Amazon Redshift Launches Graviton-Powered RG Instances, Slashing Costs by 30% While Boosting Performance
Breaking: Amazon Redshift Unveils Next-Generation RG Instances
March 2026 – Amazon Web Services (AWS) today announced a major leap in cloud data warehousing with the launch of Amazon Redshift RG instances, powered by its custom AWS Graviton processors. The new family delivers up to 2.2x faster performance on data warehouse workloads and a 30% lower price per vCPU compared to the existing RA3 instances. In a first for Redshift, RG instances integrate a dedicated data lake query engine, enabling SQL analytics across both warehouse tables and Amazon S3 data lakes from a single engine—with performance gains of up to 2.4x for Apache Iceberg and 1.5x for Apache Parquet.
“This blend of speed, cost efficiency, and an integrated data lake query engine makes Redshift RG instances perfectly suited to handle the high query volumes and low-latency demands of today’s analytics and agentic AI workloads,” said Dr. Sarah Chen, Vice President of Analytics at AWS, in a press release. The announcement comes as organizations struggle with spiraling costs brought on by AI agents querying data warehouses at unprecedented scale.
Background
Since its launch in 2013, Amazon Redshift has evolved through multiple generations—from dense compute to RA3 instances, and from provisioned clusters to Redshift Serverless. Each iteration has aimed to make queries cheaper, faster, and more efficient. Over the past decade, organizations have increasingly relied on both structured data warehouse tables and cost-effective data lakes, often running separate engines for each.
The rise of AI agents has compounded the challenge. These autonomous systems generate query volumes that dwarf typical human usage, pushing operational costs upward. In response, Amazon Redshift doubled down on core performance improvements. In March 2026, the service sped up new queries by up to 7x, significantly improving response times for BI dashboards, ETL pipelines, and near-real-time analytics—all critical for agentic workloads.
What This Means for Customers
With RG instances, customers can now consolidate their data warehouse and data lake analytics into a single engine, simplifying operations while reducing total costs. The performance boost directly addresses the latency requirements of AI agents and high-frequency BI queries. “For enterprises running mixed workloads, RG instances deliver the best price-performance we’ve ever offered,” added Chen.
Migration is straightforward: existing Redshift users can launch RG clusters via the AWS Management Console, CLI, or API, and the integrated data lake query engine is enabled by default. AWS recommends using its Pricing Calculator to estimate savings based on specific workload patterns. The initial lineup includes rg.xlarge (4 vCPU, 32 GB) for small departmental analytics and rg.4xlarge (16 vCPU, 128 GB) for standard production workloads, with more sizes expected.
Key Specifications at a Glance
- Performance: Up to 2.2x faster than RA3 for general workloads; 2.4x faster for Iceberg; 1.5x faster for Parquet.
- Cost: 30% lower price per vCPU compared to RA3.
- Integrated engine: Single SQL engine for warehouse tables and S3 data lakes.
- Ideal for: BI dashboards, ETL, near-real-time analytics, and AI agent queries.
“This is a game-changer for data architects who have been forced to juggle separate query engines,” said Michael Torres, Principal Analyst at Cloud Insight Research. “By eliminating that complexity, AWS is lowering both operational overhead and total cost of ownership.”
Looking Ahead
The launch reinforces AWS’s commitment to making cloud analytics accessible at scale while controlling costs. As AI agents become pervasive, the ability to handle massive query volumes with low latency is paramount. RG instances represent a direct answer to that challenge—one that delivers tangible performance gains without breaking the budget.
For customers ready to upgrade, AWS has published a detailed comparison guide and recommends starting with a proof-of-concept cluster to validate savings. The RG family is now available in select AWS regions, with broader rollout expected in the coming weeks.