CoreWeave Touts Multi-Rack Vera Rubin NVL72 System Online
CoreWeave has brought online its multi-rack Vera Rubin NVL72 system, connecting hundreds of next-generation Rubin GPUs into a single scale-out cluster. This milestone strengthens CoreWeave's pitch as one of the earliest cloud platforms able to commercialize Nvidia's newest AI hardware.
The company operates specialized cloud infrastructure designed for artificial-intelligence workloads, leasing high-performance computing capacity to customers training and running large AI models. Its business depends heavily on access to Nvidia GPUs, rapid infrastructure deployment, and keeping expensive computing assets highly utilized.
'CoreWeave was the first AI cloud provider to validate and bring up a Vera Rubin NVL72, demonstrating that this advanced rack-scale architecture could operate as a reliable, high-performance cloud service,' said Chen Goldberg, executive vice president of product & engineering at CoreWeave. A single NVL72 rack combines 72 Rubin GPUs with 36 Vera CPUs.
The company also introduced upgrades to its AI Object Storage platform, including cross-region write acceleration and a new Archive tier. Its Local Object Transport Accelerator can reduce latency by as much as eight times through managed caching.