0
Naver Cloud showcases its defense AX (AI Transformation) strategy, including its internally developed "HyperCLOVA X SEED 4B" model, at the 2026 Korea Institute of Military Science and Technology Annual Conference held in Jeju from June 10 to 12. / Naver Cloud

Naver Cloud is accelerating its push into the defense AI market by unveiling a defense-specialized AI model capable of operating in resource-constrained computing environments, such as drones and tactical vehicles. Moving beyond data center-centric generative AI, the company is expanding its defense AI business by introducing a lightweight AI model applicable to actual battlefield environments.

Naver Cloud announced on June 15 that it recently unveiled its internally developed lightweight omnimodal model, "HyperCLOVA X SEED 4B," at the 2026 Korea Institute of Military Science and Technology Annual Conference held in Jeju.

HyperCLOVA X SEED 4B is an omnimodal AI model that simultaneously processes text, images, video, and audio. It is specifically designed to perform integrated analysis on diverse battlefield intelligence, including drone footage, satellite imagery, radio communications, and operational documents.

A key feature of this model is the integration of Naver's proprietary vision encoder, "HyperCLOVA X CLIP," and an audio encoder. It was trained on Korean-language documents and data, making it optimal for domestic military environments.

Given that real-time processing and security are paramount in defense AI, systems must run stably not only in data centers but also within constrained computing environments like drones, unmanned systems, and tactical vehicles. Naver Cloud explained that it reduced the model size by half compared to standard 8B models using pruning and knowledge distillation techniques, while actually improving performance over previous versions.

Accordingly, the model is expected to be deployed across various defense sectors, including object detection based on drone and coastal surveillance footage, satellite imagery analysis, automatic military equipment recognition, and battlefield map analysis. It can also be utilized for automating Intelligence, Surveillance, and Reconnaissance (ISR), unmanned systems, and building integrated situational awareness systems.

Naver Cloud also disclosed its defense AI roadmap during the event. The company plans to establish a sovereign AI-based defense AI system that allows infrastructure, AI models, and agents to operate within air-gapped networks, with plans to expand AI agents across all military sectors by 2030.

After laying the foundation for defense AI and expanding its participation in military projects through this year, Naver Cloud intends to advance autonomous operational support systems to drive military-wide AI transformation.

"The defense sector requires both robust security and operational autonomy," a Naver Cloud official said. "We plan to support the military's AI transformation through a sovereign AI-based defense full-stack capable of operating even within air-gapped network environments."

                                                                                                            Lee Seo-yeon