Advantech launches AIR-411 edge AI HPC system

Advantech has introduced the AIR-411, a compact edge artificial intelligence (AI) high-performance computing (HPC) system engineered for compute-intensive industrial applications. Targeting semiconductor inspection, medical imaging workstations, multi-channel video analytics, and on-premises large language models (LLMs), the platform combines advanced processing capabilities with a 10-year long-term support lifecycle. Powered by the latest Intel Core 200S (Bartlett Lake-S) processors featuring up to 24 cores and 32 threads, the system delivers up to 36 MB of L3 cache to manage demanding virtualization and multi-container workloads directly at the industrial edge.

Scalable GPU acceleration in a compact chassis
Despite its compact 17-liter chassis, the AIR-411 incorporates a PCIe Gen4 x16 expansion slot capable of supporting a full-height, full-length graphics processing unit (GPU) drawing up to 300W. This flexibility accommodates a wide array of AI accelerators, including the NVIDIA RTX PRO 6000 Blackwell Max-Q, Intel Arc Pro B60, AMD Radeon PRO W7900/W7800, and Qualcomm Cloud AI 100 Ultra. To ensure reliable continuous operation in space-constrained deployments, the system features an intelligent fan control mechanism that optimizes thermal management while maintaining ultra-quiet operation below 38 dB. Extensive connectivity options—including dual Ethernet, triple-display support, M.2 NVMe storage, and optional GPIO interfaces—simplify integration with complex industrial peripherals and high-resolution camera sensors.

Streamlined model deployment with WEDA
To accelerate application development, the AIR-411 supports operating systems such as Windows 11 IoT Enterprise 2024 LTSC and Ubuntu Desktop 24.04 LTS, alongside the Advantech Edge AI SDK. When integrated with Advantech’s WEDA platform, the system enables centralized data and model management across distributed edge networks. This software ecosystem allows IT teams to centrally deploy validated models, seamlessly orchestrate AI workloads, and push software updates to multiple edge nodes simultaneously. For specialized hardware requirements, Advantech’s European Design and Manufacturing Services (DMS) team offers comprehensive local hardware customization.

Additional Context: Edge LLMs and thermal management
This section details technical specifications and competitive benchmarking not included in the original product announcement.

Deploying Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) workloads directly at the edge marks a significant architectural shift for industrial and healthcare facilities. Traditionally, these generative AI models relied on remote cloud data centers, exposing sensitive information—such as proprietary semiconductor schematics or patient medical scans—to data sovereignty and cybersecurity risks during transit. By leveraging high-core-count processors like the Intel Bartlett Lake-S alongside high-VRAM 300W workstation GPUs natively on-premises, platforms like the AIR-411 allow facilities to process complex AI inference completely offline with near-zero latency. Furthermore, securely dissipating 300 watts of GPU thermal output from a restrictive 17-liter enclosure—a fraction of the internal volume of a standard 45-liter ATX mid-tower desktop—requires highly optimized acoustic and fluid dynamic engineering to prevent catastrophic thermal throttling during sustained, 24/7 automated optical inspection tasks.

Edited by an industrial journalist, Lekshman Ramdas, with AI assistance.

www.advantech.com