Class Central is learner-supported. When you buy through links on our site, we may earn an affiliate commission.

Microsoft

Power Next-Generation AI Workloads with NVIDIA Blackwell on Azure - BRKSP490

Microsoft via YouTube

Overview

Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore this 44-minute conference talk from Microsoft Ignite 2025 that delves into the revolutionary NVIDIA Blackwell architecture and its integration with Microsoft Azure for powering next-generation AI workloads. Discover how the era of agentic and physical AI is transforming computing through groundbreaking advancements in reasoning, multimodal inference, robotics simulation, and visual AI agents. Learn about the purpose-built collaboration between NVIDIA Blackwell Ultra and NVIDIA RTX PRO 6000 Blackwell technologies that tackle diverse AI use cases while leveraging Azure's security and flexibility. Examine AI factories and the fine-tuning lifecycle within Azure's collaborative environment, understand how to build AI super factories using Microsoft's cloud infrastructure, and explore multi-dimensional AI design frameworks. Dive into extreme co-design principles and Mixture of Experts (MoE) model communication strategies, analyze NVLink acceleration and Pareto curve performance optimization techniques, and investigate confidential computing capabilities on H100 GPUs for secure execution environments. Discover the RTX Pro 6000 Blackwell Server Edition's capabilities for AI and visualization workloads, explore the Omniverse platform's role in enabling physical AI simulation, and examine real-world use cases in manufacturing and warehouse optimization. Gain insights into how these technologies deliver smarter, faster, and more efficient AI solutions for enterprise applications.

Syllabus

0:00 - AI Factories and the Fine‑Tuning Lifecycle with Azure Collaboration
00:08:48 - Building AI Super Factories with Microsoft Azure
00:11:22 - Introducing the Acronym and Framework for Multi-Dimensional AI Design
00:13:54 - Extreme Co-Design and Mixture of Experts MoE Model Communication
00:14:56 - NVLink Acceleration and Pareto Curve Analysis for Performance Gains
00:21:57 - Confidential Computing on H100 GPU for Secure Execution
00:22:11 - Introduction of RTX Pro 6000 Blackwell Server Edition for AI and Visualization Workloads
00:33:17 - Omniverse platform enabling simulation for physical AI
00:34:29 - Omniverse use cases: manufacturing and warehouse optimization

Taught by

Microsoft Ignite

Reviews

Start your review of Power Next-Generation AI Workloads with NVIDIA Blackwell on Azure - BRKSP490

Never Stop Learning.

Get personalized course recommendations, track subjects and courses with reminders, and more.

Someone learning on their laptop while sitting on the floor.