ABCS25: The AI Evolution: How Emerging AI Trends are Transforming the Cloud Services by Marc Stampfli

The rapid evolution of Artificial Intelligence (AI), from Generative AI to autonomous AI Agents, is fundamentally reshaping cloud computing paradigms. This session explores how the latest developments in AI—particularly the emergence of intelligent, autonomous AI agents—are driving transformative changes in cloud architectures, deployment models, and service delivery. With a specific focus on Azure cloud environments, we examine how developers and technology leaders can leverage new AI capabilities to unlock significant business opportunities, enhance operational efficiency, and accelerate innovation. We will discuss key technological breakthroughs, practical implications, strategic opportunities, and critical challenges, offering actionable insights and forward-looking guidance tailored specifically for cloud developers and architects.
🙂 MARC STAMPFLI ⚡️ Business Director Switzerland @ NVIDIA

Transcript

1. The AI Evolution: How Emerging AI Trends are Transforming the Cloud Services Marc Stampfli, Business Director Switzerland E-Mail: mstampfli@nvidia.com Linkedin: https://www.linkedin.com/in/marcstampfli
2. “20 years ago, all of this [artificial intelligence] was science fiction. 10 years ago, it was a dream. Today, we are living it.“ — Keynote CES 2025, Jensen Huang, Co-founder and CEO NVIDIA, 6th of January 2025 LIVING THE INDUSTRIAL REVOLUTION
3. Widespread Societal Transformation: Industrial revolutions shift societies from one dominant mode of production to another (e.g., from agrarian to industrial), altering how people work, live, and interact. Technological Foundations: Each revolution is anchored in core technologies (e.g., steam engines, electricity, computers) that enable new industries and ways of organizing work. Economic and Social Change: These revolutions lead to new industries, urbanization, changes in labor, and shifts in wealth and power. Long-Term, Measurable Impact: The changes are enduring, global in reach, and measurable in terms of productivity, economic growth, and societal structure. WHAT QUALIFIES FOR INDUSTRIAL REVOLUTION
4. ©SEGA 1993 NVIDIA BEGINS 1999 INVENTED GPU
5. 2006 INVENTED CUDA
6. EVOLUTION OF ARTIFICIAL INTELLIGENCE
7. PERCEPTION AI HAND CODING INSTRUCTIONS CPU SOFTWARE MACHINE LEARNING NEURAL NETWORKS GPU AI
8. Full-Stack Autonomous Driving Platform
9. FROM PERCEPTION TO GENERATION
10. What make generative AI different? •Paradigm shift from instruction-based computing to intention-based computing. •Generate completely original artifacts (tokens) that are like real content.
11. Sparking the iPhone Moment of AI.
12. Generative AI Is Transforming Life Sciences and the Pharmaceutical Industry
13. A New Computing Platform Sparks New Industrial Revolution
14. INTELLIGENT AGENTS EVERYWHERE
15. Calculator Database Search Semantic Search Web Search Generate Podcast User or Machine Tool Use Calculate Search Generate Prompt Response Critique Plan Calculate Search Generate Mechanics of an AI Agent
18. NEXT EVOLUTION IS PHYSICAL
19. “Pick up the piece of toast from the toaster and place it on the white plate to the right of the toaster.” Text Tokens Image Tokens Physical AI Model Action Tokens What is Physical AI?
21. NVIDIA AI ENTERPRISE – THE OPERATING SYSTEM FOR AI
22. Cluster-Scale Software System Software Chip Software DSL DSL DSL ... CUDA • DOCA • NCCL DSL Pioneering Accelerated Computing Modern computing started sixty years ago with the IBM System/360. For the last decade—as the performance scaling of general-purpose processing has slowed while computing demand has continued— an exponentially growing performance- to-demand deficit has built up. NVIDIA accelerated computing has created a path forward at just the right time. Accelerated computing starts with the most advanced processors and ends with AI factories. From chip architecture to advanced networking to acceleration libraries, NVIDIA builds the entire computing system at data center scale to produce intelligence at scale.
