← All hardware
275 TOPS of robotics-grade AI in a 110mm cube.
Pros
- Proven, mature platform with deep ecosystem support
- Large 64GB memory and high bandwidth for complex pipelines
- Single carrier board emulates the whole Orin family
Cons
- Pricey at $1,999
- Now eclipsed in raw compute by Jetson Thor
- 60W peak power needs active cooling and robust supply
✓ Where it shines / best for
- Prototyping advanced autonomous robots and AMRs
- Multi-sensor, multi-stream perception and edge inference
- Developers building on the NVIDIA Jetson/Isaac robotics stack
✕ Not the best fit for
- Budget or hobbyist projects (Orin Nano kits are far cheaper)
- Large generative/transformer models that need Thor-class compute
- Battery-only ultra-low-power designs
Features
- ✓ Up to 275 INT8 TOPS of AI compute
- ✓ 2048-core NVIDIA Ampere GPU with 64 Tensor Cores
- ✓ 12-core Arm Cortex-A78AE CPU
- ✓ 64GB LPDDR5 unified memory
- ✓ Configurable power modes: 15W, 30W, 50W, plus MAXN mode up to 60W
- ✓ Runs the full NVIDIA JetPack SDK (CUDA, cuDNN, TensorRT, Isaac, DeepStream)
- ✓ Emulates any Jetson Orin module for prototyping; rich I/O (PCIe, USB, CSI camera lanes, GPIO)
Pricing
| Plan | Price | Billing | Notes |
|---|---|---|---|
| AGX Orin 64GB Developer Kit | $1,999 | one-time | MSRP; ships with 64GB module on reference carrier board, 90W USB-C supply, Wi-Fi/BT NIC |
Pricing verified from the official source. Prices change often — confirm on the vendor's site before buying.
Specifications
| power | 15-60W |
| memory | 64GB LPDDR5, 204.8 GB/s |
| form_factor | 110mm cube developer kit |
| architecture | NVIDIA Ampere GPU (2048 CUDA cores, 64 Tensor cores) + 12-core Arm Cortex-A78AE |
| ai_performance | 275 INT8 TOPS |
Sponsored
A full review is being generated for this product and will appear here shortly.