Artificial intelligence (AI) semiconductor startup Mobilint will supply its self-developed Neural Processing Unit (NPU) to about 50 Navy vessels.
Mobilint said on the 19th that it will supply its NPU for the Navy's "NPU-based intelligent video surveillance system construction project." The project will build AI-based video analysis infrastructure on Navy vessels, with about 50 ships targeted.
An NPU is a semiconductor specialized for AI computation. It can run AI models that analyze video or images in real time at low power, so it is also used in edge environments such as closed-circuit (CC) TV.
Under this project, a system will be built to analyze data collected by video equipment installed on vessels with NPU-based AI. It is an Edge AI approach that performs AI computation on site without relying on external data centers or the cloud.
With this supply, Mobilint has expanded the NPU use cases it has secured in the public sector into the defense sector. The defense sector requires long-term stable system operation, and security and reliability are critical, making it a market where the real-world operating performance of AI Semiconductor can be verified.
Mobilint is developing the AI accelerator "ARIES" and the system-on-chip (SoC) "REGULUS." ARIES targets servers and on-premises environments, while REGULUS is aimed mainly at edge devices such as cameras and robots.
Shin Dong-ju, CEO of Mobilint, said, "This project will show that domestically produced NPUs can be used even in the real operating environments of defense sites," and added, "Based on experience accumulated in the public and industrial sectors, we will expand applications into various areas, including defense."