FluidStack is a New York-based “neocloud” provider — a new breed of company leasing high-performance GPU compute capacity to organisations developing AI models and workloads. Founded in 2019, the company originally connected users with idle GPU capacity from underutilised servers, before pivoting to build out dedicated AI infrastructure for the AI boom.
FluidStack partners with leading AI labs, governments, and enterprises to deploy large-scale compute infrastructure at speed. The company manages a network of over 100,000 GPUs and counts clients such as Meta and Mistral among its customers. It sits at the intersection of two powerful secular trends — the buildout of AI infrastructure and the surging demand for dedicated, high-performance compute.
