Key facts
- Naver Cloud consortium selected for government cybersecurity AI model project.
- Project aims to develop a sovereign AI model tailored to South Korea's security environment.
- Naver Cloud beat rival consortium led by SK Telecom.
- Consortium includes 32 companies and institutions, including LG CNS and LG Uplus.
- Government to provide 256 Nvidia B200 GPUs; consortium to add more GPUs.
- Models will focus on both defensive and offensive cybersecurity tasks.
- Training data will exceed 830 terabytes.
- Models to be tested across seven critical industries.
A consortium led by Naver Cloud Corp. has been chosen by South Korea's Ministry of Science and ICT to develop a specialized artificial intelligence (AI) foundation model for cybersecurity. The project, which aims to enhance national security against escalating cyber threats, saw Naver Cloud's bid outperform a rival consortium headed by SK Telecom Co. The selection was based on factors including Naver Cloud's extensive industry-academia-research partnerships, technological capabilities, and development strategy.
The Naver Cloud-led consortium comprises 32 companies and institutions, including prominent entities like LG CNS Co., LG Uplus Corp., KEPCO KDN Co., Korea Hydro & Nuclear Power Co., and the Financial Security Institute. To support the development over the next 10 months, the South Korean government will provide the consortium with up to 256 Nvidia B200 graphics processing units (GPUs).
In addition to the government-provided hardware, the consortium plans to utilize 4,000 B200 GPUs from Naver Cloud and 256 H200 GPUs from LG. This infrastructure will be used to train two distinct cybersecurity AI models: one based on Naver's HyperCLOVA X, focusing on defensive tasks, and another based on LG's EXAONE, concentrating on offensive capabilities. These models are designed to complement each other in cyberattack detection, analysis, and response.
The training process will involve approximately 830 terabytes of data sourced from Naver Cloud, LG CNS, and organizations managing critical national infrastructure and industrial systems. Following training, the models will undergo testing in seven key sectors: energy, finance, science and technology, telecommunications, semiconductors, defense, and aerospace, with the goal of strengthening AI-driven cybersecurity across these vital industries. The ministry is expected to finalize a contract with the consortium later this month, marking the commencement of full-scale development.