NVIDIA Investor Day Presentation Deck

INFERENCE IS HARD Need for Optimized Al Infrastructure Model Complexity High Throughput Real-Time Requirements Many Use Cases INFERENCE PLATFORM Amazon SageMaker K LEADERSHIP End-to-End Acceleration Amazon Elastic Container Service (ECS) HTTP || Azure Machine Learning Hypercale Scheduler Model Compilars TRITON INFERENCE SERVER JL Google Vertex Al Amazon Elastic Container Azure Kubernetes Service (ECS) Service (AKS) →阿里云 Alibaba Cloud. PAI-EAS M Google Kubernetes Engine (GKE) NVIDIA AI INFERENCE IN FULL THROTTLE FY19 RAPID SHIFT TO GPU 9X in Just 3 Years Il FY20 FY21 CSP Inference Revenue FY22

View entire presentation