Samsung SDS (Samsung SDS) will support AI research at domestic universities and government-funded research institutions by utilizing NVIDIA's latest graphics processing unit (GPU), the B300.
On the 11th, Samsung SDS announced that it was selected as a GPU resource provider for the "2026 AI Research Computing Support Project" led by the Ministry of Science and ICT and the Information and Communications Technology Planning and Evaluation Institute (IITP), with services starting on the 20th of last month.
This project provides AI research GPUs to universities and government-funded research institutions that find it difficult to secure large-scale computing infrastructure independently. The project's scale is approximately 15.3 billion won, and services will be provided until March next year.
During the selection process for the supplier, evaluations were conducted on GPU supply plans, resource scale, research environments, operational support systems, and security capabilities. Among multiple selected providers, Samsung SDS was chosen as the top-ranked provider.
Samsung SDS provides a B300-based GPU cluster alongside NVIDIA's H100. The B300, which applies the Blackwell Ultra architecture, is a product that enhances computational performance and memory capacity required for large-scale AI model training and inference.
It has also built clusters that allow multiple GPUs to be used as a single resource, along with high-speed interconnect networks. CPU, memory, storage, network, and development environments are pre-configured so researchers can focus on their work without needing to set up infrastructure separately.
In March, Samsung SDS launched B300-based GPU-as-a-Service (GPUaaS) through its Samsung Cloud Platform (SCP). Last month, it also unveiled an NPU service based on Fugro AI's second-generation neural processing unit (NPU), "Renegade."
Lee Ho-jun, Executive Vice President and Head of the Cloud Services Business Division at Samsung SDS, stated, "We will provide stable services to enable domestic researchers to achieve results based on world-class AI infrastructure."
in TECH
