AMD has expanded its Ryzen AI embedded P100 series processor lineup with new x86 chips aimed at industrial and edge AI systems such as factory automation, robotics, and medical imaging.
The new part joins the previously announced P100 series models and targets higher CPU core counts, more graphics compute, and better AI inference throughput within the same ball grid array package format.
what’s new
The new processor offers 8 to 12 “Zen 5” CPU cores, RDNA 3.5 graphics, and a neural processing unit based on the XDNA 2 architecture.
AMD listed up to 80 total system TOPS for AI inference. It is estimated to deliver up to 2x more CPU cores, up to 8x more GPU computing, and an estimated 36% more system TOPS compared to previous P100 series processors in the same package.
With support for an industrial temperature range of -40°C to 105°C, 24/7 continuous operation, and a 10-year life cycle, these processors are positioned for real-time, always-on deployments that require deterministic performance.
ROCm and software
A key addition is support for AMD ROCm, an open-source software stack for GPU computing and AI workloads. This gives embedded developers access to standard AI frameworks and open source compilers, runtimes, and libraries.
ROCm uses Heterogeneous Computing Interface for Portability (HIP), which AMD says separates GPU programming from specific hardware and reduces dependence on a single vendor’s software stack.
This processor combines CPU, GPU, and NPU functionality on a single chip, allowing workloads to be split between compute engines. AMD said this also supports predictable delays when the system runs mixed workloads.
virtualization stack
AMD also outlined a packaged virtualization reference stack for industrial mixed-critical applications. It is built on the Xen hypervisor and is designed to run Linux, Windows, Ubuntu, and RTOS environments in isolated domains.
AMD says its Zen 5 CPU cores provide the isolation and performance headroom to consolidate multiple workloads onto a single platform with deterministic multitasking behavior.
Target market
AMD highlighted three application categories for its expanded P100 series line: industrial PCs for machine vision, autonomous mobile robots, and 3D health imaging systems.
For industrial PCs, AMD says the processor can integrate programmable logic controller, machine vision, and human-machine interface functionality into a single system, and uses integrated graphics and NPUs for multi-camera vision processing and dashboard visualization.
For robotics, AMD talked about using the CPU for navigation, motion control, and route planning, while the GPU processes multiple camera feeds for spatial awareness and visual SLAM workloads. They also cited integrated memory between the CPU and GPU to reduce latency, positioning NPUs for always-on, low-power inference tasks such as object detection and scene understanding.
When it comes to medical image processing, AMD says the processor can perform 3D image processing for ultrasound and endoscopic systems at the edge and support clinical reasoning and Q&A workflows. The platform is positioned as a way for medical device manufacturers to integrate imaging, analysis, and reporting on embedded x86 systems with long product lifecycles.
In performance comparisons, AMD said the P100 series is expected to deliver up to 39% higher multi-threaded performance and up to 2.1x higher total system TOPS compared to the previous generation Ryzen Embedded 8000 series.
Partner system
Production solutions based on Ryzen AI Embedded P100 processors are available from original design manufacturer partners including Advantech, congatec, and Kontron.
“Advantech is proud to announce a comprehensive lineup powered by the scalable AMD Ryzen AI Embedded P100 processor portfolio. Featuring computer-on-modules, single-board computers, edge AI and intelligent systems, this portfolio leverages an enhanced unified AI architecture to enable highly efficient multitasking to drive next-generation edge AI advancements,” said Aaron Su, vice president of Embedded IoT Sector at Advantech.
Congatec pointed out configuration options spanning CPU core count and graphics scaling.
“With the launch of the AMD Ryzen AI Embedded P100 series, congatec expands its computer-on-module portfolio for embedded computing and edge applications with a versatile platform. By offering 4 to 12 CPU cores and scalable GPU performance, it allows customers to precisely match performance, power, and cost to specific application needs. From industrial automation to AI-accelerated systems, edge This exceptional level of flexibility is essential as workloads become more diverse,” said Florian Drittenthaler, Product Line Manager at congatec.
Kontron linked the chip to its mini-ITX product, which is based on the P100 part.
“The AMD Ryzen AI Embedded platform is a game-changer for industrial and AI-driven applications at the edge. Our P100-based K4131-Px mITX features 4- to 12-core APUs, allowing us to offer our customers a suite of solutions that deliver high computing performance and AI acceleration in the same compact footprint.” Thomas Stanik, Senior Sales and Business Development Manager, Kontron He said:
According to AMD, the Ryzen AI Embedded P100 series processors with 8 to 12 cores are currently shipping samples, with mass production shipments scheduled to begin in July 2026.
