Artificial intelligence (AI) infrastructure corporations Elice said on the 4th it was selected as a participating company in the "AI computing resource utilization base reinforcement project" promoted by the Ministry of Science and ICT and the National IT Industry Promotion Agency (NIPA), and will build and operate a B300-optimized modular AI data center.
Elice will deploy a total of 2,560 B300 graphics processing units (GPUs) across four modular AI data center buildings, installing 640 units in each. A company official said, "We completed a data center design optimized for ultra-high power and high-density environments, and optimized the cluster architecture, cooling, network, and power efficiency for the latest B300 GPU."
The goal is to maintain a power usage effectiveness (PUE) of 1.1 at the data center. It will also apply a direct liquid cooling (DLC) method and an ultra-high-speed network architecture optimized for large-scale GPU clusters.
The cooling system will use a hot-water cooling method that utilizes cooling water at over 40 degrees and apply outside-air cooling. The measure aims to reduce the power consumption required for cooling and improve operational efficiency.
Chief Executive Kim Jae-won of Elice said, "Through this project, we will build a modular AI data center optimized for high-performance GPU environments and provide a stable and efficient AI computing infrastructure."