First-Ever 100% Domestic AI Semiconductor Deployed in Defense Sector
Proven Performance with Reduced Power Consumption

DeepX, a company specializing in ultra-low power on-device artificial intelligence (AI) semiconductors, is set to deploy its Neural Processing Unit (NPU) in the AI security and surveillance system at an Air Force base in Korea. This marks the first case in which a military video surveillance system, previously built primarily with foreign-made Graphics Processing Units (GPUs), has been completely replaced by domestically produced AI semiconductors.


On August 5, DeepX announced that it will supply its products to the Air Force base AI security and surveillance system as part of the “Pilot Project for NPU-Based Intelligent Video Surveillance Systems,” led by the Ministry of National Defense.

An image depicting DEEPX's domestic NPU-based AI video surveillance technology. DEEPX

An image depicting DEEPX's domestic NPU-based AI video surveillance technology. DEEPX

View original image

This project is being carried out in collaboration with KMNET, a specialized military network company, and PINTEL, an AI video analysis company. DeepX is providing the AI computation platform that runs PINTEL’s video analysis software. The final contract was signed on July 30 following an open competitive bidding process.


The system is designed to provide around-the-clock, real-time monitoring of critical security zones, such as airfield runways and perimeter fences at the Air Force base. When the AI detects movements of people or vehicles or identifies abnormal behavior, surveillance personnel assess the threat and respond accordingly.


The project deploys the “DX-H1 V-NPU,” which processes video decoding and AI inference simultaneously, and the “DX-H1 NPU card for edge servers,” which delivers performance up to 100 tera operations per second (100 TOPS). Since these solutions allow for multi-channel CCTV video analysis using only a single NPU without any GPU, significant reductions in power consumption and overall system deployment costs are anticipated.


Following its previous project converting Ulsan’s municipal CCTV system to public AI, DeepX continues to expand its presence into the Air Force sector, accelerating the commercialization of on-device AI within public safety and defense.


DeepX CEO Kim Nokwon commented, “This project is a meaningful example of implementing an AI video surveillance system previously centered on foreign-made GPUs using domestically developed NPU technology. The V-NPU solution, capable of processing everything from video decoding to AI inference within a single platform, is expected to become a new alternative to enhance the efficiency and cost-effectiveness of AI video analysis infrastructure.”



He added, “Looking ahead, we will expand the application of on-device AI to various sectors requiring AI-based video analysis—including public safety, defense, smart cities, industrial sites, and critical national infrastructure—to further grow the ecosystem for domestic AI semiconductors.”


This content was produced with the assistance of AI translation services.

© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.

Today’s Briefing