Elice Cloud Unveils Solution Optimized for Open AI Environments

Elice Group has unveiled a new solution for 'Elice Cloud,' which is set to bring innovation to the AI cloud market, thereby paving the way for the safe and convenient utilization of various AI models, including open-source ones, in a user-friendly cloud environment. The core feature introduced is the 'ML API Library,' which significantly upgrades existing ML APIs to lower the barrier to utilizing open-source AI models. Previously, complex environment settings and fine-tuning were essential for each open-source model, but Elice Cloud has streamlined this process. It provides customized APIs that allow immediate use of general-purpose and open-source AI models like DeepSeek R1 and Meta Llama in an optimized GPU/NPU environment without separate installation or configuration. This simplifies the entire process from AI chatbot development to management and deployment, maximizing development efficiency with high-performance computing services utilizing the latest NPUs from FuriosaAI and Rebellions. Elice Cloud operates in an 'Elice AI PMDC (Portable Modular Data Center)' environment, a private cloud based on edge data centers, for the secure processing of sensitive data. Its modular design allows for physical isolation, completely blocking the risk of external leakage, and flexibly responds to changes in AI demand. Furthermore, it provides sustainable AI infrastructure with excellent energy efficiency (PUE of 1.27). Currently, they are designing and building 1MW-class AI PMDCs (per unit) with the goal of operation within the year, having introduced space-optimized design and Direct Liquid Cooling (DLC) technology for high-density power and thermal management. This signifies industry-leading infrastructure that stably supplies 160kW of power per rack and perfectly supports the latest high-performance GPUs such as Nvidia B100. Kim Jae-won, CEO of Elice Group, stated, "We have already designed and built AI PMDCs capable of accommodating 5,000 high-performance GPU H100s, which will be the largest scale among domestic cloud companies." He further expressed his ambition to "fully launch the Elice Cloud business through GPU acquisition, R&D, and technology investment."