Amazon Redshift has been a cornerstone of cloud data warehousing since its launch in 2013, delivering enterprise-grade analytics at a fraction of on-premises costs. Now, with the introduction of AWS Graviton-based RG instances, Redshift is setting a new benchmark for performance, cost efficiency, and integrated data lake querying. This article unpacks the top 10 things you need to know about this groundbreaking release, from hardware upgrades to real-world use cases.
1. A Legacy of Continuous Innovation
From its dense compute roots to RA3 instances and serverless options, Redshift has always evolved to lower costs and accelerate queries. The new RG instances represent the latest leap—powered by AWS’s custom Graviton processors—delivering up to 2.2x faster performance than RA3 instances for traditional data warehouse workloads.
2. Why Data Lakes and Warehouses Need to Converge
Modern analytics often require querying both structured warehouse tables and cost-effective data lakes (e.g., Amazon S3). Organizations also face skyrocketing query volumes from AI agents, which demand low latency and high throughput. RG instances address this by combining a robust warehouse engine with an integrated data lake query engine, allowing you to query both from a single system with no data movement.
3. AI Agents Are Driving Unprecedented Query Loads
AI agents—autonomous programs that analyze data at machine speed—can generate query volumes dwarfing human analysts. This strains traditional architectures and spirals costs. Redshift RG instances are built to handle such workloads, with near-real-time response and the ability to scale efficiently under heavy agent-driven querying.
4. Up to 7x Faster for BI Dashboards and ETL
In March 2026, Redshift already improved performance for low-latency SQL queries by up to 7x. This translates to snappier BI dashboards, faster ETL pipelines, and responsive near-real-time analytics. RG instances build on this foundation, ensuring that even complex, concurrent queries stay quick.
5. Meet the New RG Instance Family
RG instances are a brand-new family powered by AWS Graviton processors. They are designed specifically for today’s combined data warehouse and data lake workloads. With options like rg.xlarge (4 vCPU, 32 GB memory) and rg.4xlarge (16 vCPU, 128 GB memory), they offer a modern replacement for older RA3 instances.
6. Performance Leap: 2.2x Faster Than RA3
For core data warehouse operations, RG instances deliver up to 2.2x the speed of comparably configured RA3 instances. This means you can process more queries in less time, reducing wait times for both human users and AI agents. The improvement stems from Graviton’s efficiency and architectural optimizations in Redshift.
7. Integrated Data Lake Query Engine Boosts Lakehouse Queries
The integrated data lake query engine—enabled by default—allows you to run SQL directly on Amazon S3 data lakes. Performance gains are substantial: up to 2.4x faster for Apache Iceberg tables and up to 1.5x faster for Apache Parquet compared to RA3 instances. This eliminates the need for separate query engines and simplifies your stack.
8. Cost Savings: 30% Lower Price per vCPU
While delivering superior performance, RG instances also cut costs. The price per vCPU is 30% lower than RA3 instances. Combined with faster query execution, total analytics spend drops significantly, especially for organizations with heavy data lake usage or AI agent workloads. Use the AWS Pricing Calculator to estimate savings for your specific patterns.
9. Simplified Operations: One Engine, No Data Movement
With the integrated data lake query engine, you no longer need to move data between the warehouse and data lake. Query both Redshift tables and S3 data lakes from a single SQL endpoint, reducing operational complexity and latency. This unified approach streamlines governance and security policies.
10. Getting Started: Migration and Launch Options
Launching new clusters or migrating existing RA3 workloads to RG instances is straightforward via the AWS Management Console, CLI, or API. Simply select the new RG instance type, and the data lake engine is active by default. Migrating from RA3 to RG is recommended for most production workloads to realize immediate performance and cost benefits.
Conclusion
The introduction of Amazon Redshift RG instances marks a pivotal moment for cloud analytics. By combining Graviton’s power efficiency with a unified warehouse-and-lake query engine, Redshift enables faster, cheaper, and simpler analytics for human and AI-driven workloads alike. Whether you’re running BI dashboards, ETL pipelines, or autonomous agents, RG instances provide the foundation for next-generation data processing. Ready to upgrade? Start exploring today in the AWS console.