Naver Cloud Selected for National Security AI Project

The Ministry of Science and ICT (MSIT) has selected the Naver Cloud consortium as the final executing agency for the national project 'Development of a Cybersecurity-Specialized AI Foundation Model'. The consortium includes major security and AI companies such as VIDRAFT, LG CNS, and ESTsecurity, as well as academic and research institutions like Korea University and KAIST.

Project Scale and Support Details

The selected team will conduct development for 10 months starting from September 2026 and will receive support in the form of 256 NVIDIA B200 GPUs. The support is divided into 5-month intervals, operating on a phased basis where the remaining duration is determined based on interim evaluation results.

VIDRAFT's Role: AI Safety Diagnostics

VIDRAFT provides its AI safety diagnostic technology 'AX-RAY' as a core capability. AX-RAY is an 'AI that diagnoses AI', precisely checking the reliability, safety, and hallucination risks of trained models. It plays a role in filtering out model vulnerabilities before deployment in environments where errors are fatal, such as cybersecurity.

Consortium Composition and Evaluation

The Naver Cloud consortium received high ratings for its balanced composition involving industry, academia, and research, as well as the excellence of its Agent and Harness design strategies. This project aims to build trustworthy national-level security AI for public, defense, financial, and energy sectors where data cannot be transmitted externally.