NVIDIA Jetson AGX Thor Developer Kit
NVIDIA Jetson AGX Thor Developer Kit
Jetson Developer Kits
Jetson Thor AGX Developer Kit

The Jetson Thor Developer kit delivers unmatched real-time performance for robotics in a compact, power-efficient design, powered by the NVIDIA Blackwell GPU and 128GB memory for up to 2070 FP4 TFLOPS within a 130W envelope—achieving 7.5x higher AI performance and 3.5x better efficiency than Jetson AGX Orin.

Features:

Integrates Blackwell Multi-Instance GPU (MIG) technology and a 14-core Arm Neoverse-V3AE CPU, plus advanced accelerators such as the third-generation PVA, optical flow accelerator, and dual encoders/decoders for low-latency, real-time workloads.

Provides extensive I/O, including a QSFP slot (4×25 GbE), Multi-GbE RJ45, and multiple USB ports for high-speed sensor fusion and seamless humanoid-robotics integration.

Optimized for generative reasoning and multimodal sensor processing, supporting leading AI frameworks like Isaac GR00TN, NVIDIA Metropolis, Holoscan, and Cosmos Reason for tasks such as video search and summarization.

Supported by a comprehensive partner ecosystem providing carrier boards, sensors, and system software across robotics, retail, industrial, medical, and smart-space applications, enabling rapid prototyping of next-generation physical AI systems.

Jetson AGX Thor Developer Kit Included Components
NVIDIA Jetson AGX Thor Developer Kit
What are the recommended uses for Jetson AGX Thor Developer Kit?

The Jetson AGX Thor can be used for multiple purposes such as Artificial Intelligence, Autonomous Machines, Edge Computing, Generative AI, Humanoid Robotics, Industrial Automation, Physical AI, Robotics, Vision AI.

Jetson Developer Kits Technical Specifications
Jetson AGX Thor Developer Kit Jetson AGX Orin Developer Kit Jetson Orin Nano Super Developer Kit
AI Performance 2070 TFOPS (FP4-Sparse) 275 TOPS 67 TOPS
GPU 2560-core NVIDIA Blackwell architecture GPU with fifth-gen Tensor Cores Multi-Instance GPU (MIG) with 10 TPCs 2048-core NVIDIA Ampere architecture GPU with 64 Tensor Cores 1024-core NVIDIA Ampere architecture GPU with 32 Tensor Cores
GPU Max Frequency 1.57 GHz 1.3 GHz 1020MHz
CPU 14-core Arm® NeoverseR-V3AE 64-bit CPU
1 MB L2 cache per core
16 MB shared system L3 cache
12-core Arm® Cortex-A78AE v8.2 64-bit CPU
3MB L2 + 6MB L3
6-core Arm® Cortex-A78AE v8.2 64-bit CPU
1.5MB L2 + 4MB L3
CPU Max Frequency Tick Crimefighter Sorta Tick
The 2.6 GHz 2.2 GHz 1.7 GHz
Memory 128 GB 256-bit LPDDR5X 273 GB/s 64GB 256-bit LPDDR5 204.8GB/s 8GB 128-bit LPDDR5 102 GB/s
Storage 1 TB NVMe M.2 Key M Slot 64GB eMMC 5.1 (SD Card Slot & external NVMe via M.2 Key M)
Power 40W-130W 15W-60W 7W-15W-25W