Microsoft Azure has moved ahead of its cloud rivals, becoming the first provider to begin validating Nvidia’s Vera Rubin NVL72 system. CEO Satya Nadella made the announcement Friday afternoon in a post on X, calling it “another big step in building the next generation of AI infrastructure.”
The Vera Rubin NVL72 is a rack-scale platform that packs 72 Rubin GPUs and 36 custom Arm-based Vera CPUs into a single system. The GPUs are connected via sixth-generation NVLink fabric, delivering 260 terabytes per second of bandwidth.
Microsoft Corporation, MSFT
That is a serious jump in raw compute. Each NVL72 rack is capable of up to 3.6 exaflops of performance — roughly five times more than the previous GB200-based systems it is set to replace.
The key phrase here is “co-design.” Microsoft says it has been working with Nvidia across interconnects, memory systems, thermals, packaging, and rack-scale architecture for years. That means the Rubin hardware slots directly into Azure’s existing infrastructure — no rework needed.
Microsoft’s Azure data centers, including facilities in Wisconsin and Atlanta, were specifically engineered to handle the power density and liquid-cooling demands of NVL72 racks. This kind of infrastructure planning doesn’t happen overnight.
The company had to redesign power and liquid-cooling systems across multiple sites to handle the increasing watt density these new racks bring. That investment is now paying off with validated hardware others are still waiting to spin up.
A BlackRock-led consortium, backed by Microsoft and Nvidia, recently moved to acquire Aligned Data Centers in a $40 billion deal aimed at expanding global capacity ahead of this next hardware wave.
Microsoft being first doesn’t mean it’s alone for long. Amazon Web Services, Google, CoreWeave, Nebius, and Oracle are all expected to deploy Vera Rubin systems — most of them in the second half of 2026.
Bernstein analysts have pointed to Microsoft’s “first-to-validate” position as a reflection of its broader cloud and SaaS efficiency gains, which they describe using a “Rule of 37.3%” outperformance metric.
MSFT was down 1.57% and NVDA fell 1.58% on the day the news was published, in line with broader market softness rather than any negative reaction to the announcement.
Rubin Ultra, the next iteration of the platform, is expected to follow in 2027.
The post Microsoft (MSFT) Becomes First Cloud Provider to Validate Nvidia’s Most Powerful AI Chip appeared first on CoinCentral.


A new discussion emerging within the global Pi Network community is shedding light on how pioneers may eventually