NHN Cloud sells out GPU resources, unveils AI brand 'FactoryX'

This article is by Lee Jae-lim and read by an artificial voice. NHN Cloud's AI chip clusters are running at full capacity, with GPU resources sold out for the year and customers already lining up to reserve capacity for next year, the cloud provider announced Tuesday. The company also unveiled a new brand called "FactoryX," which covers the full pipeline from GPU procurement to AI agent deployment. "All of our GPUs are operating at 100 percent," said NHN Cloud CEO Kim Dong-hoon at a press conference held in central Seoul. "GPU bookings are nearly sold out through the end of this year. We're in talks with several major Korean conglomerates about filling any capacity that may open up next year." He added that across the domestic market, supply continues to fall short of demand, prompting customers to prebook capacity and enter into long-term contracts. NHN Cloud currently operates the National AI Data Center in Gwangju, powered by a mix of Nvidia H100 GPUs and domestically developed neural processing units. Its FactoryX Seoul facility, located in the Yeongdeungpo district of western Seoul, runs a cluster of 7,656 B200 chips which deliver 27.4 exaflops. The company is also planning to build a new data center in Pohang, North Gyeongsang. For customers who have secured GPU capacity, NHN Cloud offers two platform services: GPU Live, which helps businesses maximize GPU utilization with minimal waste, and EasyMaker, a development platform for building AI applications. "The center of gravity in the AI race has shifted away from large models toward operating AI that can run reliably in real business environments while keeping costs in check," Kim said. "Through FactoryX, we will fully support AI transformation across both the public and private sectors." The company expects AI revenue to nearly triple its share this year, from 13 percent to 38 percent, and to fully catch up with its cloud business by next year — with the two contributing roughly equal shares of total revenue. NHN Cloud is also set to launch a service currently known as "Project X" in the second half the year. The platform allows businesses to deploy AI agents across their operations and is designed so that even nontechnical staff can configure customized agents using plain language commands. The agents integrate with a company's existing internal systems and data to automate workflows while keeping costs and security risks low. Beyond Korea, NHN Cloud has its sights on Japan's AI infrastructure market, which Kim sees as lagging behind Korea — and therefore ripe for expansion. "We've been meeting frequently with Japanese firms and believe Korean companies hold a meaningful technological edge in AI and cloud," he said. NHN Cloud plans to pursue data center investment and GPU deployment in Japan, with that push expected to begin between 2027 and 2028.