Ryzen processors power physical AI workloads
Optimized for physical AI, AMD’s Ryzen AI Embedded X100 series processors combine CPU, GPU, and NPU resources on a single embedded SoC. With up to 16 AMD Zen 5 CPU cores, an integrated GPU, power-efficient NPU, and unified memory, the processors handle perception, reasoning, and real-time control workloads in robotics, industrial automation, aerospace and defense, and other embedded systems.
AMD says the series delivers up to 2.1× higher multithreaded CPU performance, 1.7× higher graphics performance, and 3.5× higher AI token generation with 1.4× faster time-to-first-token than Intel Core Ultra Series 3 processors. The company also says the processors provide up to 3× higher peak FP32 performance than the NVIDIA Jetson T5000 and an average 1.7× faster beamforming for cardiac ultrasound than the NVIDIA RTX 4000 Ada.
The Ryzen AI Embedded X100 series supports an open software stack with Linux, the AMD ROCm GPU software stack, the Xen Hypervisor, and AI frameworks including PyTorch, ONNX, and TensorFlow. AMD also provides tools to migrate CUDA codebases to ROCm.
Production availability of the Ryzen AI Embedded X100 series is expected in Q4 2026. Learn more about the processors on the product page or in AMD’s technical blog.
The post Ryzen processors power physical AI workloads appeared first on EDN.
