
Jensen Huang took the stage at COMPUTEX Taipei on June 1, 2026 to unveil Nvidia's Vera Rubin architecture, a new AI superchip platform that delivers ten times the inference throughput of its predecessor, Blackwell. The NVL72 rack combines 72 Rubin GPUs with 36 Vera CPUs linked by NVLink 6, targeting large language model inference at data-center scale. Huang also previewed the Rubin Ultra and Feynman architectures on Nvidia's roadmap through 2028.

Entertainment · Jun 1 · L1-L4