VAST Data has announced an expansion of its partnership with Google Cloud to make its AI Operating System (AI OS) available as a managed service for organisations using hybrid cloud environments. The collaboration is intended to address data management and accessibility challenges faced by enterprises deploying artificial intelligence workloads across both on-premises and cloud infrastructure.
The VAST AI OS uses the company’s DataSpace technology to establish a unified data environment that connects clusters running in different locations. This configuration enables data to be accessed from multiple sites without requiring large-scale migrations. In a demonstration, VAST linked clusters located in the United States and Japan, showing that data could be shared between the two in near real time while running inference workloads using vLLM.
VAST said the partnership aims to reduce the complexity associated with moving and managing large datasets for AI systems. The approach allows enterprises to choose where to run workloads based on performance or compliance needs while maintaining consistent governance policies across environments.
Google Cloud’s Nirav Mehta, Vice President of Compute Platform, said the integration is designed to make it easier for customers to deploy and operate VAST’s data platform on Google Cloud’s infrastructure.
Recent testing has shown that the VAST AI OS can connect to Google Cloud’s Tensor Processing Unit (TPU) virtual machines with model load speeds similar to local NVMe storage. When tested using Meta’s Llama-3.1-8B-Instruct model, the system maintained consistent performance even during initial load phases, suggesting it can support high-throughput AI workloads across distributed environments.
VAST’s platform provides a single data environment that can span on-premises systems and Google Cloud. It supports data streaming, access management, and policy enforcement, aiming to simplify the operational challenges of hybrid AI deployment. The service is now available on Google Cloud for customers seeking to integrate their existing data infrastructure with cloud-based AI resources.
