Amazon Redshift Launches Graviton-Powered RG Instances: Up to 2.2x Faster, 30% Cheaper, with Integrated Data Lake Query Engine

Breaking News – Amazon Redshift today announced the general availability of its new RG instance family, powered by AWS Graviton processors and featuring an integrated data lake query engine. The new instances deliver up to 2.2x faster performance than the current RA3 generation, at 30% lower price per vCPU, and enable customers to query both warehouse tables and Amazon S3 data lakes from a single engine.

The announcement comes as organizations face skyrocketing query volumes from both human analysts and AI agents. "Amazon Redshift RG instances are built to handle the scale and speed demands of modern analytics and agentic AI workloads," said a company spokesperson. "We've combined the cost-efficiency of AWS Graviton with a unified query engine that accelerates data lake queries up to 2.4x for Apache Iceberg and 1.5x for Apache Parquet."

Background

Since 2013, Amazon Redshift has evolved through multiple architectural generations—from dense compute to RA3, and from provisioned to serverless—each making queries cheaper, faster, and more efficient. Over the past decade, customers increasingly rely on both data warehouse tables for structured, frequently accessed data and data lakes for cost-effective storage of diverse datasets.

The rise of AI agents has further strained resources, as these autonomous systems query data warehouses at scales "that dwarf typical human usage," leading to spiraling operational costs. In March 2026, Redshift already improved new query speeds by up to 7x to support BI dashboards, ETL workloads, and near-real-time analytics.

What This Means

For enterprises running mixed data warehouse and data lake workloads, RG instances offer a compelling value proposition. The integrated data lake query engine eliminates the need for separate engines, reducing total analytics costs while simplifying operations. Customers can run SQL across all their data from a single system, whether querying Redshift tables or Apache Iceberg and Parquet files stored in Amazon S3.

The performance gains are especially significant for high-volume, low-latency applications. "This blend of speed, cost efficiency, and an integrated data lake query engine makes Redshift RG instances well-suited to handle the high query volumes and low-latency requirements of today's analytics and agentic AI workloads," the spokesperson added.

Instance Comparison

Amazon recommends the following migration path from current RA3 instances to new RG instances:

• ra3.xlplus → rg.xlarge: 4 vCPUs, 32 GB memory – ideal for small cluster departmental analytics.
• ra3.4xlarge → rg.4xlarge: 12 to 16 vCPUs (1.33:1 ratio), 96 to 128 GB memory (1.33:1) – for standard production workloads with medium data volumes.

Customers are advised to use the AWS Pricing Calculator with their specific workload patterns to estimate savings.

Getting Started

Amazon Redshift RG instances are available now through the AWS Management Console, AWS CLI, or AWS API. New clusters can be launched directly, or existing clusters can be migrated. The integrated data lake query engine is enabled by default, requiring no additional configuration.

For more details, refer to the Amazon Redshift documentation.