Nvidia CEO Jensen Huang said the company has begun full-scale production of its next-generation AI infrastructure platform, Vera Rubin, marking a major step in Nvidia’s long-term data center roadmap. According to Huang, Vera Rubin delivers performance that is roughly five times higher than the current Blackwell platform, which is only now ramping across hyperscalers and enterprise customers.
The new architecture is designed to support increasingly complex AI workloads, including large-scale model training, reasoning systems, and advanced inference. Nvidia positions Vera Rubin as a foundation for the next wave of generative AI, robotics, and autonomous systems, where compute density and energy efficiency are critical constraints.
The announcement underscores Nvidia’s aggressive cadence in AI hardware, as demand from cloud providers, governments, and enterprises continues to outpace supply. The early production start suggests customers are already preparing deployments well ahead of broader availability.
