Amazon Redshift launches RG instances powered by AWS Graviton

Amazon Web Services has announced the general availability of RG instances for Amazon Redshift, marking a significant hardware upgrade for the cloud data warehouse service. The new provisioned cluster nodes are powered by AWS Graviton processors and deliver up to 2.4x faster performance than the previous generation RA3 instances while offering 30% lower cost per vCPU. The RG instances feature a custom-built vectorized data lake query engine that can process Apache Iceberg and Parquet data directly on cluster nodes, eliminating the need for Redshift Spectrum's separate scanning infrastructure and its associated per-terabyte charges. The performance improvements span multiple workload types, with RG instances delivering up to 2.2x faster processing for structured data warehouse workloads on Redshift Managed Storage, up to 2.4x acceleration for Apache Iceberg queries, and up to 1.5x speed improvements for Parquet workloads. The architecture includes several technical enhancements including a purpose-built I/O subsystem with smart prefetch capabilities, NVMe caching, vectorized Parquet scans, and advanced file and partition-level pruning. Additional features include Just-in-Time Analyze for automatic query optimization and intelligent NVMe caching to reduce data lake round-trips for frequently accessed datasets. AWS is launching RG instances in two sizes: rg.xlarge and rg.4xlarge, with existing RA3 clusters able to migrate through Snapshot & Restore, Elastic Resize, or Classic Resize options. The instances are available with flexible pricing including On-Demand and Reserved Instance options with one-year and three-year terms.