CoreWeave Puts NVIDIA Vera Rubin NVL72 Into Production: Cognition Reports 4.8x Over GB200, Vera CPU Racks and Forge Follow
CoreWeave used its Fully Connected 2026 conference in San Francisco to put NVIDIA Vera Rubin NVL72 into production, with Cognition as the first customer running live workloads on it. Cognition’s engineers report up to 4.8x the total token throughput per GPU of a GB200 NVL72 baseline on SWE-2 inference at matched interactivity, and 3.8x the The post CoreWeave Puts NVIDIA Vera Rubin NVL72 Into Production: Cognition Reports 4.8x Over GB200, Vera CPU Racks and Forge Follow appeared first on StorageReview.com.
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.