NOW AVAILABLE FOR PRE-ORDER

AGX OrinField Kit

275 TOPS of AI performance.64GB unified memory.Production-ready out of the box.

The ultimate edge AI platform for robotics, autonomous systems, and industrial applications. Pre-configured with 30+ AI frameworks and applications.

275TOPS
AI Performance
64GB
LPDDR5 Memory
2048cores
CUDA Cores
100+hrs
Setup Time Saved

We believe

"The future of AI

isn't in

the cloud—

it's at the

edge,

where

decisions

happen in

milliseconds..."

275TOPS
AI Performance
<10ms
Inference Latency
0cloud
Dependency

Everything You Need

Complete professional edge AI platform. Unbox and start deploying immediately.

1
275 TOPS

Jetson AGX Orin 64GB

NVIDIA's flagship edge AI computer with 275 TOPS performance, 2048 CUDA cores, and 64GB unified memory.

2
7000 MB/s

512GB NVMe SSD

High-speed storage with pre-configured JetPack 6.2, AI frameworks, and 30+ production-ready applications.

3
802.11ax

WiFi 6 + Bluetooth

Dual-mode wireless with AP+STA parallel operation for simultaneous hotspot and internet connectivity.

4
65W

Power Supply Unit

Industrial-grade 65W power adapter with international plugs. Reliable power delivery for all workloads.

5
Full Set

Complete Cable Kit

USB-C cables, Ethernet cable, and GPIO breakout accessories for rapid prototyping and deployment.

6
Included

Developer Support

Access to documentation, GitHub repos, Discord community, and 90 days of email support.

Pre-Integrated Frameworks

All frameworks optimized for AGX Orin's Ampere architecture

NVIDIA
Docker
Ultralytics
Roboflow
Ollama
OpenAI
Home Assistant
LiveKit

Technical Deep Dive

Detailed specifications of every component in the AGX Orin Field Kit.

NVIDIA Jetson AGX Orin Developer Kit

The most powerful AI computer for autonomous machines. The Jetson AGX Orin delivers server-class AI performance with up to 275 TOPS in an embedded form factor, enabling breakthroughs in robotics, advanced vision, and generative AI at the edge.

275
TOPS INT8
Peak AI Performance

Core Specifications

  • AI Performance275 TOPS (INT8)
  • GPU2048-core Ampere + 64 Tensor Cores
  • CPU12-core Arm Cortex-A78AE
  • Memory64GB LPDDR5 (256 GB/s)
  • DL Accelerator2x NVDLA v2.0
Power Modes
15-60W
Vision Accelerator
PVA v2.0
Storage
64GB eMMC
Form Factor
110x110mm
2.75x More Powerful Than Orin Nano

Run larger LLMs, multiple concurrent vision pipelines, and complex multi-modal AI workloads. The AGX Orin is designed for production-grade robotics, industrial automation, and advanced research applications.

Pre-configured NVMe SSD

High-performance 512GB NVMe SSD with fully optimized JetPack 6.2 environment. Pre-loaded with CUDA-accelerated frameworks, AI models, and production-ready applications. Save over 100+ hours of setup and compilation time.

100+ Hours Saved

Skip configuration hell. Start building immediately.

512GB
Storage
7000
MB/s Read

AI Frameworks

All major frameworks compiled and optimized for AGX Orin's architecture.

  • PyTorch 2.3 + TorchVision
  • TensorRT 10.x Optimizations
  • ONNX Runtime + cuDNN
  • llama.cpp (CUDA)

Ready-to-Run Apps

Production-grade applications pre-installed and tested.

  • Ollama + Open WebUI
  • DepthAI + OAK-D SDK
  • Roboflow Inference Server
  • YOLO v8/v11 Ultralytics

Pre-installed Software Stack

JetPack 6.2
OS
CUDA 12.6
Compute
cuDNN 9.x
DNN
TensorRT 10
Inference
Docker + Compose
Container
ROS 2 Humble
Robotics
Jupyter Lab
Dev
VS Code Server
IDE

Dual-Mode Wireless Connectivity

AP + STA Parallel Mode

Simultaneous Hotspot & Internet Connection

The Field Kit is optimized to run both Access Point (AP) and Station (STA) WiFi modes in parallel. Create a seamless hotspot for direct device control while maintaining an internet connection.

Access Point (AP)

Broadcasts a dedicated network for direct low-latency control from your laptop, tablet, or robot controller.

Station Mode (STA)

Connects to existing WiFi networks for internet access, OTA updates, cloud telemetry, and remote SSH access.

Included Connectivity
WiFi 6 (802.11ax)
Bluetooth 5.0
Gigabit Ethernet
USB 3.2 Gen 2
40-pin GPIO
CAN Bus Support
Limited Launch Pricing

Ready to Build the
Future of Edge AI?

Join researchers, roboticists, and engineers who are pushing the boundaries of what's possible with edge AI.

Stay in the Loop

Updates on new features, edge AI tutorials, and product announcements.

No spam. Unsubscribe anytime.

All specifications subject to change. Performance may vary based on configuration, workload, and environmental conditions. NVIDIA, Jetson, CUDA, and TensorRT are trademarks of NVIDIA Corporation.