Edge AI HPCs for Vision AI and GenAI

The HPCs enable on-premise deployment and fine-tuning of AI models, including large language models (LLMs) and AI vision inference applications. Powered by AMD Ryzen 8000 series embedded processors, the 'AIR-410' supports high-end GPUs, including the 'Nvidia RTX PRO 6000 Blackwell Max-Q Workstation Edition', AMD Radeon PRO W7900 and Qualcomm Cloud AI 100 Ultra, which provide different AI capabilities for various edge AI applications. Thanks to its compact three-slot FHFL GPU form factor and quiet operation (< 38.2 dBA), it is particularly suitable for on-premise semiconductor inspection, medical imaging and LLM inference. Key features include an 850 W power supply, support for up to 96 Gbytes of DDR5 RAM and 14 Tbytes of storage, and extensive I/O interfaces. Equipped with Advantech's 'Edge AI SDK', pre-validated AI models and a programming-free GUI, the HPC accelerates AI deployment at the Edge.

The scalable and AI-ready 'AIR-420' supports two high-performance GPUs for on-premise GenAI post-training. Designed for demanding AI inference and fine-tuning workloads at the edge, it features a compact 28.6L chassis and supports two 600W high-performance GPU cards. Combined with AMD Ryzen 7000/9000 series processors, it offers exceptional AI acceleration in a small footprint, according to the vendor. With an optimized CPU/PCIe layout, a redundant power supply system and the integrated Advantech software tools 'aiSSD', 'aiDAPTIV+' and the no-code platform 'GenAI Studio', the AIR-420 enables efficient on-premise LLM deployment and customization. GenAI Studio is continuously updated to support the latest LLMs and offers both full-parameter and parameter-efficient fine-tuning (e.g. PEFT, LoRA), allowing organizations to quickly adapt AI models to specific edge applications.