Name: Nvidia Developer Kit Jetson AGX Thor™ 128 GB 14 x 1.57 GHz
Brand: Nvidia
SKU: 3466340
Availability: InStock

A sleek black electronic device with ports on the side, accompanied by a circuit board with a central chip, tilted against the device.

A sleek, modern-looking black electronic device with vented top and multiple ports on the side, likely a gaming console or media player.

Exploded view of a computer component showing layers: outer casing, circuit board, chip, and internal parts. Displayed on a dark background.

1/3

EAN:

Part number:

Item no:

7 Piece(s)

Sales and shipping:

Conrad ElectronicTerms & Conditions

Total (€ 7,825.00 / piece)

€ 7,825.00

incl. VAT

Only 7 Piece(s) available

Only available for business customers

14-day return policy included

Report legal concerns

Technical specifications

AI power: 2,070 TFLOPS (FP4 - SPARE)
Module: Jetson T5000
GPU: Nvidia GPU with 2,560 computing units and 96 tenor computing units of the fifth generation / multi-instance GPU (MIG) with 10 TPCs
CPU: ARM® Neoverse® V3AE 14-core 64-bit CPU / 1 MB L2 cache per Kern 16 MB shared system L3 cache
128 GB, 256-bit LPDDR5X / 273 GB / s memory
Network: 1x 5 GbE RJ45 port / 1x QSFP28 (4x 25 GbE)
Connections: QSFP-connection / HDMI-connection / DisplayPort / 2 x USB-A 3.2 | 2x USB-C 3.1 / Gigabit-Ethernet / 2x 13-pin CAN-header / Microfit power connector
Data storage: 1 TB NVRAM M.2 Key M slot
Power 40 W – 130 W.

Product type: SBC

Processor cores (No.): 14 x

: 128 GB

Clock speed: 1.57 GHz

Memory capacity: 128 GB

Internal ports: M.2

Documents & Downloads

User & Safety Instructions

User & Safety Instructions 3466340 Nvidia Developer Kit Jetson AGX Thor™ 128 GB 14 x 1.57 GHz

Description

The ultimate platform for physical AI and humanoid robotics.

A compact power pack for advanced AI and robotics
The Nvidia® JETSON Thor™ Series modules provide the ultimate platform for physical AI and robotics. They offer AI computing power of up to 2,070 FP4 TFLOPS and 128 GB RAM with configurable power consumption between 40 W and 130 W. They offer more than 7.5 times the AI computing power of Nvidia AGX ORIN™ and 3.5 times better energy efficiency.

Supercomputer for humanoid use
Speed up generative AI and large transformation models on the Edge with the 2,070 FP4 TFLOPS Blackwell graphics card.

High speed sensor processing
Capture high-speed sensor data for real-time performance with 4x 25 GbE network ports, a camera relief module, and a holoscan sensor bridge.

KI-Software for Robotics
Discover a solution designed for humanoid robotics and physical AI applications based on the Nvidia Isaac™ platform and GR00T base models.

Robust security
Provide end-to-end security for the entire computing platform, all AI models, and the entire edge-to-cloud pipeline.

This text is machine translated.