23. NVIDIA H100, H200 NVIDIA InfiniBand Networking Azure AI Foundry NVIDIA NIM Azure Kubernetes Service NVIDIA AI Enterprise NVIDIA GB200 NVL72 NVIDIA AI Enterprise NVIDIA NIM NVIDIA Omniverse Cloud NVIDIA DGX Cloud Azure OpenAI Service NVIDIA H200 Azure Confidential Computing NVIDIA H100 Azure Machine Learning NVIDIA NeMo Enterprises Azure Databricks NVIDIA RAPIDS Azure Container Apps NVIDIA A100 Azure Maps NVIDIA cuOpt Microsoft Bing NVIDIA CV-CUDA Microsoft 365 Copilot NVIDIA Triton Microsoft Teams NVIDIA Triton Windows 365 NVIDIA RTX vWS GSIs and SDPs Startups ISVs Confidential Computing NVIDIA Platforms in Azure Bringing NVIDIA Ecosystem to Azure NVIDIA Powers and Accelerates Azure
24. 3 things to know about NVIDIA We are an ISV We deliver a software platform for developers, ISVs and Integrators to build AI enabled solutions and digital twins. We sell via marketplace Two software platforms: • NVIDIA AI Enterprise • NVIDIA Omniverse Available in marketplace. We address complex AI and Digital Twins problems AI applications Vertical models, optimized models and blueprints… Industrial Metaverse 3D pipeline integration, physical simulation, IOT, robotics…
25. Industrial Metaverse What we can offer together • Layout Planning • Process Simulation • Remote Monitoring • Industrial Inspection • Robotics Simulation • Industrial Co-Pilot AI at scale • Preoptimized models and blueprints • Enterprise support • Full control of AI stack • Predictable costs Vertical Solutions • Healthcare: drug discovery, medical imagery,… • FSI: Fraud detection, customer service, trading, insurance risk evalution… • Smart Space: computer vision, simulation,… • HPC, Retail, AEC, Transport, Energy,…
26. Azure Tools AI Foundry, AKS, Container Apps, Azure Databricks,Fabric, Azure IOT Azure Accelerated Infrastructure Running the NVIDIA AI Enterprise accelerated platform on Azure Solving the problem that cannot be solved out-of-the-box in Azure Azure differentiators • InfiniBand networking • Confidential computing on H100 • H100 NVL GPU • vGPU on A10 • Omniverse Cloud Software available as licenses or managed service • NVIDIA AI Enterprise and DGX Cloud • NVIDIA Omniverse Enterprise and Omniverse Cloud Available in Azure Marketplace • MACC eligible • Sold via marketplace GB200 NVL72 H100/H200 H100 NVL RTX Pro Blackwell A10 Workloads CPU + GPU Scale up for AI Factories AI Training, large inference and HPC Generative AI inference Inference, workstations, rendering Coming Soon Inference, workstations, rendering Fast, low latency, InfiniBand with NVIDIA Networking NVIDIA AI NVIDIA Omniverse NVIDIA Software Platform Managed or Unmanaged
27. NVIDIA NIM: Optimized AI Models Run Up to 5X Faster Community models – partner models – NVIDIA models NVIDIA INFERENCE MICROSERVICE Computer Vision Biology Vision Language RAG Speech Simulation Regional Language Digital Human Language Pre-Trained AI Models Packaged and Optimized to Run Across CUDA Installed Base https://build.nvidia.com
28. NVIDIA NIM Optimized Inference Microservices Accelerated runtime for generative AI Easy to Use Speed time to market with prebuilt, continuously maintained microservices Performance Optimize throughput to maximize token generation for lower TCO and ROI Enterprise Supported Gain confidence with API stability, security patching, quality assurance and support Portable Deploy anywhere with security and control of AI applications and data Optimized inference engines NVIDIA NIM Domain specific code Support for custom models Industry standard APIs Prebuilt container and Helm chart
29. The ChatGPT Moment for Humanoid Robotics is coming.