Amazon Redshift said its new AWS Graviton-based RG instances run data warehouse workloads up to 2.2x as fast as RA3 instances while cutting price per vCPU by 30%.
The launch targets rising query volumes from analytics and AI agents by combining warehouse and data lake SQL in one engine, with Iceberg queries up to 2.4x faster and Parquet up to 1.5x faster than RA3.
Redshift now executes data lake queries on cluster nodes inside the customer VPC, eliminating the need for Spectrum and removing the $5-per-terabyte scanning fee.
AWS said customers can launch new RG clusters or migrate existing ones with Elastic Resize in 10-15 minutes of downtime for compatible setups, or use Snapshot and Restore without changing external tables or query syntax.
RG instances are available now across multiple AWS regions in the Americas, Europe, Asia Pacific and the Middle East, as Amazon pushes lower-cost infrastructure for mixed analytics and agentic AI workloads.
Beyond the hype, what are the hidden trade-offs when querying data lakes directly with Redshift's new RG instances?
How will Amazon’s Graviton-powered Redshift reshape the competitive landscape against rivals like Snowflake and Databricks?
With Graviton's energy savings, is sustainable computing becoming the ultimate advantage in the cloud data warehouse wars?