Lablup is moving to expand its overseas business in earnest, leveraging full-stack artificial intelligence (AI) technology that covers everything from graphics processing unit (GPU) resource management to AI model training, inference and service operations.
Shin Jung-kyu, CEO of Lablup, outlined the company's post-listing growth strategy at an IPO press conference held at the Conrad Seoul hotel in Yeouido, Seoul, on Oct. 19. Lablup is scheduled to list on the KOSDAQ market on Nov. 3.
“Building on our experience collaborating across various cloud environments, we will expand our business model by providing Lablup's solutions and technologies to neocloud operators in Japan, Southeast Asia and Europe,” Shin said.
The company's core competitive strength lies in its vendor-agnostic, full-stack AI infrastructure technology. Its Backend.AI platform provides integrated management of AI accelerators and infrastructure from multiple hardware manufacturers. Lablup has improved resource utilization efficiency through its proprietary GPU partitioning virtualization and AI workload orchestration technologies.
“Lablup provides optimized solutions regardless of the hardware or accelerators being used,” Shin said. “Our strength lies in integrating everything from GPU optimization to the user interface into a single platform without being tied to a specific vendor.”
Backend.AI supports more than 11 types of AI accelerators, including GPUs from NVIDIA, AMD and Intel, as well as neural processing units (NPUs). It provides integrated management of GPU resources, AI model development and training, inference and serving across various environments, including on-premises systems, cloud infrastructure and closed networks.
Lablup has accumulated GPU operations experience at more than 120 sites in South Korea and overseas, including Samsung Electronics, Hyundai Mobis, KT and Shinhan Bank. The company also supported AI infrastructure operations for the Upstage consortium under the government's Sovereign AI Foundation Model development project.
Following its listing, Lablup plans to accelerate its expansion into the global AI infrastructure market. To address growing demand for sovereign AI, the company aims to secure neocloud operators in emerging markets, including Japan, as key customers. It also plans to expand its overseas sales network through partnerships with local resellers in the United States, the United Kingdom and Singapore.
Lablup is also pursuing business expansion through new products. The company plans to introduce a token factory platform integrating AI:GO, which manages AI operations in edge environments, with Continuum Suite, which manages AI model request distribution, token usage and costs.
Proceeds from the IPO will be used to advance research and development and expand infrastructure.
“We will also use the funds raised to acquire testing equipment to validate our platform across various hardware environments,” Shin said. “We also plan to use the proceeds to expand our overseas business, including in the United States.”
Lablup plans to offer 2.18 million shares in its IPO. The indicative offering price range is KRW19,000 to KRW22,000 per share, with total expected proceeds of KRW41.4 billion to KRW48 billion. Public subscription for retail investors is scheduled for Oct. 22-23.
