NVIDIA IGX Thor Developer Kit Mini
The NVIDIA IGX Thor Developer Kit Mini is a compact, high-performance AI development platform designed for industrial robotics, medical systems, and physical AI applications. Built on the IGX T5000 module with NVIDIA Blackwell GPU architecture, it delivers up to 2070 FP4 TFLOPS of AI compute performance within a 130W power envelope, making it ideal for running the latest generative AI models at the edge with enterprise-grade reliability and functional safety.
Why Choose the IGX Thor Developer Kit Mini
-
Blackwell Architecture GPU: 2560-core NVIDIA Blackwell iGPU with fifth-generation Tensor Cores for exceptional AI inference
-
Powerful 14-Core CPU: Arm Neoverse V3AE 64-bit processor with Multi-Instance GPU (MIG) technology for low-latency, real-time applications
-
128GB Memory: 256-bit LPDDR5X memory with 273GB/s bandwidth for demanding AI workloads
-
Functional Safety: Integrated Functional Safety Island (FSI) and onboard safety microcontroller (MCU) for industrial and medical certification support
-
Accelerator Suite: Third-generation Programmable Vision Accelerator (PVA), dual encoders/decoders, optical flow accelerator, and more
-
10-Year Enterprise Support: NVIDIA AI Enterprise software with long-term lifecycle support for mission-critical deployments
High-Speed Connectivity for Sensor Fusion
The developer kit provides extensive I/O options for high-speed sensor fusion and edge applications. Features include a QSFP28+ connector supporting 4×25GbE, Gigabit Ethernet RJ45, HDMI and DisplayPort outputs, dual USB-A 3.2 ports, dual USB-C 3.1 ports with Power Delivery support up to 140W, and additional connectivity interfaces including CAN headers for industrial communication.
Enterprise AI Software Stack
Run the complete NVIDIA AI Enterprise software stack at the edge, including NVIDIA Isaac for robotics, Metropolis for visual AI agents, Holoscan for sensor processing, and NVIDIA Halos safety AI agents. The platform supports generative AI models from Vision-Language-Action (VLA) models like NVIDIA Isaac GR00T N to popular LLMs and VLMs such as NVIDIA Cosmos Reason, with seamless cloud-to-edge deployment via NVIDIA NIM.
Industrial & Medical Applications
Purpose-built for industrial and medical environments, the IGX Thor Developer Kit Mini delivers enterprise-grade performance, security, and functional safety with a 10-year lifecycle. Ideal for autonomous robots, physical AI agents, smart hospitals, manufacturing automation, and edge AI applications requiring real-time sensor processing and AI inference with functional safety certification pathways including ISO 26262, IEC 61508, IEC 60601, and IEC 62304.
What's in the Box
- NVIDIA IGX T5000 module with heat sink and reference carrier board
- 140W DC power supply with USB-C connector
- 802.11ax wireless network interface controller (Wi-Fi 6)
- 1TB NVMe storage populated in M.2 Key-M slot
- Safety microcontroller (MCU)
- Quick Start Guide
Full Specifications
Processor & Graphics
| GPU |
2560-core NVIDIA Blackwell architecture iGPU with fifth-gen Tensor Cores |
| CPU |
14-core Arm Neoverse V3AE 64-bit processor |
| AI Performance |
Up to 2070 FP4 TFLOPS |
| Memory |
128GB 256-bit LPDDR5X, 273GB/s bandwidth |
Storage & Expansion
| Primary Storage |
1TB NVMe (M.2 Key-M slot) |
| Secondary Storage |
UFS (Universal Flash Storage) |
I/O Interfaces
| High-Speed Networking |
QSFP28+ connector (4×25 Gbps per port) |
| Ethernet |
RJ45 Gigabit Ethernet (5 Gbps) |
| USB Type-A |
2× USB 3.2 Gen 2 (10 Gbps) |
| USB Type-C |
2× USB 3.1 (5 Gbps) with PD Sink (140W) |
| Video Output |
HDMI, DisplayPort |
| Industrial I/O |
2× 13-pin CAN headers |
| Power Input |
Microfit power jack (9-28V DC, up to 8A) |
| Debug Port |
USB-C (behind lid cover) |
Wireless Connectivity
| Wi-Fi |
802.11ax (Wi-Fi 6) - Realtek module |
| Bluetooth |
Included |
Power & Environment
| Power Consumption |
130W typical |
| Power Supply |
140W USB-C DC adapter (included) |
| Power Input Range |
9-28V DC (up to 8A via Microfit jack) |
Safety & Reliability
| Functional Safety |
Functional Safety Island (FSI) integrated into Thor SoC |
| Safety MCU |
Onboard safety microcontroller |
| Lifecycle Support |
10-year NVIDIA AI Enterprise support |
| Certification Support |
ISO 26262, IEC 61508, IEC 60601, IEC 62304 |
Software
| Operating System |
IGX Software (Linux-based) |
| AI Frameworks |
NVIDIA AI Enterprise, NVIDIA NIM |
| Applications |
NVIDIA Isaac (robotics), Metropolis (visual AI), Holoscan (sensor processing), Halos (safety AI agents) |
Specifications are sourced from the official manufacturer product page and may be updated.