Rutronik Expands Portfolio with Edge AI COM Module

Edge computing systems in industrial automation, medical technology and transportation platforms increasingly require integrated AI acceleration within standardized module formats. Rutronik has added the ADLINK cExpress-R8 to its range, a COM Express^® R3.1 Type 6 Compact module developed for AI-driven edge deployments.

COM Express R3.1 platform for AI inference
The ADLINK cExpress-R8 is implemented in the COM Express^® R3.1 Type 6 Compact form factor (95 × 95 mm), enabling integration via carrier boards in embedded system designs. The R3.1 specification defines mechanical dimensions, pinout and high-speed interfaces, supporting scalable edge computing architectures.

The module is based on AMD Ryzen™ Embedded 8000 Series processors with up to eight Zen 4 CPU cores. It integrates RDNA™ 3 GPU architecture and an XDNA™ NPU for AI acceleration, delivering up to 40 TOPS of combined CPU, GPU and NPU inference performance. This heterogeneous architecture enables local execution of vision models, object recognition and analytics workloads in latency-sensitive environments, reducing dependence on centralized processing within a digital supply chain of distributed industrial systems.

High-bandwidth memory and industrial connectivity
For data-intensive workloads, the cExpress-R8 supports up to 96 GB DDR5-5600 memory, with ECC and non-ECC options. ECC memory is particularly relevant in medical imaging and industrial control systems where error detection and correction contribute to operational stability.

The module provides 16 PCIe Gen4 lanes for high-speed expansion, four USB 3.2 Gen2 ports, four USB 2.0 ports and four SATA interfaces at 6 Gb/s. Networking is handled via 2.5 GbE with optional TSN support, enabling deterministic Ethernet communication in automation and robotics environments. A TPM 2.0 module can be implemented to establish hardware-based security functions.

Support for up to four independent displays via DisplayPort, HDMI, eDP/LVDS or VGA allows integration into multi-screen operator panels, diagnostic systems or edge-based video analytics platforms.

Application scope and lifecycle planning
The module is intended for industrial automation and robotics, medical imaging and diagnostics, transportation and mobility systems, smart surveillance and video analytics, as well as retail POS/POI systems and smart infrastructure nodes. In these environments, edge AI processing can lower network loads and improve response times compared with cloud-dependent architectures.

ADLINK specifies a 10-year product lifecycle for the cExpress-R8, supporting long-term system design, certification and maintenance planning in embedded applications.

www.rutronik24.com