Anonymous
Mar 14, 2026
Rating: 4/5
Great Board, Steep Learning Curve
Hardware is fantastic, but optimizing code to fully utilize the Blackwell architecture requires deep knowledge of CUDA. Once you get it working, though, the performance is stellar.
Description
Reviews
| Brand | - |
| Category | Robotics / STEM Teaching Tools / Development, Learning, Evaluation & Industrial Control Boards |
| Origin | - |
| Processor | NVIDIA Thor SoC with Blackwell GPU architecture |
| AI Performance | Up to 2000 TOPS (INT8) |
| CPU | ARM Cortex-A78AE v8.2 64-bit CPU |
| Memory | 64GB LPDDR5 |
| Storage | Support for NVMe SSD via M.2 Key M slot |
| Connectivity | Dual 10GbE, Wi-Fi 6E, Bluetooth 5.3 |
| Video I/O | 8x MIPI CSI-2 camera interfaces |
| Display Output | HDMI 2.1, DisplayPort 1.4a |
| Power Consumption | Configurable TDP: 15W - 70W |
| Operating System | NVIDIA JetPack SDK (Linux based) |
Anonymous
Mar 14, 2026
Rating: 4/5
Great Board, Steep Learning Curve
Hardware is fantastic, but optimizing code to fully utilize the Blackwell architecture requires deep knowledge of CUDA. Once you get it working, though, the performance is stellar.
Anonymous
Mar 14, 2026
Rating: 5/5
Incredible Power for Autonomous Robots
The AGX Thor is a beast. We are using it for our next-gen warehouse robot and the inference speed is unmatched. The 2000 TOPS allows us to run multiple complex perception models simultaneously without lag.
Anonymous
Mar 12, 2026
Rating: 4/5
Solid Developer Kit
The kit arrives well-packaged with all necessary cables. Setup was straightforward using the SD card image. Only minor issue was initial driver compatibility with an older Ubuntu version, but quickly resolved.
Anonymous
Mar 11, 2026
Rating: 5/5
Perfect for Research Labs
Our university lab upgraded from the Orin to the Thor. The development environment is seamless, and the documentation from NVIDIA is top-notch. Highly recommended for serious AI research.
Anonymous
Mar 11, 2026
Rating: 5/5
Future-Proof Investment
Expensive, yes, but worth every penny for industrial applications. The thermal management is efficient even under full load, and the connectivity options cover all our sensor needs.
Anonymous
Mar 01, 2026
Rating: 5/5
Best in Class for Edge AI
We tested several competitors, and nothing comes close to the Thor's density and efficiency. It handles real-time video analytics for four 4K streams effortlessly.
Q: Is this compatible with previous Jetson sensors?
A: Most MIPI CSI-2 sensors designed for Jetson Orin and Xavier are compatible, but we recommend checking the specific sensor driver support in the latest JetPack SDK release notes.
Q: What is the warranty period for the developer kit?
A: The NVIDIA Jetson AGX Thor Developer Kit comes with a standard one-year limited hardware warranty from the date of purchase.
Q: Can I run Windows on this board?
A: No, the NVIDIA Jetson platform is optimized for Linux (Ubuntu) and runs the JetPack SDK. It does not natively support Windows OS.
Q: Does it support Docker containers?
A: Absolutely. The JetPack SDK includes full support for Docker, and many pre-built AI container images are available via the NVIDIA NGC catalog.
Q: Does this kit include a power supply?
A: Yes, the Development Kit includes a compatible AC adapter and power cord suitable for the board's maximum TDP configuration.