AWS Glue 6.0 Ships With 30% Lower Pricing and Full Apache Iceberg v3 Support

Overview

Amazon Web Services has announced the general availability of AWS Glue 6.0, a new version of its serverless data-integration service that delivers “30% lower pricing than previous AWS Glue versions and introduc[es] full support for Apache Iceberg v3 features,” according to AWS’s official announcement. The release, detailed in a blog post by AWS News Blog lead blogger Channy Yun, is generally available today in all AWS Regions where AWS Glue operates, a fact independently confirmed by Database Trends and Applications.

What We Know

• AWS Glue 6.0 is “built on a fully modernized runtime, Apache Spark 4.1, Python 3.13, and Scala 2.13, delivering faster performance,” according to AWS, a set of version numbers also reported by DBTA.
• “AWS Glue provides the most complete Iceberg v3 implementation on any fully serverless managed Spark service, along with new capabilities that simplify ETL authoring, improve PySpark performance, and enable real-time streaming with single-digit millisecond latency,” the company said, according to both AWS and DBTA, which both quote the statement identically.
• The Iceberg support is built on Iceberg 1.11.0, and its headline feature is a new VARIANT data type with shredding support, which AWS says “achieves faster query read performance compared to traditional string data type columns for semi-structured data.” According to AWS, VARIANT shredding lets users “store and query JSON, logs, and event data without flattening schemas, eliminating duplicate data copies, custom parsing code, and pipeline breakage when schemas change.”
• Additional Iceberg v3 capabilities include new “Geometry and Geography data types” for spatial processing, “nanosecond-precision timestamps” for IoT and financial workloads, and “unknown type handling” that lets pipelines tolerate evolving upstream schemas without failing.
• On the Spark 4.1 runtime side, AWS introduced Spark Declarative Pipelines, where “data engineers declare transformations, specifying what data should look like, while the engine automatically determines execution order and optimization,” as well as Arrow-native execution for Python UDFs and UDTFs, which AWS says “eliminates serialization overhead between Python and the JVM, improving PySpark performance for complex transformations.”
• For streaming workloads, AWS Glue 6.0 adds a real-time streaming mode built on “Spark 4.1’s Real-Time Mode with Glue-optimized execution,” which AWS says supports “real-time event processing, low-latency data transformation pipelines, and time-sensitive data routing.”
• Migrating to the new version requires no API changes: users select AWS Glue 6.0 through the existing --glue-version parameter in the create-job or update-job APIs, and AWS says existing jobs can also be moved over using a Spark upgrade agent in AWS Glue Studio or an auto-upgrade feature.

What We Don’t Know

• AWS has not published specific benchmark figures quantifying the query-performance gains from VARIANT shredding or the exact latency achieved by the new real-time streaming mode beyond the general “single-digit millisecond” characterization.
• Neither AWS nor DBTA detailed how the 30% price reduction breaks down across AWS Glue’s different billable components — per-second crawler and ETL job compute versus AWS Glue Data Catalog storage and access fees.

Analysis

AWS Glue 6.0’s emphasis on Apache Iceberg v3 places it squarely in the ongoing shift among cloud data platforms toward open table formats as the default layer for analytics and AI workloads, rather than proprietary storage formats tied to a single vendor. By building the release on Iceberg 1.11.0 and marketing it as the “most complete Iceberg v3 implementation on any fully serverless managed Spark service,” AWS is positioning managed Glue jobs as a direct on-ramp for organizations that have already standardized their data lakes on Iceberg, competing with other engines and cloud services that support the same open specification. The choice to require no API changes for the upgrade — letting existing pipelines opt in via a version parameter or an auto-upgrade path — also reflects a now-familiar pattern among major cloud providers of trying to lower the switching cost for incremental infrastructure upgrades, encouraging faster adoption of both the price cut and the newer Spark 4.1 runtime underneath it.