Credit: Amazon

Amazon Web Services (AWS) has announced two major enhancements for Amazon S3 Tables: a new Intelligent-Tiering storage class that automatically optimizes data storage costs, and built-in replication support to maintain synchronized Apache Iceberg table replicas across Regions and accounts. Both features are available today in all Regions where S3 Tables are supported.

Automatic Cost Optimization with Intelligent-Tiering

Organizations working with large Iceberg-based datasets often struggle with manually managing storage cost as data ages. The new S3 Tables Intelligent-Tiering storage class addresses this by automatically placing files into the most cost-efficient access tier based on observed access patterns.

Three Low-Latency Tiers

• Frequent Access – Ideal for active datasets

• Infrequent Access – ~40% cheaper than Frequent Access

• Archive Instant Access – ~68% cheaper than Infrequent Access

How It Works

• After 30 days of no access, files move to Infrequent Access

• After 90 days of no access, they shift to Archive Instant Access

• Transitioning occurs automatically, with no changes required in applications

• Table maintenance tasks (snapshot expiration, compaction, cleanup) operate normally without affecting tier transitions

Compaction is optimized for cost: it only processes data in the Frequent Access tier, avoiding expensive reads from colder tiers.

Configuration

• Existing tables continue using Standard storage by default

• New tables can specify Intelligent-Tiering or inherit the storage class from the bucket default

• Admins can set Intelligent-Tiering as the bucket-level default for all new S3 Tables

Cross-Region and Cross-Account Replication for S3 Tables

AWS also introduced native replication for Amazon S3 Tables—removing the need to manually build complex synchronization systems for Iceberg metadata and data files.

Key Capabilities

• Create read-only replica tables in another Region or account

• Maintain snapshot lineage and chronological update order

• Automatically replicate table updates and object changes

• Enable global read access for distributed teams

• Support compliance, disaster recovery, and multi-Region architectures

This simplifies constructing global data lakes and ensures consistent, up-to-date Iceberg tables regardless of geography.

Pricing Details

Intelligent-Tiering

• No additional overhead to enable

• Pay only for storage per tier

• Monitoring via AWS CUR, CloudWatch, and S3 metrics

Replication

Charges include:

• Storage in destination tables

• Replication PUT requests

• Table commits

• Object monitoring for replicated files

• Inter-Region data transfer fees (when applicable)

Customers can refer to the Amazon S3 pricing page for full details.

Availability

Both new features—Intelligent-Tiering storage class and S3 Tables replication—are available today in all AWS Regions supporting Amazon S3 Tables.