CoreWeave: GPU Cloud and AI Infrastructure

An AI cloud with an operating software layer

CoreWeave provides infrastructure and software for training, serving and managing AI workloads. Its own history dates the business to 2017 and its public listing to March 2025. The platform combines accelerated computing with the operational tools needed to run large deployments. [1]

In January 2026, NVIDIA invested $2 billion in CoreWeave common stock. The announcement also outlined collaboration on infrastructure and software and a goal of more than 5 GW of AI factories by 2030. The investment was a specific equity transaction; the larger capacity figure was a future buildout objective. [2]

CoreWeave also relies on other developers for physical facilities. Applied Digital's October 2025 first ready-for-service phase delivered 50 MW within a larger leased campus. That tenant relationship illustrates how CoreWeave's computing service can use leased infrastructure without owning every facility or treating all contracted megawatts as already delivered. [3]