VAST data partners with Megaport to enhance AI infrastructure

VAST Data and Megaport collaborate to streamline AI workloads across hybrid and multicloud environments.

  • Tuesday, 23rd June 2026 Posted 3 days ago in by Katy Hill

VAST Data has announced a partnership with Megaport, in which Megaport plans to use VAST’s AI-oriented data platform as part of its infrastructure development.

Megaport, which provides network and computing services, is expanding its offerings beyond automated connectivity into integrated compute and GPU services, particularly following its acquisition of Latitude.sh. As part of this expansion, VAST will contribute enterprise data services intended to support Megaport’s infrastructure.

The combined aim is to enable customers to deploy and manage AI workloads across distributed environments, including hybrid, multi-cloud, and geographically dispersed infrastructure setups. This reflects a broader industry trend where AI workloads increasingly operate across multiple clouds, data centres, and regions.

A key challenge in these environments is that network, compute, and data layers are often managed separately, which can create operational complexity. Megaport and VAST are addressing this by aligning their respective capabilities: Megaport provides programmable connectivity across more than 1,100 data centres, while Latitude.sh provides bare-metal and GPU compute services, and VAST contributes unified data services intended to work across distributed environments.

Within this structure, VAST’s DataSpace is positioned as a way to provide a global namespace for data access and management across different environments, including on-premises systems, public clouds, and edge locations. The goal is to reduce data fragmentation and enable more consistent access across locations where workloads are running.

Overall, the collaboration combines connectivity, compute, and data infrastructure components into a more integrated model intended to support distributed AI workload deployment and management across multiple environments.