Chipmaker Advanced Micro Devices (AMD) has expanded its Ryzen AI Embedded P100 Series processors to support demanding industrial and edge artificial intelligence (AI) workloads such as factory automation, autonomous robots, and medical imaging systems.
The expanded lineup is designed for environments that require real-time AI processing, predictable performance, and long-term reliability for systems that operate around the clock.
According to AMD, the new processors deliver up to twice the CPU core count, eight times more graphics processing unit (GPU) compute, and an estimated 36% higher system tera operations per second (TOPS) compared with earlier P100 Series chips that use the same compact ball grid array (BGA) package.
The processors feature eight to 12 “Zen 5” CPU cores, support up to 80 system TOPS, and combine RDNA 3.5 graphics with a neural processing unit (NPU) built on the XDNA 2 architecture. This architecture enables AI inference with lower latency and improved power efficiency.
AMD said the chips support industrial temperature ranges from -40°C to 105°C, continuous 24/7 operation, and 10-year product life cycles, making them suitable for industrial deployments that require long-term stability.
In factories, the processors can consolidate programmable logic controllers (PLCs), machine vision systems, and human-machine interface (HMI) dashboards into a single industrial PC. The integrated GPU and NPU can accelerate multi-camera vision processing and anomaly detection using AI models such as DeepSORT, RAFT-Stereo, CenterPoint, and PaDiM.
For mobile robots and autonomous machines, the CPUs handle navigation and motion control, while the GPU processes camera feeds for spatial awareness and visual simultaneous localization and mapping (SLAM). The NPU provides always-on AI inference for tasks such as object detection using models including YOLOv12 and MobileSAM.
In healthcare environments, the processors support 3D imaging and clinical AI for ultrasound systems, endoscopes, tissue classification, and tumor detection. AMD said healthcare original equipment manufacturers (OEMs) can run imaging, AI analysis, and reporting on a single embedded platform.
The processors also support the ROCm open software ecosystem, allowing developers to run standard AI frameworks without rewriting code while avoiding vendor lock-in. AMD said the architecture enables efficient workload distribution across CPU, GPU, and NPU components to improve performance for mixed AI and industrial workloads.
Compared with the previous Ryzen Embedded 8000 Series, AMD estimates the new P100 Series processors deliver up to 39% higher multithreaded performance and 2.1 times higher total system AI compute performance.