en.Wedoany.com Reported - Korea Information Engineering Services Co., Ltd. recently held an "HPE DeepX Visual AI Business Meeting" for channel partners, showcasing a visual AI solution that integrates DeepX Neural Processing Units (NPUs) with HPE servers.

Provided by this IT infrastructure specialist, the solution's hardware architecture is designed to reduce the high costs associated with the initial deployment and maintenance of visual AI systems, which typically constitute a major barrier to large-scale adoption. By integrating DeepX's ultra-low-power AI accelerators, the NPU embeds the LPDDR5 controller directly into the chip, a design that the company says helps eliminate system bottlenecks.

According to Korea Information Engineering, the platform can reduce annual power consumption by approximately 86% compared to traditional GPU-based servers, thereby lowering operational expenses. Additionally, the chip price is about 30% to 50% lower than comparable GPUs, and combined with a reduced infrastructure footprint, it can significantly cut deployment costs, reducing total cost of ownership (TCO) by up to 90%.

The system is scalable and easy to deploy, utilizing standard PCIe form factor accelerator cards, with a single server supporting up to 256 channels. The company states that compatibility testing with HPE servers has been completed, enabling a plug-and-play deployment model. Korea Information Engineering signed a formal distribution agreement with fabless edge AI semiconductor company DeepX in 2025 as part of its business diversification strategy. The visual AI solution has completed a proof of concept at a major South Korean enterprise.

This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com