architecture with 1792 NVIDIA CUDA® cores and 56 Tensor Cores
CPU
8-core Arm® Cortex®-A78AE v8.2 64-bit CPU 2MB L2 + 4MB L3
CPU Frequency
2.2 GHz
Storage Capacity
64GB eMMC 5.1
See all specifications

More from NVIDIA

View all →

Related Products

View all →