Fluidstack

Fluidstack operates as a specialized infrastructure provider for frontier artificial intelligence systems, focusing on the rapid deployment of massive-scale compute clusters measured in gigawatts. Rather than offering generic cloud computing services, the company targets the most demanding AI workloads that require tens of thousands of high-performance accelerators (such as GPUs or TPUs) working in concert. Its core function is to acquire, design, build, and operate data centers from the ground up, covering the entire lifecycle from securing power purchase agreements and navigating utility grid interconnection to constructing bespoke facilities, installing cooling systems, and managing the hardware and software stack that makes clusters production-ready. A distinguishing capability is its ability to compress the typical 18- to 24-month timeline for building a gigawatt-level AI data center into roughly six months. This acceleration is achieved through proprietary pre-engineering, modular construction techniques, tight integration with chip suppliers and networking vendors, and a lean operational approach that eliminates delays common in traditional data center builds. The company’s current flagship project involves deploying a $50 billion compute build-out for Anthropic, described as one of the largest infrastructure projects in United States history. This project entails securing contiguous land parcels, negotiating with local utilities for multi-gigawatt power allocations, constructing hundreds of thousands of square feet of server halls, and installing advanced liquid cooling and high-bandwidth interconnects to handle the extreme heat and traffic generated by training large language models. Beyond construction, Fluidstack also operates the finished clusters. This incl