Advantech Announces Functional Safety-Compliant Physical AI Platform 'MIC-735'

Advantech Co., Ltd. (Headquarters: Taitung District, Taipei, Taiwan; hereinafter "Advantech") announces the launch of its latest edge AI platform, the 'MIC-AI Series,' powered by NVIDIA® Jetson Thor™. Leveraging OpenClaw and NVIDIA Nemotron™, it enables agent-based AI at the edge.

▶ Advantech Co., Ltd – Press Release

Press Release: English Version

*Below is the translated version of this press release.

Advantech Announces Functional Safety-Compliant Physical AI Platform 'MIC-735'

On June 22, 2026, Advantech Co., Ltd. (Taiwan, TWSE: 2395; hereinafter "Advantech") announced significant progress toward certification by the 'NVIDIA Halos Validation Lab' for its next-generation edge AI platform 'MIC-735,' powered by NVIDIA IGX Thor™. It is positioned as one of the first platforms to support NVIDIA's safety-focused Physical AI framework.

Accelerating Physical AI Adoption in Robotics and Automation

The application of 'Physical AI' is rapidly expanding to humanoid robots, AMRs (Autonomous Mobile Robots), autonomous machinery, and various intelligent systems operating in real-world environments. Consequently, functional safety, deterministic operation for reliable, low-latency control, and compliance with international standards have become essential requirements to enable safe human collaboration and dependable on-site operations.

The NVIDIA Halos AI System Inspection Lab certification program provides a sophisticated framework for systematically evaluating and validating the safety design of entire systems. Utilizing this program streamlines compliance with major international functional safety standards and supports the realization of reliable, secure designs. Through this validation, MIC-735 has established alignment with key global safety standards, including 'IEC 61508' and 'ISO 13849,' significantly enhancing its ability to meet various regulations in the industrial and robotics sectors. Furthermore, adopting this approach simplifies safety-related design and development processes, reducing the engineering effort required for final certification.

Accessing the Next-Generation Safety-Centric AI Market Through Strong Ecosystem Collaboration

Functional safety is a critical foundation for the practical implementation of Physical AI systems in robotics and industrial fields, where machines must operate reliably and interact safely in unpredictable real-world conditions. MIC-735 adopts the NVIDIA IGX™ platform with functional safety capabilities through collaboration with the NVIDIA Halos AI Lab framework. This contributes to the expansion of a safety-focused Physical AI ecosystem and provides an optimal solution for developing next-generation robotic applications.

Based on the NVIDIA IGX Thor™ T5000 architecture, MIC-735 is a high-performance platform that integrates internal and external sensor data to enable real-time processing. This achieves high-precision environmental awareness and robust safety functions across the entire system. Furthermore, this design provides a Physical AI framework that integrates safety and control, enabling immediate risk reduction in robotics and industrial automation by reliably processing external safety signals.

An Immediately Deployable, Safety-Compliant Physical AI Platform

MIC-735 goes beyond hardware design, supporting a 10-year long-term lifecycle and the NVIDIA AI Enterprise-IGX software stack. This enables rapid development of AI applications using NVIDIA NIM microservices, NVIDIA Isaac's foundation models and various development frameworks, and NVIDIA Metropolis. Furthermore, by integrating the NVIDIA Holoscan platform (including Holoscan Sensor Bridge technology), it achieves real-time sensing and high-bandwidth, high-precision synchronized communication between multiple sensors and AI processors.

Designed for operation in dynamic on-site environments, the product supports real-time safety control, including certified emergency stop (E-Stop) functionality over secure IP networks. Ecosystem features such as the Endpoint Controller enable integrated fleet-wide management and immediate command execution. Moreover, by utilizing 'Wireless E-Stop Pro,' direct local control or emergency stops can be performed on individual AMRs (Autonomous Mobile Robots) during emergencies, enabling rapid on-site risk response.

Advantech's MIC-735 is scheduled to complete certification at the NVIDIA Halos Validation Lab in the near future, establishing its reliability and positioning as a functional safety-compliant (FuSa-ready) platform.

The Industry's First NVIDIA IGX Thor-Powered, Production-Ready NVIDIA-Certified System

NVIDIA-Certified Systems symbolize the 'gold standard'—not just as a safety benchmark, but for superior platform performance, security, scalability, and seamless software integration. Advantech's 'MIC-735' has passed all these rigorous evaluations and has been officially certified by NVIDIA as an immediately deployable AI platform. This allows adopting companies to significantly shorten system development timelines and minimize technical barriers during development and integration. This achievement underscores Advantech's leadership in the Physical AI domain and provides a reliable, high-performance architecture for next-generation robotics, vision AI, and cutting-edge medical applications.

Exhibiting at COMPUTEX 2026

Advantech exhibited at COMPUTEX 2026, held in Taipei, Taiwan, from June 2 to 5, 2026, introducing its next-generation edge AI platform at Booth #K0603a. For more details, please visit Advantech's official website or contact our sales representatives.

*Company names, service names, and product names mentioned in this release are registered trademarks or trademarks of their respective companies.

*Specifications of this product are subject to change without prior notice.