Global Lowest-Cost GPUaaS Elice Cloud Supports AI Research for Universities Worldwide

With the rapid acceleration of AI Transformation (AX), GPU as a Service (GPUaaS)—which allows users to rent high-performance GPUs as needed—is gaining attention. GPUaaS offers scalable access to GPU resources via the cloud, making it especially attractive to AI startups, universities, and research institutions that face challenges in building their own infrastructure or investing heavily in GPU hardware.

Elice Cloud, an AI-specialized cloud service by Elice, provides high-performance GPUs with customizable specifications on demand and without waiting times. This has made it especially popular among AI researchers at universities both in Korea and overseas, for whom time and cost efficiency are critical.

Researcher at the University of Minnesota submits paper to top-tier international journal using Elice Cloud

Recently, Zae Myung Kim, Ph.D. student in the Minnesota NLP group at the University of Minnesota Twin Cities, developed a large language model (LLM) training technique in collaboration with Elice Cloud. He cited several reasons for choosing Elice Cloud, including the ability to use high-performance H100 GPUs at the lowest global price with no wait time, and the pre-configured library environment, which significantly reduced setup time.

Kim emphasized that minimizing research time was critical for paper submission, making Elice Cloud the optimal choice. By accelerating his experiments with Elice Cloud and lowering costs, Kim successfully submitted his paper to NeurIPS (Neural Information Processing Systems)—widely recognized as the most prestigious conference in the AI field.

Elice Cloud users have also expressed high satisfaction with Elice Cloud DataHub, which efficiently manages large-scale datasets used in AI R&D. Donghwan Ji, a master’s student at the AGI Lab in the Department of Artificial Intelligence at Korea University, used Elice Cloud for multimodal LLM research. After experiencing its efficiency in undergraduate coursework, Ji expanded its use to his research. Using the DataHub, he was able to run more than two high-performance GPU experiments in parallel, significantly reducing research time and maximizing productivity.

Elice Cloud minimizes costs for AI researchers by offering high-performance GPUs at globally lowest prices

With its dynamic allocation technology, Elice Cloud allows users to scale resources flexibly, dramatically enhancing cost and resource efficiency for individual researchers, startups, and universities. Furthermore, its cluster technology can bundle high-performance GPUs like H100 to meet the resource-intensive needs of government and enterprise clients.

Jae Won Kim, CEO of Elice, stated,
“We initially developed Elice Cloud between 2021 and 2023 to enhance our internal cost efficiency. In 2023, we expanded to external customers, starting with the A100. As a company focused on AI technology development, we understand how critical it is to access high-performance GPU resources affordably and efficiently. We will continue to advance our AI cloud infrastructure to support AI research at both individual and national levels.”