SoftBank, a leading Japanese multinational investment holding company, has recently introduced Infrinia AI Cloud OS, a specialized software stack designed specifically for AI data centers. Developed by SoftBank’s Infrinia team, this innovative OS allows data center operators to offer Kubernetes-as-a-service (KaaS) in multi-tenant environments, as well as provide inference-as-a-service (Inf-aaS). This enables customers to access LLMs through simple APIs that can seamlessly integrate with existing GPU cloud offerings.
The Infrinia Cloud OS is poised to meet the increasing global demand for GPU-powered AI solutions, offering a cost-effective and streamlined alternative to internally developed options and custom stacks. By accelerating GPU cloud service deployments and supporting all stages of the AI lifecycle, from model training to real-time usage, Infrinia Cloud OS aims to simplify operations and reduce complexities for data center operators.
Initially, SoftBank plans to integrate Infrinia Cloud OS into its current GPU cloud offerings before expanding its deployment to overseas data centers and cloud platforms worldwide. The rising demand for GPU-powered AI across various industries, coupled with evolving user requirements, underscores the need for advanced GPU cloud solutions.
Infrinia AI Cloud OS is designed to address these challenges by optimizing GPU performance, simplifying management, and enhancing the deployment of GPU cloud services. With its KaaS capabilities, SoftBank’s latest software stack automates various aspects of the underlying infrastructure, from server settings to storage and networking. It also enables quick configuration of GPU clusters to accommodate different AI workloads, while automated node allocation enhances GPU-to-GPU bandwidth for distributed workloads.
The Inf-aaS component of Infrinia Cloud OS streamlines the implementation of inference workloads, making AI model inference more accessible and scalable through managed services. By reducing operational complexities and total cost of ownership, Infrinia AI Cloud OS is poised to drive the adoption of GPU-based AI infrastructure across industries worldwide.
(Image source: “SoftBank.” by MIKI Yoshihito. (#mikiyoshihito) is licensed under CC BY 2.0.)
Want to delve deeper into Cloud Computing and cybersecurity trends? Explore Cyber Security & Cloud Expo, a premier event hosted in Amsterdam, California, and London. This comprehensive event, part of the TechEx series, offers valuable insights from industry leaders. Visit the event website for more information.
CloudTech News is brought to you by TechForge Media. Discover upcoming enterprise technology events and webinars on the TechForge Media website.
