Dell Technologies confirmed that it is now shipping its liquid-cooled PowerEdge XE9812 rack systems equipped with NVIDIA Vera Rubin NVL72 to CoreWeave, marking one of the earliest public deployments of NVIDIA’s next-generation AI infrastructure platform. The announcement surfaced during NVIDIA GTC at COMPUTEX, where NVIDIA CEO Jensen Huang referenced the shipment while discussing the build-out of next-generation AI factories.
The deployment places Dell among the first OEMs bringing rack-scale Vera Rubin infrastructure into production. The PowerEdge XE9812 is designed for dense AI training and inference workloads and integrates direct liquid cooling to manage the thermal demands of NVIDIA’s Vera Rubin NVL72 platform. CoreWeave confirmed that it has a “fully working NVIDIA Vera Rubin NVL72” system online, crediting close engineering coordination across hardware, software, networking, and operations.
The shipment also signals how quickly the AI infrastructure supply chain is moving from Blackwell deployments toward Rubin-based platforms. CoreWeave has emerged as one of the fastest-moving cloud operators for large-scale GPU deployment, while Dell continues expanding its role as a systems supplier for hyperscale AI clusters. The use of liquid-cooled rack-scale infrastructure underscores how thermal design, networking, and power delivery are becoming core differentiators as AI deployments move from pilot projects to full production environments.
- Dell confirmed shipment of liquid-cooled PowerEdge XE9812 racks to CoreWeave
- Systems feature NVIDIA Vera Rubin NVL72, NVIDIA’s next-generation accelerated computing platform
- Jensen Huang referenced the shipment during his COMPUTEX appearance at NVIDIA GTC
- CoreWeave says it is among the first to bring a fully operational Vera Rubin NVL72 system online
- Deployment highlights increasing use of direct liquid cooling for high-density AI compute infrastructure
- Signals broader industry transition from Blackwell-era systems toward Rubin-based AI clusters
