BioGenix stores multiple terabytes of raw and preprocessed training datasets in an S3 bucket. Datasets are frequently used during active development but become infrequently accessed after 90 days. The team wants a low‑management, cost‑optimized storage strategy that minimizes retrieval surprises when occasional re-training happens. Which S3 configuration is the best fit?

S3 Intelligent-Tiering is the best fit because it automatically moves objects between access tiers based on changing access patterns, optimizing costs without performance impact or retrieval fees. This is ideal for datasets with fluctuating access. Adding a lifecycle rule to transition to Glacier Deep Archive after a long retention period (e.g., 365 days) provides further cost savings for archival data that is rarely accessed, while still allowing for occasional, cost-effective retrieval. Configuring a lifecycle rule to move all objects to Standard-Infrequent Access after 30 days is less optimal as it doesn't account for potential re-access within that 90-day window, leading to retrieval costs. S3 One Zone-IA is not suitable as it lacks resilience (data is stored in a single Availability Zone) and doesn't offer the same cost optimization for varying access patterns as Intelligent-Tiering. Discarding Intelligent-Tiering is incorrect; it is designed to detect access patterns and automate tiering.