Advantech MIC-735 Enables Safety Physical AI

As Physical AI expands into humanoid robotics, AMR, autonomous machines, and intelligent systems operating in real-world environments, the need for functional safety, deterministic operation, and international standards compliance continues to grow.   

NVIDIA Halos AI System Inspection Lab Certification program provides a structured framework for evaluating system-level safety design and supports alignment with major international functional safety standards. Through this validation process, MIC-735 establishes alignment with key safety frameworks, including, IEC 61508, and ISO 13849, helping customers better address compliance requirements across industrial and robotics applications. This approach enables customers to streamline safety-related development and reduce the complexity and effort required for downstream certification processes.

Enable future access to safety-critical AI markets through ecosystem alignment

Functional safety is a critical foundation for deploying Physical AI systems in robotics and industrial environments, where machines must reliably operate and interact within unpredictable real-world conditions. Through its alignment with NVIDIA Halos AI Lab framework, MIC-735 is powered by NVIDIA IGX platform with functional safety capability to expand NVIDIA’s safety-focused ecosystem for physical AI workloads, making it a leading choice for robotics application. 

MIC-735 is based on NVIDIA IGX Thor T5000 platform architecture and integrates both onboard and external sensor data to strengthen system-level safety capabilities. This design enables a coordinated safety and control framework for physical AI systems, allowing external safety signals to be reliably processed within the system architecture to support real-time risk mitigation in robotics and industrial automation applications.

Deployment-Ready Safe Physical AI Platform

Beyond hardware design, the MIC-735 supports a 10-year lifecycle and NVIDIA AI Enterprise-IGX software stack, enabling customers to accelerate applications using NVIDIA NIM microservices, NVIDIA Isaac open robotics development platform’s foundation models, libraries, and frameworks, NVIDIA Metropolis for vision AI, and integrates the NVIDIA Holoscan platform, including NVIDIA Holoscan Sensor Bridge technologies for real-time sensing, enabling high-bandwidth and synchronized communication between multiple sensors and AI processors. 

It is specifically designed for environments operating in complex and dynamic conditions that require real-time safety mechanisms, including certified emergency stop (E-Stop) capabilities over secure IP-based networks. Through ecosystem components such as the Endpoint Controller, fleet-wide commands can be executed, while the Wireless E-Stop Pro support localized control and emergency intervention for individual AMRs when required. The Advantech MIC-735 will soon achieve Halos lab validation, establishing it as a FuSa-ready platform.

Advantech to Showcase at COMPUTEX 2026

Advantech invites global partners, developers, and innovators to explore its next-generation edge AI platforms at COMPUTEX 2026, taking place June 2–5 in Taipei, Taiwan. To learn more about Advantech’s Physical AI solutions, visit booth #K0603a to experience safe industrial intelligence in real-world applications. For more information, please refer to the official event page.