NVIDIA announced $215.9 billion full-year revenue for 2025, with $68.1 billion generated in the fourth quarter. The company confirmed that its first Vera Rubin VR200 sample racks are now shipping to customers, with volume production expected to begin in the second half of 2026. This marks a significant milestone in NVIDIA's transition to next-generation AI infrastructure. The Vera platform integrates the Vera CPU, Rubin GPU, NVLink 6 switch, and other specialized components designed to power trillion-parameter models with unprecedented efficiency.

Compared to NVIDIA's previous-generation Blackwell systems, Vera Rubin will deliver 10x lower inference costs while requiring only one-fourth the number of GPUs. This represents a major leap in AI compute density and cost-effectiveness. NVIDIA CFO Colette Kress emphasized the platform's modular, cable-free design improves serviceability and resilience, stating every cloud provider will deploy Vera Rubin. The shift signals a broader industry pivot toward more efficient AI architectures.

The Vera Rubin deployment underscores NVIDIA's strategy to dominate both high-performance computing and generative AI markets. While specific customer names weren't disclosed, the announcement confirms NVIDIA's timeline for transitioning its flagship AI platforms. The VR200 samples represent the first tangible step toward widespread adoption of this next-generation architecture, with implications for everything from cloud computing to enterprise AI deployments.