AWS Glue 6.0 launches with 30% price cut and Apache Iceberg v3 support

AWS Glue 6.0, the general-availability release of Amazon's serverless data-integration service, pairs a headline 30% price reduction with a substantial modernization of its open-table-format support. The standout is full Apache Iceberg v3 support, including the new VARIANT data type with automatic shredding — a feature that lets semi-structured data be stored and queried efficiently without rigid schema definition.

Under the hood, Glue 6.0 upgrades to Apache Spark 4.1, Python 3.13, and Scala 2.13, alongside newer Apache Hudi and Delta Lake versions. For data engineering teams, the combination of a 30% cost cut and current-generation runtimes is a straightforward win, reducing both bill and technical debt for ETL and lakehouse workloads.

The timing fits AWS's broader data-and-agent narrative this week, which included native vector search in DynamoDB, query-aware RAG compression on Bedrock, and an Agentic Data Operations Platform reference architecture that compresses data-source onboarding from weeks to hours. Glue 6.0's Iceberg v3 support is foundational plumbing for those higher-level agentic data workflows — agents need governed, performant table formats to operate against.

The competitive context is the ongoing lakehouse format war, where Iceberg has been winning mindshare against proprietary formats. AWS's rapid adoption of Iceberg v3 (including VARIANT) signals it is betting on open-format interoperability rather than lock-in at the storage layer. What to watch is real-world VARIANT performance at scale and whether the 30% price cut pressures competing managed-Spark offerings.