NHN Cloud is accelerating its transformation by positioning artificial intelligence (AI) infrastructure as a core growth pillar. The company plans to capture AI transformation (AX) demand from both public and private sectors with its next-generation AI full-stack brand "NHN FactoryX," which integrates AI infrastructure, platforms, and services. Over the medium to long term, NHN Cloud aims to raise the AI business to more than 50% of its total revenue.
"NHN Cloud will grow around two pillars: our existing cloud business and FactoryX," Kim Dong-hoon, CEO of NHN Cloud, said at a press conference held at The Plaza Hotel in Sogong-dong, Seoul, on Wednesday. "FactoryX represents the culmination of experience we have accumulated in Korea's GPU infrastructure market since 2019, and it will support the AI transformation across public and private sectors."
FactoryX focuses on solving the problem of corporate AI projects getting stuck at the proof-of-concept (PoC) stage. The company aims to help businesses apply AI services to actual operations and business activities by offering a three-stage integrated execution environment that spans graphics processing unit (GPU) procurement, operation, and AI agent deployment. "Only seven out of 100 Korean companies properly utilize GPUs," Kim said. "The success or failure of AI business depends on how stably and efficiently the infrastructure is operated."
NHN Cloud highlighted its experience of integrating H100 GPUs with domestically produced neural processing units (NPUs) at the Gwangju National AI Data Center as a core competitive advantage. Through the government's "AI Computing Resource Utilization Base Strengthening Project," the company has built 7,656 Nvidia B200 GPUs and commercialized Korea's largest single GPU cluster with 4,080 units, securing capabilities to operate large-scale AI infrastructure. By adopting a 100% liquid-cooled GPU cooling system, the company said it lowered the annual GPU failure rate to about one-third of that of conventional air-cooled systems, improving stability.
NHN Cloud also unveiled its in-house platforms designed to efficiently utilize expensive GPU resources without waste. "GPU Live" automatically separates training and inference workloads and increases GPU utilization through dynamic resource allocation. "AI EasyMaker" supports the entire process from model training to deployment and operation. In the service domain, the company introduced "ProjectX," an AI agent execution environment that allows even non-developers to design enterprise-customized AI agents using natural language.
Through these three core environments, NHN Cloud aims to raise the AI business's share of total revenue from 13% last year to 38% this year and to more than 50% next year. "Through FactoryX, we will support companies in stably executing AI and connecting it to real business growth," Kim said.
