GMI Cloud
GMI Cloud provides an AI-native inference cloud, built on NVIDIA technology. It offers a unified platform for compute, inference, and agents, designed for production-scale AI infrastructure. The service includes GPU clusters, optimised inference, and agent runtime, supporting both serverless and dedicated GPU deployments. GMI Cloud supplies GPU infrastructure, including dedicated bare metal GPUs and cluster orchestration. They offer transparent pricing for NVIDIA H100, H200, and Blackwell platforms, focusing on performance and reliability for AI workloads. The platform supports serverless inference by default, with automatic scaling and cost optimisation.
At a glance
Facilities tracked
2
Markets
0
Countries
1
Total potential
50+ MW
Total potential counts every megawatt once, including expansion headroom at existing sites.
Company Overview
Website
Visit WebsiteGMI Cloud facilities
Every GMI Cloud site with a page on DC Atlas.
United States
- GMI Cloud Data Center ColoradoUndisclosed · Announced
- GMI Cloud Data Center United StatesUndisclosed · Operational
Browse every operator in United States
Recent Activity
GMI Cloud seeks $300M loan for Nvidia chips at Thailand data centre
US-based data centre operator GMI Cloud is seeking a US$300 million loan to purchase Nvidia AI chips for its facility in Thailand hosted by STT GDC Pte. The chips will be used by Tencent Holdings Ltd., Bloomberg reports. GMI Cloud, an Nvidia partner, approached banks and private credit funds for the financing, with yields expected in the low teens.