Polaris AI Accelerates Domestic AI Semiconductor Ecosystem Expansion... Unveils On-Device AI and Robotics Strategy

On the 12th, a group commemorative photo was taken at the 'Domestic AI Semiconductor Utilization Site and Meeting' held at the DEEPX headquarters. Photo by Polaris AI

Polaris AI is accelerating the expansion of the domestic AI semiconductor ecosystem by presenting a strategy to broaden the application of on-device AI and robotics based on domestically produced AI semiconductors (NPUs).

Polaris AI announced that on May 12, it participated as a key demand-side company in the "Domestic AI Semiconductor Utilization Site and Meeting" hosted by the Ministry of Science and ICT and the National IT Industry Promotion Agency (NIPA) at the headquarters of DEEPX, a specialized AI semiconductor company. The meeting, presided over by Lee Do-kyu, policy chief at the Ministry of Science and ICT, was organized to discuss the status of AI semiconductor utilization by each company, future plans, and policy support directions.

The event focused on next-generation semiconductor strategies to address the rapidly advancing fields of Physical AI and Agentic AI. In particular, in the Agentic AI environment—where AI autonomously sets and executes goals—the importance of on-device AI semiconductors based on low-latency and low-power processing for real-time operations has emerged as a critical agenda.

At present, Polaris AI is advancing vision AI solutions for industrial sites using DEEPX's NPU-based edge devices. The company aims to minimize network dependency and enable immediate object recognition and anomaly detection on-site, thereby securing both real-time performance and system stability.

The company especially plans to expand the application scope of domestic NPUs by making robotics its next-generation growth engine. To this end, it will actively promote its "Physical AI" business by combining the hardware of global robotics company UBTECH Robotics with intelligent control technology powered by domestic NPUs.

For industrial and service robots to perceive their environments and autonomously perform tasks, infrastructure that enables real-time AI inference at the edge is essential. Polaris AI plans to apply domestic NPUs to UBTECH robots, expanding demonstration projects to improve operational precision, and thereby strengthen its technological competitiveness in the global robotics market.

DEEPX, a key partner in this collaboration, is a domestic fabless company with advanced low-power, high-efficiency AI semiconductor technology. DEEPX began mass production of its on-device AI semiconductor "DX-M1" in August 2025 and has been recognized for its technological prowess in the global market, including winning the CES 2026 Innovation Award.

Polaris AI intends to secure commercialization references for domestic NPUs and contribute to ecosystem expansion through its partnership with DEEPX.

A Polaris AI official who attended the meeting stated, "Accumulating real-world references in industrial settings is crucial for enhancing the competitiveness of domestic AI semiconductors," adding, "We will continue to expand the application of domestic NPUs across various fields such as industrial safety, vision AI, and robotics through public-private cooperation."

Meanwhile, Polaris Group is pursuing the convergence of AI software and robotics hardware as part of its next-generation growth strategy, "Polaris 2.0." Polaris AI is taking a central role as a core demand driver within the group to increase technological self-sufficiency in domestic AI semiconductors and aims to take the lead in the Physical AI market through collaboration with UBTECH Robotics.

In addition, Polaris AI will drive the expansion of the group-wide AI ecosystem by launching demonstration projects that apply UBTECH robotic solutions to manufacturing sites in affiliates such as Polaris Office, Polaris Uno, Polaris Sewon, and Polaris AI Pharma, with a focus on Physical AI.

This event is significant in that it concretized Polaris AI’s direction to establish itself as a key player in the era of Physical AI, leveraging its strategy for integrating on-device AI and robotics technologies based on domestic NPUs.

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