Presented by Microsoft Events, this Ignite 2025 session examines how Microsoft Azure and NVIDIA accelerate physical AI and robotics development using simulation, digital twins, and advanced AI infrastructure.

Accelerate and Scale Physical AI Development with Simulation

Presented by Microsoft Events at Microsoft Ignite 2025

Session Overview

This session brings together experts from Wandelbots, Microsoft, and NVIDIA to discuss the latest advancements in physical AI for robotics. The presentation explores how simulation, powered by NVIDIA’s accelerated computing and the NVIDIA Omniverse platform, in conjunction with Microsoft Azure and the OpenUSD open standard, enable safer, faster, and more robust robotics development.

Key Topics Covered

  • Training Physical AI Models: Approaches for teaching AI-enabled robots to operate autonomously in real-world scenarios, leveraging high-fidelity simulations to minimize real-world risk.
  • NVIDIA’s Three-Computer Approach: How simulation, training, and deployment work together to deliver effective AI robotics solutions.
  • Microsoft-NVIDIA Collaboration: Integration of NVIDIA AI infrastructure and tools on Microsoft Azure for scalable development workflows.
  • Simulation in CI/CD: Using simulation as an integrated part of continuous integration and deployment pipelines, speeding up iteration and validation cycles for robotics.
  • Edge Deployment and Security with Azure ARC: Deploying AI models and robotics solutions at the edge, with security, using Azure ARC.
  • Digital Twins and Synthetic Data: Leveraging digital twin technology and synthetic data generation to improve model accuracy, speed up skills acquisition, and enhance robustness.
  • Industrial Adoption of Skills-Based Robotics: Trends and case studies in how industries are adopting advanced robotics and hybrid AI systems.
  • Open Collaboration and OpenUSD: The importance of open standards like OpenUSD in enabling interoperability across different AI and robotics frameworks.
  • Agentic AI and Unified Programming Environments: The rise of agent-based AI and the move toward unified programming paradigms for autonomous robotics and vehicles.

Why This Matters

Simulations and digital twins are transforming how organizations develop, test, and deploy physical AI solutions. By bridging advanced AI infrastructure with practical robotics, Microsoft Azure and NVIDIA are enabling faster time-to-value, improved safety, and streamlined continuous integration/deployment processes for a new era of industrial AI.

  • Teresa Conceicao
  • Ganesh Sivaraman

Resources

Session Details

  • Session ID: BRKSP489
  • English (US)
  • Level: Intermediate (200)

Chapters

0:00 - Discussion on training physical AI models for autonomy in real-world tasks
8:22 - NVIDIA’s three-computer approach: simulation, training, deployment
9:37 - NVIDIA and Microsoft: Joint AI infrastructure on Azure
11:56 - Simulation’s role in CI/CD for robotics
20:09 - Using Azure ARC for edge deployment and security
23:35 - Industrial adoption of skill-based and hybrid robotics
29:05 - Demos: Digital twins, open standards, and collaboration
30:11 - Agentic AI and unified programming environments for robotics and vehicles

Key Takeaways

  • Cloud-powered simulation and digital twins reduce risk, speed learning, and ensure robustness in robotics.
  • Microsoft Azure and NVIDIA Omniverse offer a platform for scalable, secure, and collaborative robotics development.
  • Open standards such as OpenUSD foster interoperability and innovation across the robotics ecosystem.

For more similar sessions and resources, visit Microsoft Ignite.