Kinesis Data Streams adds streaming tables for Apache Iceberg on S3 Tables

AWS added streaming tables to Amazon Kinesis Data Streams, a fully managed feature that continuously lands streaming data as queryable Apache Iceberg tables directly on Amazon S3 Tables. The pitch is operational simplicity plus cost: built-in inline compaction reduces delivery costs by up to 25% and query costs by up to 30%, and there is no infrastructure for teams to provision or maintain.

The mechanism addresses a persistent data-engineering pain point — bridging real-time streams and analytical lakehouse tables usually requires a separate ingestion, compaction and file-management pipeline. By making Iceberg tables a native Kinesis delivery target with automatic compaction, AWS collapses that pipeline into a managed feature, which matters as Iceberg becomes the de facto open table format (Amazon Redshift also added Iceberg v3 support this cycle).

Strategically it deepens AWS's lakehouse story and its bet on open table formats as the neutral substrate for analytics and, increasingly, for feeding AI systems that query fresh operational data. Lower query and delivery costs make continuous streaming-to-lake architectures viable for more workloads.

What to watch: how the managed compaction compares to hand-tuned pipelines on large-scale streams, whether the cost-reduction claims hold under real query patterns, and adoption relative to competing managed Iceberg ingestion from Confluent, Databricks and others chasing the same real-time-lakehouse convergence.