NVIDIA is set to acquire GPU orchestration software provider Run:ai in a move geared towards enhancing cluster resource utilization for AI workloads across shared accelerated computing infrastructure. The Israeli startup focuses on promoting efficiency in managing AI computing resources.
As AI deployments grow in complexity across cloud, edge, and on-premises data centers, the need for sophisticated scheduling to optimize performance at both system and infrastructure levels is paramount. Run:ai's Kubernetes-based workload management and orchestration software enables enterprise customers to effectively manage and optimize their compute infrastructure irrespective of whether it is on-premises, in the cloud, or in hybrid environments.
By building an open platform on Kubernetes, the foundational layer for modern AI and cloud infrastructure, Run:ai caters to various industries and enterprises globally. Their platform facilitates centralized management of shared compute infrastructure, offering functionalities such as user management, resource access control, quotas, priorities, GPU pooling, and efficient GPU cluster utilization.
Omri Geller, co-founder and CEO of Run:ai, emphasized the collaboration with NVIDIA since 2020 and the shared passion for infrastructure optimization. The acquisition will further enhance NVIDIA's offerings, ensuring seamless integration of Run:ai's capabilities into the NVIDIA DGX Cloud, a comprehensive AI platform designed in partnership with leading cloud providers.
Customers using NVIDIA DGX and DGX Cloud will benefit from Run:ai's solutions tailored for AI workloads, especially large language model deployments. Moreover, the joint efforts of NVIDIA and Run:ai will aim to support a wide range of third-party solutions, maintaining flexibility and choice for customers.
Together, NVIDIA and Run:ai aim to streamline GPU utilization, enhance infrastructure management, and provide customers with a unified fabric accessing GPU solutions across different environments. This collaboration promises improved efficiency, flexibility, and enhanced management of GPU resources within an open architecture framework.
