NVIDIA Jetson Thor Module

Name: NVIDIA Jetson Thor Modul
Brand: NVIDIA
SKU: d3e064b8-9532-414f-899d-8bc2a1345639
Price: 4422 EUR
Availability: InStock

NVIDIA | A-000000-08175

€4,422.00 VAT incl. €3,685.00 VAT excl.

The NVIDIA Jetson T4000 and NVIDIA Jetson T5000 are embedded computing modules designed for advanced robotics, physical AI, embedded AI, and edge inference (Edge AI).

Quantity

Ask for a quote

Create a cart and turn it into a quote instantly!

Payment methods (see T&C)

NVIDIA Jetson T4000 / T5000 – Blackwell Embedded AI Modules for Robotics and Edge AI

Key features

NVIDIA Blackwell architecture optimized for advanced robotics, physical AI, and embedded inference
Two complementary performance levels to match the required thermal envelope and computing performance
Shared form factor and pin compatibility between Jetson T4000 and Jetson T5000 to simplify hardware development
256-bit LPDDR5X memory at 273 GB/s for intensive vision, multimodal, and generative AI workloads
Arm Neoverse-V3AE CPU with up to 14 cores depending on the version
NVMe storage support via PCIe Gen5
25 GbE networking for high-bandwidth robotic and multi-sensor architectures
Compatible with the NVIDIA software ecosystem: JetPack, TensorRT Edge-LLM, Isaac, Metropolis, Holoscan, Video Codec SDK

Applications

The NVIDIA Jetson T4000 / T5000 modules are intended for applications requiring high embedded computing power, low latency, and robust integration in Edge or robotic environments. They are particularly suitable for:

Autonomous mobile robotics
Humanoid robotics and physical AI applications
Industrial vision and real-time video analytics
Multi-sensor fusion and embedded perception
Embedded generative AI applications (LLM, VLM, VLA)
Edge AI systems for industry, healthcare, retail, and smart infrastructure
Robotic platforms with advanced needs in real-time reasoning, navigation, and interaction

Jetson T4000 and Jetson T5000 comparison

The Jetson Thor series is available here in two complementary modules. The Jetson T4000 is aimed at high-performance deployments with a more contained thermal envelope. The Jetson T5000 is intended for the most intensive applications in terms of compute, memory, and I/O bandwidth.

Module	GPU	CPU	AI Performance	Memory	Network	Power Consumption
Jetson T4000	NVIDIA Blackwell – 1,536 cores, 6 TPC	12-core Arm Neoverse-V3AE 64-bit	Up to 1,200 FP4 TFLOPS (sparse)	64 GB LPDDR5X 256-bit – 273 GB/s	3x 25GbE	40 – 70 W
Jetson T5000	NVIDIA Blackwell – 2,650 cores, 10 TPC	14-core Arm Neoverse-V3AE 64-bit	Up to 2,070 FP4 TFLOPS (sparse)	128 GB LPDDR5X 256-bit – 273 GB/s	4x 25GbE	40 – 130 W

An NVIDIA Jetson Thor platform designed for physical AI

With the Jetson T4000 and Jetson T5000, NVIDIA is targeting physical AI and robotics use cases where requirements go beyond conventional embedded vision. These modules are designed to locally run heavier workflows, including real-time reasoning, multimodal models, visual agents, perception + control pipelines, and advanced video processing.

The Jetson T4000 is notably highlighted by NVIDIA with JetPack 7.1, TensorRT Edge-LLM for efficient edge inference of LLMs/VLMs, and a unified Video Codec SDK for multimedia and real-time perception workflows. The Jetson T5000 follows the same platform logic with higher compute capability, memory, and network throughput for the most demanding systems.

Technical specifications of the NVIDIA Jetson T4000 / T5000 modules

Product family: NVIDIA Jetson Thor
GPU architecture: NVIDIA Blackwell
Vision accelerator: 1x PVA v3
Storage: NVMe storage support via PCIe; the T5000 also supports SSD via USB 3.2
PCI Express: up to 8 PCIe Gen5 lanes
USB: 3x USB 3.2 and 4x USB 2.0 on T5000
T5000 display: up to 4 shared HDMI 2.1 outputs and DisplayPort 1.4a MST
Mechanical form factor: 100 x 87 mm
Connector: 699-pin B2B connector
Cooling: Integrated Thermal Transfer Plate (TTP)

Jetson T4000 specifications

AI performance: up to 1,200 FP4 TFLOPS (sparse)
GPU: NVIDIA Blackwell GPU, 1,536 cores, Multi-Instance GPU, 6 TPC
CPU: 12-core Arm Neoverse-V3AE 64-bit
Max CPU frequency: up to 2.6 GHz
Memory: 64 GB LPDDR5X 256-bit
Memory bandwidth: 273 GB/s
Video encoding: 1x NVENC
Video decoding: 1x NVDEC
Video processing: real-time 4K encoding/decoding highlighted by NVIDIA
Network: 3x 25GbE
Other I/O: up to 5x I2S, 1x Audio Hub (AHUB), 2x DMIs, 4x UART, 3x SPI, 13x I2C, 6x PWM
Power consumption: 40 W to 70 W

Jetson T5000 specifications

AI performance: up to 2,070 FP4 TFLOPS (sparse)
GPU: NVIDIA Blackwell GPU, 2,650 cores, Multi-Instance GPU, 10 TPC
CPU: 14-core Arm Neoverse-V3AE 64-bit
CPU cache: 64 KB I-Cache + 64 KB D-Cache, 1 MB L2 per core, 16 MB shared L3
Max CPU frequency: up to 2.6 GHz
Memory: 128 GB LPDDR5X 256-bit
Memory bandwidth: 273 GB/s
Video encoding: 2x NVENC, up to 6x 4Kp60 (H.265)
Video decoding: 2x NVDEC, up to 4x 8Kp30 (H.265)
Cameras: up to 20 cameras via HSB, up to 6 cameras via 16 MIPI CSI-2 lanes, up to 32 cameras via virtual channels
Camera interfaces: C-PHY 2.1 (10.25 Gbps) and D-PHY 2.1 (40 Gbps)
Network: 4x 25GbE
Other I/O: 5x I2S, 2x AHUB, 2x DMIs, 4x UART, 4x CAN, 3x SPI, 13x I2C, 6x PWM
Power consumption: 40 W to 130 W