NVIDIA touts Vera Rubin NVL72 up to 30x throughput vs GB300
NVIDIA said its Vera Rubin NVL72 platform delivers up to 30x higher throughput per megawatt than the GB300 NVL72 in agentic AI workloads, cutting cost per million tokens by up to 35x.
NVDA Vera Rubin Delivers 30x Higher AI Throughput vs.
GB300 ⠀ • NVIDIA says its Vera Rubin NVL72 platform delivers up to 30x higher throughput per megawatt than the GB300 NVL72 in agentic AI workloads, while reducing cost per million tokens by up to 35x. • Taiwanese AI server manufacturers Foxconn, Quanta and Wistron are preparing for Vera Rubin mass production and shipments in Q4, with order visibility reportedly extending into 2028. • Foxconn says GB300 and Vera Rubin production will run in parallel during the transition, while Quanta and Wistron have already entered mass-production transition phases for the next-generation platform.