Advantech Focuses on Compact Servers for On-Site Inference and GenAI

Advantech Unveils New Compact AI Edge Platforms: AIR-410 and AIR-420

Advantech has introduced two new compact AI-focused edge platforms: the AIR-410 and AIR-420 models. These systems are designed for visual inference and the optimization of large language models (LLMs). Aimed at reducing costs, footprint, and latency, they promise strong performance for demanding industrial applications.

AI at the edge is poised for significant growth as a pivotal tool for decentralized computing, particularly for industries concerned with latency, data sovereignty, and minimizing cloud dependency. With these new systems, Advantech directly addresses the competition in on-site AI architecture, targeting sectors such as medical imaging, semiconductor inspection, and automated industrial vision.

The Advantech AIR-410 is pitched as a compact inference solution, featuring an AMD Ryzen 8000 series processor. It supports a PCIe x16 slot for a GPU (up to 550W) all within a compact, quiet form factor measuring below 38 dBA. The system supports up to 96 GB of DDR5 memory and offers 14 TB of local storage. For easy deployment at the edge, its software environment includes the Edge AI SDK and pre-validated models.

### Promising Energy Efficiency

On the other hand, the AIR-420 is tailored for post-training local LLM scenarios. It allows for two high-performance GPUs, each with 600W, and comes housed in a 28.6-liter chassis with a redundant power supply. This model features the no-code GenAI Studio environment, which is regularly updated to support fine-tuning methods like PEFT and LoRA. It is especially suited for applications requiring local adjustments to LLMs on sensitive or non-centralized data.

Both models leverage an integrated architecture that combines AMD CPUs, industrial GPUs, and optimized software. Advantech claims this approach could reduce the total cost of ownership by up to 96% compared to traditional rack-mounted infrastructure, minimizing expenses related to energy, maintenance, software licenses, and network communications. The company also highlights enhanced data security, enabled by complete localization of processing.

This strategy aims at industrial enterprises, AI solution integrators, and healthcare service providers seeking to deploy AI directly on-site. The compactness of these units facilitates installation in constrained environments, while the pre-integrated software ecosystem lessens reliance on specialized external resources.

### A Catalyst for Sovereignty and Real-Time Performance

The AIR-410 and AIR-420 offerings align with a structural trend toward the localization of AI computing, driven by regulatory requirements, data privacy needs, and real-time performance demands. By providing ready-to-use systems that integrate hardware, software, and interfaces, Advantech seeks to simplify on-site AI adoption without sacrificing flexibility or customization possibilities.

Furthermore, the company leverages its European Design and Manufacturing Services (DMS) network to offer tailored services for integrators, accommodating specific business environments, thermal constraints, or existing deployment models. This local adaptability strengthens the value proposition in a market where AI edge systems are still predominantly cloud-centric.

Scroll to Top