Microsoft (MSFT.US)Azure is racing ahead, becoming the world's first cloud service provider to begin testing NVIDIA Vera Rubin NVL72 system.

robot
Abstract generation in progress

Microsoft(MSFT.US) CEO Nadella said the company’s Azure has become the first cloud services provider to begin validating NVIDIA(NVDA.US) Vera Rubin NVL72 systems.

On Friday afternoon, Nadella posted on the X platform: “We are the first cloud platform to introduce NVIDIA Vera Rubin NVL72 systems for validation—another major leap forward in building next-generation AI infrastructure, together with NVIDIA.”

Microsoft’s long-standing partnership with NVIDIA enables it to deploy continuously iterating hardware devices more effectively in its data centers.

Rani Borkar, president of Microsoft Azure Hardware Systems, said, “Microsoft has years of market-validated experience in designing and deploying scalable AI infrastructure—an infrastructure that evolves with every major advancement in AI technology,” and “By synchronizing with each generation of NVIDIA’s accelerated computing infrastructure, Microsoft quickly integrates NVIDIA’s innovations and delivers them at scale.”

Borkar added: “The NVIDIA Rubin platform marks an important step forward for accelerated computing. From the outset, Azure’s AI data centers and superfactories are designed to fully leverage this advantage. Over the years, we have co-designed with NVIDIA interconnects, memory systems, thermal management, packaging, and rack-level architectures—meaning Rubin can be integrated directly into the Azure platform without redesign.”

Although Microsoft says it is the first company to introduce Vera Rubin for validation, multiple cloud providers are also expected to deploy the technology in 2026. This includes Amazon(AMZN.US) AWS, CoreWeave(CRWV.US), Google(GOOGL.US), Nebius(NBIS.US), and Oracle(ORCL.US).

According to NVIDIA, each Vera Rubin NVL72 rack can deliver up to 3.6 exaflops of performance—about five times the performance of the GB200-based systems. Each rack connects 72 GPUs and 36 CPUs via sixth-generation NVLink, delivering 260TB of bandwidth per second.

View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • Comment
  • Repost
  • Share
Comment
Add a comment
Add a comment
No comments
  • Pin