A domestic Neural Processing Unit (NPU) will be deployed for Air Force base perimeter surveillance. In the defense AI video surveillance field, which has been dependent on foreign Graphics Processing Units (GPUs), this marks the first case of system construction based on 100% domestic NPU, and is evaluated as the beginning of full-scale commercialization of on-device AI semiconductors in the defense sector.

DEEPX announced on the 5th that its NPU has been applied to the Air Force base AI perimeter surveillance system construction project.

This project is part of the Ministry of Defense's "NPU-based Intelligent Video Surveillance System Pilot Construction Project," conducted in cooperation between KMNET and Fintel, with the final contract executed on July 30th following open competitive bidding.

DEEPX's DX-H1 V-NPU and DX-H1 NPU card for edge servers with up to 100TOPS performance are applied in this project.

The V-NPU has a structure that integrates video decoding and AI inference processing in a single chip, designed to enable real-time analysis of multiple CCTV feeds without a separate GPU.

DEEPX supplies an AI computing platform that runs Fintel's AI video analysis software, supporting 24-hour surveillance of key perimeter areas such as runways and fence lines.

When AI detects people, vehicles, and anomalous behavior in real-time, surveillance personnel determine whether there is a threat and respond accordingly.

DEEPX has expanded its application scope from Ulsan Metropolitan City's public AI CCTV conversion project to this Air Force project.

The company stated that it can reduce GPU dependency while lowering system construction costs and power consumption, and improving operational efficiency.

Kim Nok-won, CEO of DEEPX, said, "V-NPU technology that processes everything from video decoding to AI inference on a single platform will become a new alternative to enhance the efficiency and economic viability of AI video analysis infrastructure," and added, "We will expand on-device AI applications not only to public safety and defense but also to smart cities, industrial sites, and national critical infrastructure, contributing to the expansion of the domestic AI semiconductor ecosystem."