Fluidstack operates an AI cloud platform built around massive-scale GPU infrastructure, managing over 100,000 GPUs globally. The company provides single-tenant GPU clusters and infrastructure software to AI labs, governments, and enterprises, offering what it describes as enterprise-grade isolation at hardware, network, and storage levels.
The company's platform rests on two core products. Atlas OS is a bare-metal operating system designed for AI infrastructure that handles fast provisioning and orchestration at scale. Lighthouse, an intelligent monitoring system, optimizes workloads and aims to prevent operational issues before they materialise.
Fluidstack's scale reflects significant commitments from major partners. The company has secured a €10 billion commitment to build a 1-gigawatt AI supercomputer in France, whilst a $50 billion partnership with Anthropic supports the construction of custom data centers across the United States. Headquartered in New York City since 2025, Fluidstack operates across the globe, supporting the deployment of NVIDIA H100, H200, B200, and GB200 systems.