[Edaily Reporter Han Kwangbeom ] Consortia led by SKTelecom(017670)and Naver (NAVER(035420)) Cloud have submitted bids for the government’s “Development of AI Foundation Models Specialized in Cybersecurity” project. This has set the stage for a head-to-head competition between leading ICT companies that represent the full breadth of Korea’s AI and cybersecurity capabilities.

The Ministry of Science and ICT announced on the 26th that, following the call for proposals, a total of two consortia—the SKTelecom consortium and the Naver Cloud consortium—submitted proposals. The Ministry plans to swiftly review the eligibility of the submitted documents and conduct presentation evaluations by external experts to select one final project team in early September.

A look at the consortium members reveals a massive gathering of key players from the domestic security and IT ecosystem. The SKTelecom consortium comprises 13 companies and institutions, including Upstage, SK Shields, SecuI, AHNLAB, INC., IGLOO Corporation, RaonSecure Co., Ltd., GENIANS, INC., and the industry-academic cooperation units of Korea University and Soongsil University.

The Naver Cloud consortium has built a large-scale coalition by securing 32 companies, institutions, and universities as allies, including major LG Corp. affiliates such as LG CNS (#LG CNS), LG AI Research, and LG Uplus(032640), as well as East Security, SANDS LAB Inc., Sparrow, Tori, the Korea Financial Security Institute, Korea Hydro & Nuclear Power, Seoul National University, and KAIST.

The government’s push to develop security-specialized AI is driven by a growing sense of crisis over the rapid sophistication of cyber threats, fueled by the recent proliferation of generative AI and Agent AI technologies. Critics have long pointed out that existing defense systems have reached their limits, as AI is now being exploited across the entire spectrum of cyberattacks—beyond merely identifying hacking targets—to include vulnerability scanning, attack path analysis, malware generation, and the execution of automated attacks.

Consequently, calls have grown louder for defense systems to transition to an AI-based approach to counter the AI used in attacks, and securing a proprietary security AI foundation model optimized for the domestic cyber environment has emerged as a national priority.

The single consortium ultimately selected through this call for proposals will receive 256 NVIDIA B200 GPUs (32 nodes)—NVIDIA’s latest AI accelerators—for a period of 10 months. The selected team will develop a security-specialized AI model and propose areas for the demonstration and application of agent-based security services tailored to the domestic cyber threat landscape.

The government plans to release an interim results model five months after the project begins to evaluate the technical level and performance, and then decide whether to provide follow-up support. Upon completion of development, the results will be released as open source to support the AI transformation (AX) of the domestic security industry as a whole, and starting in 2027, the government plans to fully accelerate model advancement and industry adoption through follow-up projects.