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

AWS made Glue 6.0 generally available, headlined by a 30% price reduction and full support for Apache Iceberg v3, including the new VARIANT data type with automatic shredding for semi-structured data. The release modernizes the runtime to Apache Spark 4.1, Python 3.13, and Scala 2.13, and bundles newer versions of Hudi and Delta Lake.

While less flashy than frontier-model news, Glue 6.0 matters because data engineering is the unglamorous foundation of every AI pipeline—models are only as good as the data plumbing feeding them. The 30% price cut fits a clear AWS pattern this week of cost reductions (GPT-5.6 Sol pricing, cross-region inference) aimed at defending its cloud economics against Google Cloud's 82% growth and Azure's 43%.

Apache Iceberg v3 support is the technically significant piece: the VARIANT type with automatic shredding lets teams store and query semi-structured JSON efficiently within an open table format, reducing the friction between operational and analytical data. The Spark 4.1 and Python 3.13 upgrades keep Glue current with the open-source ecosystem.

This lands amid a wave of AWS data-and-agent tooling—native vector search in DynamoDB, the Agentic Data Operations Platform (ADOP) reference architecture compressing new-source onboarding from weeks to hours, and query-aware RAG compression on Bedrock. Together they show AWS building an integrated stack where data infrastructure and agentic AI converge. Watch whether the price cuts pressure competitors' data-engineering offerings and whether Iceberg v3 adoption accelerates.