NVIDIA® Jetson AGX Xavier
- NVIDIA Jetson AGX Xavier
Overview
NVIDIA Jetson AGX Xavier brings data-center-class AI to embedded form factors. It combines a powerful GPU, AI accelerators, and an 8-core CPU with rich I/O for multi-sensor systems. Use it when you need more parallel inference, higher frame rates, and advanced perception on device.
Why Choose Jetson AGX Xavier?
- Up to 32 TOPS class AI performance in a compact module
- 8-core Arm CPU and NVIDIA GPU with Tensor Cores for modern AI workloads
- Flexible power profiles to match thermal and battery targets
- Camera-dense designs with CSI-2 lanes for multi-sensor pipelines
- Mature Jetson software stack, tools, and ecosystem
Key Features
- High throughput AI: Run multiple DNNs in parallel for perception, tracking, and analytics
- Sensor-fusion-ready: Abundant CSI-2 and high-speed I/O for cameras, lidar, and radar
- Rugged integration: Small footprint with integrated heat spreader simplifies mechanical design
- Production-grade software: JetPack SDK, CUDA, TensorRT, and containerized deployment
- Lifecycle support: Long-term availability for industrial programs
Why These Modules?
Typical Applications
- Mobile robots and AMRs
- Inspection and quality control
- Industrial vehicles
- Smart city analytic
- Autonomous systems
- Multi-camera vision
- Sensor fusion and mapping
Carriers and Cameras
- Carriers: Connect Tech Rogue and Rogue-X carrier boards for AGX Xavier
- Camera expansion: MIPI boards for Allied Vision Alvium, FPD-Link III, SDI, and HDMI ingest
- Storage and comms: NVMe via M.2, multiple USB 3.x, dual Ethernet options on supported carriers
Ordering Information
- Jetson AGX Xavier: 16 GB or 32 GB module options
Contact us for availability, lead time, recommended carriers, thermal solutions, and camera kits.
Choose Your Module
Selection Guidance
- Choose Jetson AGX Xavier when you need higher model complexity, more camera streams, or faster decision loops than Xavier NX or Orin Nano class modules
- Pair with a production-ready carrier to shorten time to market, then add camera boards to match your sensor stack
- Technical Specifications
- Resources
Technical Specifications
| Category | Interfaces |
|---|---|
| Form factor | 100 mm x 87 mm module with integrated thermal transfer plate |
| AI performance | Up to 32 TOPS class depending on model and power mode |
| CPU | 8-core Arm v8.2 |
| GPU | NVIDIA architecture with Tensor Cores |
| AI accelerators | Dual NVDLA engines |
| Memory | 16 GB or 32 GB LPDDR4x options |
| Storage | 32 GB eMMC on module |
| Camera I/O | Up to 16 lanes MIPI CSI-2 |
| High-speed I/O | PCIe Gen 4, USB 3.x, Gigabit Ethernet |
| Video | Multi-stream 4K encode and decode |
| Typical power | Profile selectable in the 10 W to 30 W range |
Values are typical for selection and may vary by exact module and software release.
Resources
Similar Products
Macnica Americas Site Map
