CoreWeave brings up multi-rack NVIDIA Vera Rubin NVL72 and adds object-storage features
CoreWeave said it has brought up multi-rack NVIDIA Vera Rubin NVL72 on CoreWeave Cloud, linking hundreds of NVIDIA Rubin GPUs into one scale-out cluster aimed at agentic AI workloads. CoreWeave also announced two CoreWeave AI Object Storage features: cross-region write acceleration and a new Archive tier. The release does not disclose megawatts, annual recurring revenue, deal value, site locations, or customer contract terms.
Why it matters
A product and operations claim about next-generation rack-scale GPU availability and storage plumbing, not a power or offtake filing.
Companies
- CoreWeave
Risks
- Bring-up is not the same as broad commercial availability or signed capacity deals.
- No independent verification of cluster size or performance claims in the release.