CoreWeave Puts NVIDIA Vera Rubin NVL72 Into Production: Cognition Reports 4.8x Over GB200, Vera CPU Racks and Forge Follow
CoreWeave puts NVIDIA Vera Rubin NVL72 into production, with Cognition running live workloads as its first customer. Cognition reports up to 4.8x the total token throughput per GPU compared with a GB200 NVL72 baseline on SWE-2 inference at matched interactivity.
Putting NVIDIA Vera Rubin NVL72 into production marks a major deployment of new AI compute hardware. The reported throughput comparison gives infrastructure teams a concrete performance figure to assess.
