Samsung SDS said on the 23rd that it launched GPUaaS (GPU subscription service) based on Nvidia's latest graphics processing unit (GPU) "B300 (Blackwell Ultra)" through its in-house cloud Samsung Cloud Platform (SCP), the first in Korea.
The service launch is a strategy to meet the surging demand for high-performance computing as corporations move beyond developing artificial intelligence (AI) models to the "AI inference" stage of applying them to real services.
The B300 GPU is equipped with 12-stack HBM3E (high-bandwidth memory), providing 288GB (gigabytes) of memory capacity and 8TB (terabytes) per second of bandwidth per GPU. Based on this, in AI inference areas that require complex computation, memory performance has improved compared with the H100 by 3.6 times in capacity and 2.4 times in bandwidth.
Accordingly, the company said the data bottleneck—where overall performance drops because memory data transfer is slower than the GPU's fast compute speed when running large language models (LLMs)—has been dramatically improved.
Samsung SDS proactively introduced GPUaaS based on the A100 in 2021 and the H100 in 2023, leading the formation of a GPUaaS ecosystem so that GPUs can be used as AI-dedicated infrastructure for cloud-based infrastructure build-out, operations, and customer services.
A company official said, "Customers adopting SCP B300 GPUaaS can efficiently handle large AI models through high-capacity memory, minimizing latency for high-performance AI services such as AI agents and image, video, and code generation and analysis."
They also emphasized that using a subscription model—paying for only what is needed—can lower initial investment risk and optimize expense. Even amid tight GPU supply, the latest Nvidia architecture can be immediately adopted for work through SCP, and sensitive corporate data can be processed in a secure cloud environment combined with Samsung SDS's security capabilities.
Samsung SDS plans in the third quarter of this year to launch a serverless inference service, in which only the amount of tokens used is paid with no separate infrastructure usage fee when applying AI models, and an AI training service that automatically and instantly distributes AI training when developers input code and data.
Lee Ho-jun, head of the Cloud Services Business Unit (executive vice president) at Samsung SDS, said, "Based on SCP's GPU efficiency capabilities, such as resource optimization and energy savings, we will actively support AX (AI transformation) by providing the first-in-Korea B300 GPU service to customers seeking to apply AI to their work in large corporations, mid-sized and small businesses, and the public sector."