GMO Internet, Inc., a member of the GMO Internet Group, has decided to offer the NVIDIA HGX B300, a cloud service equipped with NVIDIA’s next-generation NVIDIA Blackwell Ultra GPU , as part of its high-performance GPU cloud service, GMO GPU Cloud, built with NVIDIA technology.
This service, one of the first of its kind in Japan, will involve the procurement of 25 servers equipped with NVIDIA B300 (a total of 200 servers), and is scheduled to be rolled out sequentially from fiscal year 2025.
Since its launch in November 2024, GMO GPU Cloud has been adopted as a platform for many AI developments, including robotics and autonomous driving AI. The introduction of the NVIDIA B300 instances marks the addition of bare metal to the lineup as part of GPU cloud services’ commitment to cutting-edge technology. GMO Internet will provide a stable infrastructure from within Japan capable of handling all cutting-edge AI workloads, including generative AI, large-scale language models (LLMs), image generation, and robotics.
Also Read: FLOW Digital to Launch New Data Center in Central Tokyo
In recent years, the development of proprietary large-scale language models (LLMs) by companies and the rapid evolution of AI and robotics in industrial fields have led to a growing demand for large-scale computing resources that enable advanced computational processing. To meet the demands of larger AI models, real-time inference, and support for diverse generation formats, cloud infrastructure must be updated as a constantly evolving “dynamic technology platform,” a key to technological innovation.
On April 15, 2024, GMO Internet received certification from the Ministry of Economy, Trade and Industry (METI) for its supply assurance plan for the “Cloud Program” of designated important materials under the Economic Security Promotion Act. As of November 22, 2024, GMO Internet has been providing services using the NVIDIA H200 GPU , one of the fastest GPUs in Japan, as a cloud service provider.
In light of these market conditions and technical requirements, we have decided to review our plan to invest in additional NVIDIA H200 GPUs, which was announced on May 14, 2025, and switch to the higher-performance next-generation GPU, the NVIDIA B300.
As a result, GMO GPU Cloud will strengthen its support for the further computational performance, scalability, and low-latency communications required for future AI development and operation, and will develop as a more advanced infrastructure platform.
SOURCE: PRTimes