Operational Grid with renewable certificates

CoreWeave AI Data Centers (US)

Livingston, Livingston, NJ; Plano, TX; Manassas, VA; Chicago, IL, United States

Capacity
500MW
Operators
CoreWeave · NVIDIA
AI Focus
GPU cloud for AI training and inference (primarily NVIDIA H100/H200)
Year
2022

CoreWeave is the largest independent GPU cloud in the United States, built from the ground up for AI workloads. Founded as a cryptocurrency mining operation in 2017, the company pivoted to NVIDIA GPU compute and became the preferred cloud for AI companies that needed large-scale GPU access before hyperscalers could scale OpenAI-compatible capacity. By 2025, CoreWeave had raised $19 billion in equity and debt, signed long-term capacity contracts with Microsoft (worth $10+ billion), Meta, and IBM, and operated data centers across New Jersey, Texas, Virginia, and Illinois. The company completed its IPO in early 2025 with a $23 billion valuation. CoreWeave’s infrastructure is built around NVIDIA H100 and H200 GPU clusters, with an emphasis on InfiniBand networking for large training runs. Its customer base includes AI startups, research labs, and enterprises that need GPU capacity without hyperscaler lock-in.

Coordinates
40.7975, -74.3151