AI 시대, 최적의 GPU 인프라 선택 방안 - On-Premise vs Cloud

Learn about optimal GPU infrastructure selection strategies for the AI era in this 21-minute conference talk from SK AI SUMMIT 2024. Explore how GPUs have become a core competitive advantage in AI infrastructure and discover rational selection criteria for GPU deployment. Examine TCO minimization strategies based on NVIDIA's GPU release cycles and usage patterns across AI serving, fine-tuning, and foundation models. Gain insights into the unique advantages of GPUaaS (GPU-as-a-Service) being developed through SK Telecom's collaboration with Lambda, aimed at creating an accessible global service for customers.