Overview
- A contract expansion disclosed Thursday puts about $21 billion toward dedicated AI cloud capacity that runs through December 2032.
- CoreWeave will deliver GPU-heavy capacity across multiple locations with some of the first deployments of Nvidia’s Vera Rubin platform to run inference at scale.
- The new pact builds on a prior ~$14.2 billion agreement, taking Meta’s total commitment to roughly $35 billion and boosting CoreWeave’s multi‑year revenue visibility, including a reported $66.8 billion backlog.
- To fund the buildout, CoreWeave launched more than $4 billion in debt offerings, including $3 billion in convertible notes due 2032 and $1.25 billion in senior notes due 2031 for data centers, GPUs, and possible refinancing.
- Meta characterized the move as a portfolio approach that uses specialized providers for immediate compute needs while its own data centers are built out over several years.