CoreWeave, Inc. Class A Common StockCoreWeave launched multi-rack NVIDIA Vera Rubin NVL72 clusters and upgraded AI Object Storage with cross-region write acceleration and a new Archive tier.
CoreWeave has introduced multi-rack NVIDIA Vera Rubin NVL72 clusters designed to run training and inference across hundreds of GPUs as one system. The company also upgraded its AI Object Storage with cross-region write acceleration to support high throughput data pipelines for agentic AI workloads, and added a new Archive storage tier targeting colder AI datasets to align long term model data retention with lower cost storage options. CoreWeave operates as a US-based cloud infrastructure provider focused on high performance compute, placing it directly in the buildout of GPU-heavy data centers that many AI developers rely on. The company said the Rubin clusters and storage upgrades support its role as foundational plumbing for compute hungry agentic systems, tying GPU density directly to data locality and higher throughput. CoreWeave has leaned on diverse funding sources, including recent convertible notes and equity offerings, to push more capacity into the ground rather than reshape its business model.
CoreWeave, Inc. Class A Common StockCoreWeave launched multi-rack NVIDIA Vera Rubin NVL72 clusters and upgraded AI Object Storage with cross-region write acceleration and a new Archive tier.
NVIDIA CorporationCoreWeave's new clusters are built on NVIDIA's Vera Rubin NVL72 platform, indicating demand for NVIDIA's GPU systems.