AMD targets industrial edge AI with new Ryzen embedded chips built for real-time inference
AMD has launched its Ryzen AI Embedded P100 Series processors, delivering scalable and efficient AI compute (AI-on-Chip) for industrial and edge AI deployments.
These next-gen processors span up to 2x CPU cores, 8x GPU compute, and a whopping 36% better system tera operations over its predecessors.
A key application is industrial automation, followed by mobile robotics and medical imaging, adding support for real-time AI processing and dependability for always-on environments.
The processors combine “Zen 5” CPU cores, RDNA 3.5 graphics, and XDNA 2-based neural processing units (NPUs) for low-latency, power-efficient AI inference.
“Advantech is proud to announce a comprehensive lineup powered by the scalable AMD Ryzen AI Embedded P100 processor portfolio,” says Aaron Su, vice president, Embedded IoT Sector, Advantech. “Featuring Computer-on-Modules, Single Board Computers, and Edge AI and Intelligent Systems, this portfolio leverages an enhanced integrated AI architecture to deliver high-efficiency multitasking that drives next-gen Edge AI advancement.”
Use cases extend to intelligent machine vision, autonomous robotics and 3D health imaging, enabling advanced AI models such as YOLOv12, U-Net and Llama 3.2-Vision.
The AMD ROCm open-source software stack provides new possibilities for developers to run state-of-the-art AI frameworks with advantages over vendor lock-in.
Moreover, the architecture enables efficient workload partitioning and virtualization as well as mixed-criticality applications, which can streamline development for next-gen embedded systems.
The P100 Series is gaining support from industry partners such as Advantech, congatec, and Kontron, where production shipments of the final boards are projected to start by mid-2026.
Why edge architectures fail (and how to design around it)
Article Topics
AI inference | AMD | edge AI | edge computing | embedded processors | Industrial AI
Comments