This page was machine-translated and may differ from the original. View original

Congatec Launches AMD Ryzen AI Embedded X100-Based COM-HPC Module

Google 우선 소스Published2026.07.27 11:00

Up to 16 Cores, 59 TOPS Inference Performance, Rugged Embedded Platform Targeting Physical AI and Robotics
Congatec, a company specializing in embedded and edge computing modules, has unveiled a COM-HPC client-size C module equipped with AMD Ryzen AI Embedded X100 series processors.

Congatec announced on the 27th that it has released a new COM-HPC client module, 'conga-HPC/cRX1'.

This product is provided based on the application-ready aReady.COM and is included in AMD's Kria AI System of Modulars (SOM) portfolio.

It features up to 16 AMD Zen 5 CPU cores and is configured with an AMD Radeon RDNA 3.5 integrated GPU with up to 40 compute units and a dedicated XDNA 2 NPU.

The integrated GPU provides up to 59 TOPS of performance based on INT8 AI inference and up to 29.7 TFLOPS of performance based on FP32 computation.

The dedicated NPU provides up to 50 TOPS of additional AI processing performance, enabling it to handle always-on workloads such as object detection, speech recognition, and image processing.

Large-scale language models (LLM) without a separate AI accelerator card The company explained that it can be run in a local environment.

Up to 128GB of memory is mounted via onboard soldering, and up to 512GB of NVMe storage is available as an option.

The Thermal Design Power (TDP) is 55W by default and can be set from 45W to 120W.

The operating temperature is supported from minus 40 degrees to plus 85 degrees based on industrial models.

It was developed in accordance with the IEC 62,443-4-1 standard and incorporates a reliable platform module (TPM 2.0).

Congatec announced that it also supports the response to the EU Cyber Resilience Regulation (CRA), which is scheduled to take effect in December 2027.

The operating systems support Windows 11, Windows 11 IoT Enterprise, and Linux, and CtrlX OS, Ubuntu Pro, and Kontron OS are pre-configured on a license basis.

It also includes the AMD ROCm open-source software stack to support the development and deployment of AI applications.

Florian Dritenthaler, Product Line Manager at Congatec, said, “OEMs can implement systems without separate AI accelerator cards and expect to reduce Total Cost of Ownership (TCO).”

Summit Shah, General Manager of Product Management and Marketing for AMD’s Adaptive and Embedded Computing Group, stated, “By integrating CPUs, GPUs, and NPUs into a single platform, we provide the deterministic performance and efficiency required for AI implementation in edge environments.”
본 기사에 대한 정정·반론·추후보도 청구는 보도 청구 안내를, 그간 게재된 보도문은 정정·반론보도 모아보기를 참고해 주세요.
명세환 기자
명세환 기자