Digital Event Horizon
Microsoft Azure has unveiled what it claims to be the world's first production NVIDIA GB300 NVL72 cluster, a supercomputer-scale system designed for the most demanding AI inference workloads. The new system features over 4,600 NVIDIA Blackwell Ultra GPUs and is designed to provide unparalleled performance for large-scale training and inference tasks. According to Microsoft Azure, this achievement marks a major milestone in building the infrastructure that will unlock the future of artificial intelligence.
Microsoft Azure has unveiled a supercomputer-scale system built for AI inference workloads, called NVIDIA GB300 NVL72. The system features over 4,600 GPUs and delivers record-setting performance using NVFP4. The cluster provides unparalleled performance for large-scale training and inference tasks, with leadership performance on new benchmarks. The NVIDIA GB300 NVL72 is a liquid-cooled, rack-scale powerhouse that integrates GPUs and CPUs into a single unit, providing massive memory space. The system uses a two-tiered NVIDIA networking architecture for scale-up and scale-out performance. Microsoft's collaboration with NVIDIA has resulted in the first at-scale production of this technology, showcasing their commitment to AI infrastructure.
Microsoft Azure has recently made headlines by unveiling what it claims to be the world's first production NVIDIA GB300 NVL72 cluster, a supercomputer-scale system purpose-built for the most demanding AI inference workloads. This groundbreaking achievement marks a major milestone in building the infrastructure that will unlock the future of artificial intelligence.
The new system, which features over 4,600 NVIDIA Blackwell Ultra GPUs connected via the NVIDIA Quantum-X800 InfiniBand networking platform, is designed to provide unparalleled performance for large-scale training and inference tasks. According to Microsoft Azure, this supercomputer-scale cluster delivers record-setting performance using NVFP4 and achieves leadership performance on all newly introduced benchmarks like the Llama 3.1 405B model.
The NVIDIA GB300 NVL72 system at the heart of the new cluster is a liquid-cooled, rack-scale powerhouse that integrates 72 NVIDIA Blackwell Ultra GPUs and 36 NVIDIA Grace CPUs into a single, cohesive unit. This design provides a staggering 37 terabytes of fast memory and 1.44 exaflops of FP4 Tensor Core performance per VM, creating a massive, unified memory space essential for reasoning models, agentic AI systems, and complex multimodal generative AI.
To connect the system's components, Microsoft Azure relies on a two-tiered NVIDIA networking architecture designed for both scale-up performance within the rack and scale-out performance across the entire cluster. Within each GB300 NVL72 rack, the fifth-generation NVIDIA NVLink Switch fabric provides 130 TB/s of direct, all-to-all bandwidth between the 72 Blackwell Ultra GPUs, transforming the entire rack into a single, unified accelerator with a shared memory pool.
To scale beyond the rack, the cluster uses the NVIDIA Quantum-X800 InfiniBand platform, purpose-built for trillion-parameter-scale AI. Featuring NVIDIA ConnectX-8 SuperNICs and Quantum-X800 switches, this platform provides 800 Gb/s of bandwidth per GPU, ensuring seamless communication across all 4,608 GPUs.
This achievement is the result of years of deep partnership between NVIDIA and Microsoft, who have collaborated to build AI infrastructure for the world's most demanding AI workloads. The new system marks another leadership moment, ensuring that leading-edge AI drives innovation in the United States.
"Nidhi Chappell, corporate vice president of Microsoft Azure AI Infrastructure, stated, "Delivering the industry’s first at-scale NVIDIA GB300 NVL72 production cluster for frontier AI is an achievement that goes beyond powerful silicon — it reflects Microsoft Azure and NVIDIA’s shared commitment to optimize all parts of the modern AI data center." She further emphasized that their collaboration helps ensure customers like OpenAI can deploy next-generation infrastructure at unprecedented scale and speed."
Microsoft's use of custom liquid cooling and power distribution, as well as a reengineered software stack for orchestration and storage, showcases the company's dedication to delivering cutting-edge solutions.
The news was shared in an official press release on Microsoft Azure’s blog. For those interested in learning more about this announcement, it is recommended that they visit the link provided in the original article.
Related Information:
https://www.digitaleventhorizon.com/articles/Microsoft-Azure-Unveils-Worlds-First-NVIDIA-GB300-NVL72-Supercomputing-Cluster-for-OpenAI-deh.shtml
https://blogs.nvidia.com/blog/microsoft-azure-worlds-first-gb300-nvl72-supercomputing-cluster-openai/
https://azure.microsoft.com/en-us/blog/microsoft-azure-delivers-the-first-large-scale-cluster-with-nvidia-gb300-nvl72-for-openai-workloads/
Published: Wed Oct 15 02:32:29 2025 by llama3.2 3B Q4_K_M
