Introduction
Amazon Redshift has been a cornerstone of cloud data warehousing since 2013, consistently pushing the boundaries of cost-efficiency and performance. As data volumes explode and AI agents query warehouses at unprecedented scales, the need for a robust, cost-effective solution has never been greater. Today, Amazon Redshift introduces a new instance family—RG instances—powered by AWS Graviton processors, designed to meet the demands of human-driven analytics and autonomous AI workloads alike.
Next-Generation Performance with AWS Graviton
The RG instance family leverages the energy-efficient ARM-based AWS Graviton processors to deliver a significant leap in performance. Compared to the previous generation RA3 instances, RG instances run data warehouse workloads up to 2.2x faster while reducing cost per vCPU by 30%. This means organizations can process more queries in less time, at a lower total cost.
Performance Benchmarks
In benchmark testing, RG instances equipped with the integrated data lake query engine achieved remarkable speedups for common data formats:
- Up to 2.4x faster for Apache Iceberg tables
- Up to 1.5x faster for Apache Parquet tables
Integrated Data Lake Query Engine
Beyond raw compute power, RG instances come with a built-in data lake query engine. This removes the need for separate query services or complex ETL pipelines when analyzing data stored in Amazon S3. You can run SQL analytics across your data warehouse and data lake from a single engine, simplifying operations and reducing total analytics costs. The engine is enabled by default, so you can start querying your lake immediately without additional setup.
Instance Comparison: RG vs RA3
The following table shows recommended migration paths from existing RA3 instances to the new RG family, along with key specifications:
| Current RA3 Instance | Recommended RG Instance | vCPU | Memory (GB) | Primary Use Case |
|---|---|---|---|---|
| ra3.xlplus | rg.xlarge | 4 | 32 | Small cluster departmental analytics |
| ra3.4xlarge | rg.4xlarge | 12 → 16 (1.33:1) | 96 GB → 128 GB (1.33:1) | Standard production workloads, medium data volumes |
For accurate cost estimates, use the AWS Pricing Calculator with your specific workload patterns.
Ideal for Modern Workloads: BI, ETL, and AI Agents
Amazon Redshift has continued to evolve to handle high-velocity, low-latency queries. In March 2026, Redshift improved performance for new queries by up to 7 times, directly benefiting BI dashboards, ETL pipelines, and near-real-time analytics. The new RG instances take this further by delivering that performance at a lower cost, making them a natural fit for autonomous, goal-seeking AI agents that generate massive query volumes.
Whether you are refreshing a Tableau dashboard, running hourly ETL jobs, or feeding real-time insights to an AI agent, RG instances provide the speed and scalability required without budget surprises.
Getting Started with Redshift RG Instances
You can begin using RG instances immediately via familiar AWS tools:
- AWS Management Console – launch or migrate clusters with a few clicks
- AWS CLI – script your deployments
- AWS API – integrate with automation workflows
Conclusion
Amazon Redshift RG instances represent a major step forward in cloud data warehousing, combining the energy efficiency of AWS Graviton with an integrated data lake engine. They deliver up to 2.2x faster performance at 30% lower vCPU cost, while simplifying analytics across warehouses and data lakes. For organizations dealing with growing data volumes and AI-driven query loads, RG instances offer a compelling blend of speed, cost savings, and operational simplicity.