DataCenterNews Canada - Specialist news for cloud & data center decision-makers
Isometric unified ai cloud control plane kubernetes infra network

VAST Data unveils Polaris to unify AI cloud control

Thu, 26th Feb 2026

VAST Data has launched Polaris, a management layer it describes as a global control plane for distributed AI data infrastructure across public cloud, neocloud and on-premises datacentres.

The product targets organisations that run AI training, inference and data collection across locations and providers. Polaris sits above VAST deployments, coordinating provisioning and ongoing operations across those environments.

Multi-cloud and hybrid patterns have become common in the region. A report cited by VAST Data found 89% of APAC enterprises deploy workloads across multiple public clouds, while 72% operate in hybrid cloud models. That mix can split AI pipelines across geographies and platforms, and increases governance and cost-management demands as data moves between clusters and edge sites.

Control plane

Polaris adds a centralised service delivery layer for VAST environments, automating deployment and lifecycle management and orchestrating operations across multiple clusters. It is designed to treat separate environments as one operational platform, rather than standalone systems managed independently.

VAST positions Polaris as complementary to its AI Operating System and includes it at no extra cost. It is part of VAST cloud deployments, with additional multi-cluster orchestration functions planned for later releases.

Polaris also sits alongside VAST DataSpace, which presents a global namespace across clusters through a distributed fabric. In VAST's description, DataSpace focuses on how data is presented across locations, while Polaris focuses on how clusters are deployed and maintained across cloud and on-premises infrastructure.

"AI infrastructure has outgrown the idea of a single deployment in a single location," said Jonsi Stefansson, General Manager, Cloud Solutions, VAST Data. "Polaris establishes a global control plane that makes distributed AI infrastructure operationally coherent. It allows enterprises to deploy, scale and govern VAST clusters on any cloud or on-premises, all while managing them as one system."

How it works

Polaris is a multi-tenant, Kubernetes-based control plane. A lightweight agent runs on every VAST node, automating provisioning inside customer accounts and coordinating upgrades, expansion and node replacement from a central management layer.

According to VAST, the platform integrates with cloud marketplaces for subscription and entitlement, and includes identity integration, role-based access control and audit logging. These controls are aimed at organisations that need consistent operational policies across multiple cloud regions and on-premises sites.

Polaris can run under different management models: operated by VAST, a partner, or the customer. VAST also says it supports cloud service provider partners and sovereign deployments.

Another goal is scaling across infrastructure types, including deployments spanning large cloud providers and neocloud operators. Polaris also supports multi-site, multi-cluster architectures under centralised management.

Operational model

Polaris uses what VAST calls an intent-driven approach. Administrators define a desired infrastructure state, and Polaris coordinates cloud-native services and VAST software to match those settings and maintain them over time.

Deployments can be created, expanded and governed through a single API and interface. VAST says the approach keeps configuration and policy enforcement consistent across regions and providers, which can be important in AI environments where organisations want common security and data controls while using different compute pools for training and inference.

The launch reflects continued focus on operational coordination for AI workloads that span multiple locations. Many organisations place training where GPU capacity and pricing meet project needs, while inference runs closer to end users or data sources. Edge sites may also collect and pre-process data, creating fragmented operations across clusters and providers.

VAST says Polaris is intended to provide a single operational view across these deployments, aligning with evolving AI pipelines in which teams run long-lived systems rather than projects built and dismantled for each workload.

Polaris is available as part of VAST cloud deployments, with expanded multi-cluster orchestration features planned for future releases.