AWS Glue 6.0 Delivers 30% Price Cut and Full Apache Iceberg v3 Support

Amazon Web Services has announced the general availability of AWS Glue 6.0, introducing a major upgrade to its serverless data integration engine alongside an across-the-board 30% price reduction per Data Processing Unit (DPU) hour. The new release transitions Glue to an AWS-optimized runtime based on Apache Spark 4.1, Python 3.13, and Scala 2.13. Crucially, Glue 6.0 provides comprehensive support for the Apache Iceberg v3 specification (built on Iceberg 1.11.0), which brings native support for the VARIANT data type with automated shredding, deletion vectors for accelerated row-level mutations, nanosecond-precision timestamps, and native geospatial geometry and geography types. For cloud practitioners, data engineers, and FinOps leaders, this release provides immediate relief on data processing bills while solving architectural bottlenecks. In modern lakehouses, semi-structured formats like JSON payloads, application logs, and nested sensor events have traditionally forced engineers into an unpleasant trade-off: either manually maintain flattened relational schemas that break upon upstream schema changes, or store raw strings that cripple query performance. Iceberg v3's VARIANT type with automatic sub-field shredding resolves this dilemma by decomposing nested payloads into columnar storage transparently. Furthermore, the 30% price drop compounds with Spark 4.1 execution efficiencies, enabling teams to scale ingestion throughput without linear cost inflation. This development reflects a decisive industry convergence around open table formats, specifically Apache Iceberg, as the de facto standard for cloud data architectures. As organizations dismantle legacy proprietary data warehouses in favor of multi-engine lakehouses, cloud providers must compete not only on raw runtime speed but on native table format compatibility and serverless pricing efficiency. Glue 6.0 also reflects the growing blur between batch and streaming pipelines. By incorporating Spark 4.1's Real-Time Mode to achieve single-digit millisecond latency for stateless streams alongside Spark Declarative Pipelines (SDP), AWS is positioning Glue as a unified operational layer capable of handling both heavy historical transforms and continuous event routing. In practice, data teams should immediately evaluate upgrading existing pipelines from Glue 5.1 to 6.0 to capture baseline cost savings. However, practitioners should approach migrations systematically. Moving from Spark 3.5 to Spark 4.1 introduces breaking runtime changes, stricter SQL type coercion, and Python environment shifts. Teams should utilize the AWS Glue Spark Upgrade Agent to identify script incompatibilities, validate PyArrow-native user-defined functions (UDFs), and run shadow validation pipelines. While the VARIANT data type unlocks massive performance and maintenance wins, architects must ensure downstream query engines (such as Amazon Athena, Amazon Redshift, or Trino) fully support Iceberg v3 features before altering core production schemas.