Specifications
Module
NVIDIA IGX Thor T5000 embedded
AI Performance
Up to 2070 FP4 TFLOPS
Networking
4 x 10G SFP plus 1 x 5G Ethernet
Functional Safety
Onboard Functional Safety MCU
Operating Temperature
-30 C to +60 C with ruggedised design
Sensor Interface
MIPI interface for direct camera connection or Advantech Holoscan Sensor Bridge
Sensor Bridge
NVIDIA Holoscan Sensor Bridge ready, high-bandwidth synchronised sensor communication
Safety Framework
Aligned with the NVIDIA Halos AI Lab framework
Software Stack
NVIDIA AI Enterprise-IGX with NIM microservices
Robotics Stack
NVIDIA Isaac foundation models, libraries and frameworks
Vision Stack
NVIDIA Metropolis for vision AI
Sensor Processing
NVIDIA Holoscan platform integration
Lifecycle
10-year product lifecycle support
Related SKU
MIB-735 AI inference board kit, also powered by NVIDIA IGX Thor
Platform Role
Positioned as the neural hub connecting cameras, sensors and AI models in robotics
Announced
October 29 2025 launch announcement with NVIDIA ecosystem partners
Overview
The Advantech MIC-735-IT is a deployment-ready AI inference system built on the NVIDIA IGX Thor T5000 platform module. It targets safe physical AI - robotics, factory automation and medical systems - where functional safety and deterministic sensor fusion are prerequisites rather than options.
The system embeds an onboard Functional Safety MCU that integrates both onboard and external sensor data, letting external safety signals be processed reliably inside the system architecture for real-time risk mitigation. A MIPI interface enables direct camera connectivity or integration with Advantech’s Holoscan Sensor Bridge for high-throughput, synchronised communication between multiple imaging sensors and the IGX Thor module.
Networking covers four 10G SFP ports plus 5G Ethernet, and the ruggedised design operates from -30 C to +60 C. Software support spans the NVIDIA AI Enterprise-IGX stack with NIM microservices, NVIDIA Isaac for robotics, NVIDIA Metropolis for vision AI and the NVIDIA Holoscan platform for sensor processing, with a 10-year lifecycle commitment. Alignment with the NVIDIA Halos AI Lab framework positions the MIC-735-IT for safety-focused physical AI workloads.
Key Benefits
Up to 2070 FP4 TFLOPS of Blackwell-class inference in a rugged chassis. Onboard Functional Safety MCU fuses onboard and external safety signals for real-time risk mitigation. Holoscan Sensor Bridge support synchronises multiple high-bandwidth imaging sensors. -30 C to +60 C rugged design suits factory and outdoor robotics. Full NVIDIA AI Enterprise-IGX, Isaac, Metropolis and Holoscan stack removes software bring-up work, and a 10-year lifecycle matches industrial and medical programme durations.
Applications
Autonomous mobile robots and AMRs, humanoid and collaborative robotics, factory automation with functional safety, surgical visualisation and medical imaging systems, multi-sensor fusion platforms, and physical AI systems requiring Halos-aligned safety architectures.
Request a Quote — ADVANTECH MIC-735-IT — AI INFERENCE SYSTEM POWERED BY NVIDIA IGX THOR
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →