Presented by Microsoft
Microsoft's recent announcements regarding new collaborations with long-time partner Nvidia positioned the company at the center of this year’s Nvidia GTC AI conference, held in San Jose from March 18 to 21.
This week’s AI innovation updates covered a wide range of topics, from advancements in AI infrastructure and services to new platform integrations and industry breakthroughs. Notably, Nidhi Chappell, VP of Azure Generative AI and HPC Platform at Microsoft, engaged in an exclusive conversation with Senior Writer Sharon Goldman to discuss the partnership between Microsoft, OpenAI, and Nvidia, as well as insights into the future of the market.
"Partnership is at the core of everything we do," Chappell noted. "When training large foundational models, you need scalable infrastructure that runs reliably over time. We have invested significant effort with Nvidia to ensure that our Azure OpenAI service allows enterprise customers to seamlessly integrate into their current workflows or begin new projects with our tools."
Watch the full interview below: Live from GTC: A Conversation with Microsoft | NVIDIA On-Demand. For a closer look at key conference announcements, explore Microsoft’s extensive series of panels and talks, all available for free viewing on demand.
AI Infrastructure Upgrades Through New Integrations
As workloads become more sophisticated, Microsoft is addressing growing demands with hardware innovations. Notably, Microsoft is among the first to utilize the Nvidia GB200 Grace Blackwell Superchip and the Nvidia Quantum-X800 InfiniBand networking, integrating these advancements into Azure. Additionally, the new Azure NC H100 v5 VM series is now accessible for organizations of all sizes.
The Nvidia GB200 Grace Blackwell Superchip is engineered to manage complex AI workloads and data processing, with new Azure instances enhancing performance for foundational models in natural language processing, computer vision, and speech recognition. It offers up to 16 TB/s of memory bandwidth and up to 30 times better inference for trillion-parameter models compared to the previous generation. The Nvidia Quantum-X800 InfiniBand networking further amplifies parallel computing at massive GPU scales.
Discover more about Nvidia and Microsoft integrations here.
The Azure NC H100 v5 VM series is designed for mid-range training, inference, and high-performance computing (HPC) simulations, based on the Nvidia H100 NVL platform. It features one or two Nvidia H100 94GB PCIe Tensor Core GPUs connected by NVLink, providing 600 GB/s bandwidth and 128GB/s bi-directional communication to minimize data transfer latency, enhancing the efficiency of AI and HPC applications. With Nvidia multi-instance GPU (MIG) technology, customers can partition each GPU into up to seven instances.
Learn about current customer achievements.
Transformative Advances in Healthcare and Life Sciences
AI has catalyzed rapid innovations in medicine and life sciences, impacting research, drug discovery, and patient care. Microsoft Azure's expanded collaboration with Nvidia DGX Cloud and the Nvidia Clara suite enables healthcare providers, pharmaceutical companies, and medical device developers to accelerate clinical research and patient care innovations.
Organizations leveraging cloud computing and AI include Sanofi, the Broad Institute of MIT and Harvard, Flywheel, and Sophia Genetics, alongside academic medical centers like the University of Wisconsin School of Medicine and Public Health. They are effecting transformative changes in healthcare, enhancing patient care, and democratizing AI for healthcare professionals.
Learn how AI is revolutionizing the healthcare industry.
Industrial Digital Twins Gain Traction with Omniverse APIs on Azure
Nvidia Omniverse Cloud APIs are now integrated into Microsoft Azure, extending the platform's capabilities. Developers can embed core Omniverse technologies into existing design and automation applications for digital twins, aiding the testing and validation of autonomous systems such as robots and self-driving vehicles.
During his GTC keynote, Nvidia CEO Jensen Huang showcased Teamcenter X connected to Omniverse APIs, allowing the software to link design data with Nvidia generative AI APIs and utilize Omniverse RTX rendering directly within the application.
Explore how organizations are using Omniverse Cloud APIs in Azure.
Enhancing Real-Time Contextual Intelligence
Copilot for Microsoft 365, set to be available as a dedicated keyboard key on Windows 11 PCs, merges the capabilities of large language models with proprietary enterprise data. Powered by Nvidia GPUs and the Triton Inference Server, it enables real-time contextual intelligence, enhancing user creativity, productivity, and skills.
Accelerating AI Deployment
Nvidia NIM inference microservices, part of the Nvidia AI Enterprise software platform, deliver cloud-native microservices optimized for over two dozen popular foundation models. These prebuilt, run-anywhere containers, powered by Nvidia AI Enterprise inference software—including Triton Inference Server, TensorRT, and TensorRT-LLM—facilitate faster market deployment of high-performance AI applications.
Deeper Integration of Nvidia DGX Cloud with Microsoft Fabric
Microsoft and Nvidia are enhancing the integration of Microsoft Fabric, the all-in-one analytics solution, with Nvidia DGX Cloud compute. This enhancement allows Nvidia’s optimized runtimes, LLMs, and machine learning to work cohesively with Microsoft Fabric. With Fabric OneLake as the underlying storage, developers can effectively tackle data-intensive challenges, such as digital twins and weather forecasting. This collaboration also enables DGX Cloud to enhance Fabric data science and data engineering processes.
Catch up on what you missed at GTC 2024
Microsoft showcased the significant potential of its collaborations with Nvidia, affirming Azure's crucial role in crafting successful AI strategies for organizations of all sizes. Watch all of Microsoft’s panels and talks here, available for on-demand streaming.
Learn more about Microsoft and NVIDIA AI solutions:
- Azure AI Portfolio
- Azure AI Infrastructure
- NVIDIA GPU-Accelerated Computing on Microsoft Azure