03/04/2026
Are S3 Table Buckets finally ready for prime time?
When AWS introduced S3 Tables at re:Invent, they promised a more warehouse-like experience for data lakes: row-level updates, automatic optimization, and better performance without the small-file penalty.
In his latest post, Keith Gregory revisits S3 Tables a year later and puts them head-to-head with Redshift in real-world performance tests.
He breaks down:
• Where S3 Tables now feel like a traditional data warehouse
• How Athena performance compares to Redshift across multiple query patterns
• The surprising performance of plain Parquet
• The operational tradeoffs around permissions, catalogs, orchestration, and tooling
• Why S3 Tables still need maturation before they clearly beat Redshift
This is not a theoretical overview. Keith walks through actual datasets, query timings, storage comparisons, and practical integration details across Athena, Glue, and Redshift Serverless.
If you are designing or rethinking your AWS data platform, this is a must-read.
Read the full blog post:
https://ow.ly/26yB50Ykzlj