NVIDIA Jetson AGX Thor Developer Kit

Next-gen Blackwell-powered edge AI platform — 2,070 TFLOPS FP4, 2,560 CUDA cores, 128GB LPDDR5X, 14-core Arm Neoverse-V3AE CPU. 7.5x more AI compute than Jetson AGX Orin.

🚀 Express Shipping Available Across Europe & MENA

  • Check Mark Full Insurance on All Shipments
  • Check Mark Tracked Delivery & Real-Time Updates
GUARANTEED SAFE CHECKOUT
  • Stripe
  • Visa Card
  • MasterCard
  • American Express
  • Discover Card

Overview

The NVIDIA Jetson AGX Thor Developer Kit is the most powerful Jetson development platform ever created. Powered by the NVIDIA Blackwell architecture, it delivers 2,070 TFLOPS of FP4 sparse AI performance with a 2,560-core GPU and 14-core Arm Neoverse-V3AE CPU — providing 7.5x more AI compute and 3.5x better energy efficiency than the Jetson AGX Orin.

With 128GB of LPDDR5X unified memory, native FP4 quantization through a next-generation Transformer Engine, and Multi-Instance GPU (MIG) support, the Jetson AGX Thor is purpose-built for the most demanding edge AI applications: humanoid robots, autonomous vehicles, advanced medical devices, and industrial automation systems that require real-time AI reasoning.

Key Features

  • 2,070 TFLOPS FP4 Sparse: Blackwell GPU with fifth-generation Tensor Cores for edge AI at data center-class performance
  • 2,560 CUDA Cores: Massive parallel compute with MIG support (10 TPCs)
  • 128GB LPDDR5X: 273 GB/s unified memory for large AI models and sensor fusion
  • 14-Core Arm Neoverse-V3AE: High-performance CPU at 2.6 GHz for complex application logic
  • 7.5x Over Orin: Generational leap in AI compute for next-gen autonomous systems
  • 40W–130W Configurable: Flexible power envelope for diverse deployment scenarios
  • Next-Gen Transformer Engine: Dynamic FP4/FP8 switching for optimal model performance

Technical Specifications

Specification Details
GPU Architecture NVIDIA Blackwell
GPU Cores 2,560 CUDA cores, 5th Gen Tensor Cores
GPU Max Frequency 1.57 GHz
CPU 14-core Arm Neoverse-V3AE @ 2.6 GHz
Memory 128 GB 256-bit LPDDR5X (273 GB/s)
AI Performance 2,070 TFLOPS (FP4 Sparse)
Storage 1 TB NVMe M.2 (PCIe Gen5 x4)
Networking 1x 5GbE RJ45, 1x QSFP28 (4x 25GbE)
Display 1x HDMI 2.0b, 1x DisplayPort 1.4a
USB 2x USB-A 3.2 Gen2, 2x USB-C 3.1
Expansion M.2 Key M (PCIe Gen5 x4), M.2 Key E (PCIe Gen5 x1)
Camera HSB camera via QSFP, USB camera
Other I/O 2x CAN headers, PVA v3 Vision Accelerator
Power 40W – 130W (configurable)
Dimensions 243 x 112 x 57 mm

Ideal Use Cases

  • Humanoid robots and advanced autonomous machines requiring real-time AI reasoning
  • Autonomous vehicles and drones with multi-sensor fusion and path planning
  • Next-generation medical robotics and AI-powered surgical systems
  • Industrial automation with concurrent AI model execution via MIG
  • Edge AI inference running large generative models locally

Why Choose This Product?

The Jetson AGX Thor Developer Kit brings Blackwell-class AI to the edge. With 2,070 TFLOPS and 128GB of memory, it runs AI models that were previously impossible at the edge — including large vision-language models and complex multi-model pipelines. For organizations building the next generation of autonomous systems, this is the development platform that turns ambition into reality.

Interested? Contact us for pricing, early access programs, and integration support.

Reviews

There are no reviews yet.

Be the first to review “NVIDIA Jetson AGX Thor Developer Kit”

Your email address will not be published. Required fields are marked *