AWS Certified AI Practitioner AIF-C01 Practice Question

Each Region that hosts a SageMaker model must run its own training jobs and dedicated inference instances. Deploying the endpoint in a second Region therefore duplicates both the fine-tuning (if repeated) and the ongoing hosting costs for the additional fleet of instances. AWS does not automatically replicate or discount these resources across Regions, and inter-Region data transfer charges are separate considerations. Automatic, no-cost replication or multi-Region volume discounts are not provided, so the primary incremental expense is paying for the same training and endpoint capacity twice.