Naver Cloud selected for South Korean government cybersecurity AI project

A consortium led by Naver Cloud has been selected by South Korea’s Ministry of Science and ICT to develop a specialized artificial intelligence foundation model for cybersecurity. The consortium, which includes 32 institutions such as LG CNS, LG Uplus, and various universities, beat out a competing bid from SK Telecom.

The project aims to create two distinct 700B-parameter Mixture-of-Experts (MoE) models: one focused on defensive capabilities using Naver’s HyperCLOVA X, and another focused on offensive capabilities using LG’s EXAONE. To accelerate development, the consortium is utilizing massive computing power, including 4,000 Nvidia B200 GPUs from Naver Cloud, 256 H200 GPUs from LG, and 256 B200 GPUs provided by the government.

Training will involve approximately 830TB of high-quality data from sectors including power, finance, telecommunications, and defense. The resulting models will be tested across seven key industries and are intended to operate within closed networks to ensure security for critical national infrastructure. The consortium also plans to release the models in an open-source format for commercial use to foster a domestic cybersecurity AI ecosystem.