SuperAPI AI Directory

Jetson Thor - NVIDIA's Robotics AI Computing Platform

Author: AI小集 · Original source: AI工具集

What is Jetson Thor

Jetson Thor is a high-performance robotics and edge AI computing platform launched by NVIDIA, based on the latest Blackwell GPU architecture, with AI computing power up to 2070 TFLOPS, 7.5 times that of the previous generation, and equipped with 128GB of memory, supporting configurable power consumption of 40-130W. The platform is designed specifically for physical AI and robotics development, supporting multi-sensor processing, generative AI models, and real-time inference tasks. Jetson Thor integrates powerful software stacks such as NVIDIA Isaac, which can accelerate the development and deployment of applications such as humanoid robots, spatial intelligence, and video analytics, and is one of the key technologies driving the arrival of the era of general-purpose robotics.

 Jetson Thor

Main Features of Jetson Thor

  • Powerful AI computing capability: Jetson Thor is based on the Blackwell architecture, providing up to 2070 TFLOPS of AI computing power and supporting multi-instance GPU technology to meet the needs of complex AI tasks.
  • Multi-sensor processing capability: Supports 4×25GbE networking and Holoscan Sensor Bridge, enabling low-latency multi-sensor data fusion and processing.
  • Real-time inference and control: Has low-latency real-time inference capability, supports running multiple AI models on edge devices, and is suitable for high-precision real-time control scenarios such as humanoid robots.
  • Software support and ecosystem: Native support for software tools such as NVIDIA Isaac, GR00T, and Metropolis, compatible with a variety of generative AI frameworks, enabling the construction of a rich ecosystem.
  • Flexible power consumption and thermal design: Power consumption is configurable from 40-130W, and the developer kit integrates thermal design, adapting to different application scenarios.
  • Rich interfaces and expandability: Provides interfaces such as QSFP28, RJ45, USB 3.x, and PCIe Gen5, supports multiple form factors, and meets diverse expansion needs.
  • Security and reliability: Provides end-to-end security, ensuring that data and systems run securely and reliably throughout the entire process from edge to cloud.

Jetson Thor Specifications

Jetson AGX Thor Developer Kit Jetson T5000 Jetson T4000
AI Performance 2070 TFLOPS (FP4 — Sparse) 1200 TFLOPS (FP4 — Sparse)
GPU 2560-core NVIDIA Blackwell architecture GPU with 96 fifth-generation Tensor Cores 1536-core NVIDIA Blackwell architecture GPU with 64 fifth-generation Tensor Cores
Multi-Instance GPU (MIG) with 10 TPCs Multi-Instance GPU (MIG) with 6 TPCs
GPU Max Frequency 1.57 GHz
CPU 14-core Arm® Neoverse®-V3AE 64-bit CPU 12-core Arm® Neoverse®-V3AE 64-bit CPU
1 MB L2 cache per core 1 MB L2 cache per core
16 MB shared system L3 cache 16 MB shared system L3 cache
CPU Max Frequency 2.6 GHz
Vision Accelerator 1 PVA v3
Memory 128 GB 256-bit LPDDR5X 64 GB 256-bit LPDDR5X
273 GB/s 273 GB/s
Storage 1 TB NVMe M.2 Key M slot NVMe support via PCIe
SSD support via USB3.2
Video Encode 6x 4Kp60 (H.265)
12x 4Kp30 (H.265)
24x 1080p60 (H.265)
50x 1080p30 (H.265)
48x 1080p30 (H.264)
6x 4Kp60 (H.264)
Video Decode 4x 8Kp30 (H.265)
10x 4Kp60 (H.265)
22x 4Kp30 (H.265)
46x 1080p60 (H.265)
92x 1080p30 (H.265)
82x 1080p30 (H.264)
4x 4Kp60 (H.264)
Camera HSB camera via QSFP slot Up to 20 HSB cameras
USB camera Up to 6 cameras using 16x lane MIPI CSI-2
Up to 32 cameras using virtual channels
C-PHY 2.1 (10.25 Gbps)
D-PHY 2.1 (40 Gbps)
PCIe M.2 Key M slot supporting x4 PCIe 5.0 Up to Gen5 (x8 lanes)
M.2 Key E slot supporting x1 PCIe 5.0 Root port only — C1 (x1) and C3 (x2)
Root point or endpoint — C2 (x1), C4 (x8) and C5 (x4)
USB interface 2 USB-A (3.2 Gen2) xHCI host controller with integrated PHY (up to)
2 USB-C (3.1) 3 USB 3.2 interfaces
4 USB 2.0 interfaces
Network 1 5GBe RJ45 port 4 25 GbE 3 25 Gbe
1 QSFP28 (4 25 GbE)
Display 1 HDMI 2.0b 4 shared HDMI2.1
1 DisplayPort 1.4a VESA DisplayPort 1.4a—HBR2、MST
Other I/O QSFP connector 5 I2S / 2 Audio Hubs (AHUB), 2 DMIS, 4 UART, 4 CAN, 3 SPI, 13 I2C, 6 PWM outputs 5 I2S / 2 Audio Hubs (AHUB), 2 DMIS, 4 UART, 3 SPI, 13 I2C, 6 PWM outputs
M.2 Key E expansion slot (WLAN/BT, x1 PCIe, USB2.0, UART, I2C, I2S)
M.2 Key M connector (NVMe for storage)
PCIe x 4 lanes, I2C, PCIe x 2 lanes
2 13-pin CAN headers
2 6-pin automation headers
LED
JTAG connector (2 x 5 pin header)
1 fan interface – 12V, PWM and Tach
Audio panel header (2 5-pin)
Microfit power socket
RTC backup battery connector 2-pin
Power 40 W – 130 W 40 W – 130 W 40 W – 75 W
Form factor 243.19 mm x 112.40 mm x 56.88 mm 100 mm x 87 mm
Thermal Transfer Plate (TTP) and optional fan or heatsink 699-pin B2B connector
– Integrated Thermal Transfer Plate (TTP) with heat pipe

Jetson Thor price configuration

  • Jetson AGX Thor Developer Kit: priced at $3,499 (approximately 25,000 RMB), now officially on sale.
  • Jetson T5000 module: unit price of $2,999 for purchases of 1,000 units or more, now officially on sale.

Jetson Thor purchase address

  • Jetson Thor official website purchase address:https://developer.nvidia.com/buy-jetson

Jetson Thor application scenarios

  • Humanoid robots: Jetson Thor provides powerful AI computing power and real-time control capabilities, supporting humanoid robots in executing complex tasks and natural language interaction.
  • Spatial intelligence: through efficient visual AI processing capabilities, suitable for real-time video surveillance and analysis, enhancing the performance of intelligent monitoring systems.
  • Medical robots: High-precision real-time image processing and AI inference, suitable for surgical assistance and rehabilitation robots, improving healthcare efficiency and quality.
  • Industrial Automation: Jetson Thor supports complex motion control and multi-sensor data processing, suitable for industrial robotic arms and quality inspection systems, enhancing production efficiency and precision.
  • Logistics and Delivery: Real-time navigation and path planning capabilities, suitable for autonomous mobile robots in logistics warehouses, improving logistics efficiency and accuracy.