CoreWeave made NVIDIA Vera Rubin NVL72 available, with Cognition as its first customer running production workloads on the system.

CoreWeave announced the availability of NVIDIA Vera Rubin NVL72 on its cloud on September 30. Its announcement identifies Cognition as the first customer anywhere running production workloads on the system; that first-customer claim is company-attributed.

The supplied announcement says customers such as Cognition use the same operating model and tooling as their existing NVIDIA GB200 NVL72 and GB300 NVL72 fleets. This provides explicit operational context across system generations without establishing identical performance or hardware characteristics.

NVIDIA’s accompanying blog describes CoreWeave’s integration of NVIDIA compute, networking, and software into an AI-focused cloud. The supplied excerpts do not disclose the number of deployed NVL72 systems, Cognition’s contracted capacity, production-workload details, pricing, or measured performance results.

A separate company-attributed report says CoreWeave’s latest NVIDIA GPU supply was “essentially sold out” (基本售罄) and that it planned a development-tool suite called Forge for model training and optimization. Those are broader company statements; the source does not identify Vera Rubin-specific utilization or establish Forge’s general availability.