D&A leaders scaling generative AI training and inference must navigate new technologies including GPUs and AI processors, both on-premises and in the cloud. In this session, we provide a pragmatic framework to map GenAI model sizes to infrastructures across on-premises and cloud environments.