Wedoany.com Report on Mar 16th, Microsoft CEO Satya Nadella announced on the social media platform X that the company has initiated the validation of NVIDIA's Vera Rubin NVL72 system in its cloud infrastructure. Nadella stated that Azure is the first cloud platform to validate this system, marking another significant step in advancing next-generation AI infrastructure in collaboration with NVIDIA.

The Rubin platform is the successor to NVIDIA's Blackwell architecture. The NVL72 configuration integrates 72 Rubin GPUs into a rack-scale AI system, connected via the next-generation NVLink architecture and high-capacity, high-bandwidth memory. The system operates as a single logical accelerator, designed to support the training and inference of large-scale AI models. Compared to the current GB200 NVL72 generation, Rubin is expected to offer significant improvements in computational density, memory bandwidth, and interconnect performance.

Microsoft's validation work indicates that hyperscale cloud service providers are already preparing infrastructure for Rubin-class AI clusters. These platforms target multi-rack AI systems scalable to tens of thousands or even hundreds of thousands of GPUs, requiring new optical interconnects, power delivery, and liquid cooling solutions. Azure has historically been an early deployment partner for NVIDIA GPU generations, such as H100 and Blackwell.

This confirmation comes during a busy week for the AI infrastructure ecosystem. NVIDIA's GTC conference is underway in San Jose, while the optical networking industry is attending the Optical Fiber Communication Conference (OFC) in Los Angeles. This parallel scheduling highlights the increasingly close interdependence between AI computing platforms and high-speed networking technologies.

"We are the first cloud to validate the NVIDIA Vera Rubin NVL72 system, another big step with NVIDIA in building next-generation AI infrastructure," Nadella wrote on X.

Hyperscale cloud providers typically validate next-generation GPU systems months before commercial deployment to redesign data center architecture, power systems, and optical interconnect architecture. Microsoft's early Rubin validation suggests the industry's transition beyond Blackwell is already underway. The technological advancements discussed this week at GTC and OFC, including expanded optical interconnects, co-packaged optics, and 224G networking, directly support the infrastructure needed to interconnect Rubin-class AI clusters.