XIILAB Accelerates Market Expansion as National AI Infrastructure Investment Grows

Vision AI company XIILAB (189330) announced on the 24th that it will introduce its core technologies and business structure through an online vision briefing session and expand its GPU-based AI solution and digital twin businesses in line with the government’s full-scale investment in AI infrastructure.

The government plans to invest approximately ₩100 trillion over the next five years, starting in 2025, to build AI data centers, GPU servers, and smart safety systems. During this period, demand for more than 50,000 GPU servers is expected.

XIILAB anticipates strong market demand for its GPU infrastructure management solution, AstraGo, which can reduce GPU server management costs by more than 50% by optimizing automatic GPU resource allocation and task scheduling efficiency.

The company is also responding to the rapidly increasing need for image analysis AI through XAIVA, its CCTV-based real-time AI video analysis solution. XAIVA is being applied in various industrial fields, including real-time risk detection and prevention, manufacturing process quality control, and bio-GMP environment verification.

In particular, XIILAB plans to generate revenue through industry-specific solutions such as XAIVA On Device, a kiosk-type AI image analysis system, and XAIVA Micro, designed for the semiconductor sector.

In its digital twin business, XIILAB has commercialized virtual verification technology for semiconductor manufacturing lines after obtaining official reseller qualification for NVIDIA Omniverse. The company plans to accelerate digital transformation (DX) initiatives in semiconductor and manufacturing environments beginning this year.

“Now that national-level investment in AI infrastructure is fully underway, AstraGo and XAIVA—our core solutions—will be among the first to benefit,” said Yoon Se-hyuk, CEO of XIILAB. “We aim to strengthen our position in the domestic physical AI sector by building on our existing capabilities in AI video analysis and simulation, including our development of VLM, a video–language multimodal AI model.”

https://www.edaily.co.kr/News/Read?newsId=04336166642237800&mediaCodeNo=257&OutLnkChk=Y