Microsoft Azure gets its first NVIDIA Blackwell “GB200 NVL72” system to run OpenAI, marking a major milestone in AI training & development.
Microsoft Gets Over 500 Blackwell GB200 Chips in Eight Racks of GB200 NVL72 for Azure OpenAI Service
OpenAI is now getting a huge boost in performance via NVIDIA’s GB200 NVL72 “Blackwell” system. While NVIDIA has already been powering OpenAI’s AI models, the expansion to Microsoft Azure to leverage the power of NVIDIA’s flagship data center GPU will further improve the AI models through rigorous training.
NVIDIA Blackwell is here!
Today, we’re excited to bring the highly-anticipated @nvidia Blackwell GPUs to Google Cloud with the preview of A4 VMs, powered by NVIDIA HGX B200.
Learn more → https://t.co/N1oqZMzJ8M pic.twitter.com/iBTLqUQSVg
— Google Cloud Tech (@GoogleCloudTech) January 31, 2025
In an X post, the CEO of Open AI, Sam Altman, announced that Microsoft Azure is now running the first full eight racks of the GB200 NVL72 system. Sam also thanked the CEO of Microsoft, Satya Nadella, and the CEO of NVIDIA, Jensen Huang, for making this happen. Through the collaboration of these companies, the Azure OpenAI service can now position Microsoft as one of the leading cloud providers for AI workloads.
first full 8-rack GB200 NVL72 now running in azure for openai—thank you @satyanadella and jensen!
— Sam Altman (@sama) January 31, 2025
The NVIDIA Blackwell GB200 NVL72 system contains 36 Grace CPUs and 72 Blackwell GPUs in a rack-scale design. With eight such racks, there will be a total of 288 Grace CPUs paired with a whopping 576 B200 GPUs. This can produce incredible computational power as a single rack can result in up to 6,480 TFLOPS of FP32 and up to 3,240 TFLOPS of FP64 performance. This computational power can now be multiplied by eight times, resulting in almost 51,840 TFLOPS of FP32 and 25,920 TFLOPS of FP64 performance.
With such exceptional computing performance, the Azure OpenAI service can be highly beneficial for enterprises looking to scale their businesses. The NVIDIA “Blackwell” GB200 NVL72 is incredibly powerful & optimized for large-scale generative AI workloads as it offers a high memory bandwidth of up to 576 TB/s and brings parallel processing power.
OpenAI currently uses a variety of NVIDIA GPUs, including NVIDIA V100s, H100s, and A100s, but the Blackwell B200 is the most advanced GPU chip that brings significant performance uplifts over its predecessors. As Microsoft invested almost $14 Billion in OpenAI, it’s no surprise that it wants to take the lead with the most powerful machine for its customers.
