NVIDIA says CoreWeave has put Vera Rubin NVL72 into production, with token throughput increased by 4.8 times in testing
8 hour ago / Read about 0 minute
Author:小编   

NVIDIA announced that AI cloud service provider CoreWeave has put Vera Rubin NVL72 into production and begun offering related cloud services to customers. AI software company Cognition has become the first customer to run production workloads on this platform. According to Cognition's early testing, Vera Rubin NVL72 can achieve a maximum total token throughput up to 4.8 times that of GB200 NVL72 in real-world software engineering AI inference tasks. CoreWeave's testing also showed that the startup speed of AI agent sandboxes equipped with NVIDIA's Vera CPU has increased by more than three times.