NVIDIA Jetson AGX Orin specifications, memory configuration, and storage upgrade options. The Jetson platform features unified memory shared between CPU and GPU, optimised for AI and machine learning workloads.
Memory Specifications
The NVIDIA Jetson AGX Orin uses unified LPDDR5 memory shared between the CPU and GPU cores for AI inference workloads.
| Specification | Details |
|---|---|
| Memory Type | LPDDR5 |
| Memory Capacity | 8GB / 16GB (shared CPU+GPU) |
| Upgradeable | No — soldered |
| GPU Cores | 1024 CUDA cores |
| AI Performance | Up to 100 TOPS (INT8) |
| Storage | microSD + NVMe via carrier board |
Storage Upgrade Options
The best way to improve NVIDIA Jetson AGX Orin performance is through storage upgrades:
- NVMe SSD via carrier board M.2 slot for fast model loading
- microSD for OS and basic storage
- USB 3.0 SSD for dataset storage
- Consider storage speed carefully for ML training data pipelines
Recommended Products
Samsung 970 EVO Plus 500GB NVMe
Fast NVMe for Jetson carrier board. Good for ML model storage.
Check Price on Amazon UK
Fast NVMe for Jetson carrier board. Good for ML model storage.
Check Price on Amazon UK
SanDisk Extreme 64GB microSD (A2)
Reliable microSD for Jetson OS installation.
Check Price on Amazon UK
Reliable microSD for Jetson OS installation.
Check Price on Amazon UK
Frequently Asked Questions
Can I upgrade NVIDIA Jetson AGX Orin RAM?
No. NVIDIA Jetson AGX Orin has soldered memory that cannot be upgraded after manufacture. Choose the highest RAM option at purchase for maximum flexibility.
What is the best storage for NVIDIA Jetson AGX Orin?
An NVMe SSD via the carrier board M.2 slot provides the fastest storage for model loading and datasets.
Recommended NVMe SSDs
Western Digital
Western Digital Green WD ™ SN3000 NVMe™ SSD POWERED BY SANDISK 500 GB
£138.02
View Deal
Western Digital
Western Digital Black WD_BLACK™ SN7100 NVMe™ SSD POWERED BY SANDISK 1
£160.55
View Deal
Western Digital
Western Digital Black WD_BLACK™ SN850X NVMe™ SSD POWERED BY SANDISK 8
£2,420.32
View Deal
