On Might 12, 2026, we introduced the overall availability of Amazon Redshift RG situations, powered by AWS Graviton processors. RG situations are as much as 2.2x as quick for information warehouse workloads and as much as 2.4x as quick for information lake workloads, all at 30% lower cost per vCPU in comparison with RA3 situations. RG situations assist all information lake codecs supported by RA3 and remove Amazon Redshift Spectrum’s per-TB scanning costs. RG situations function a custom-built built-in vectorized question engine, making them a extra performant and cost-effective basis for unified analytics.
We’re launching with two occasion sizes: rg.xlarge and rg.4xlarge, with further sizes coming later this 12 months.
Why we constructed this
RG situations convey the facility of AWS Graviton processors to Amazon Redshift Provisioned clusters for the primary time, paired with a purpose-built vectorized question engine. By combining Graviton’s superior price-performance with the newest Amazon Redshift improvements, RG situations ship a step-change enchancment throughout two dimensions: considerably decrease price and meaningfully quicker efficiency for each warehouse and information lake workloads utilizing Apache Iceberg and Apache Parquet. We constructed RG that will help you keep away from selecting between efficiency and economics. Graviton prices much less to function, and we’re passing that profit to you whereas concurrently elevating the efficiency bar. Equally vital, we designed RG to take care of full function parity with RA3, so you’ll be able to modernize your current clusters with out rearchitecting workloads or sacrificing capabilities you rely on immediately.
This mix can be more and more essential for agentic synthetic intelligence (AI) workloads. AI brokers working at scale generate a brand new class of analytics demand: excessive volumes of distinctive, unpredictable queries that require quick, low-latency responses to maintain brokers productive. Conventional price-performance ratios make operating these workloads at scale cost-prohibitive. RG situations handle this head-on. Decrease per-vCPU pricing makes sustained high-query volumes economically viable, whereas improved question efficiency makes certain brokers get solutions quick sufficient to stay efficient. Collectively, this gives the muse for AI-driven analytics on the scale and economics that agentic workloads demand.
What’s new
RG situations: Higher efficiency, decrease price
RG situations run on AWS Graviton, Amazon’s custom-designed cloud processor constructed from the bottom as much as ship superior price-performance and vitality effectivity. This interprets straight into RG situations providing extra compute cores, larger reminiscence bandwidth, and decrease inter-process communication latency in comparison with RA3, with efficiency enhancements throughout warehouse, information lake, and blended workloads.
Graviton prices much less to function, and we’re passing that profit on to you. RG situations are priced at a 30% decrease price per vCPU in comparison with RA3. Reserved Occasion pricing follows the identical mannequin, making RG Reserved Cases equally 30% less expensive than RA3. For pricing particulars, go to the Amazon Redshift pricing web page.
Efficiency outcomes
RG situations ship quicker, extra environment friendly analytics throughout your most demanding warehouse and information lake workloads, whether or not you’re querying structured information in Amazon Redshift Managed Storage (RMS), operating analytics over Iceberg tables in Amazon Easy Storage Service (Amazon S3), or processing Parquet recordsdata at scale. Iceberg workloads see probably the most important good points, delivering as much as 2.4x quicker question execution. Parquet workloads ship as much as 1.5x quicker question execution, and RMS-based information warehouse workloads ship as much as 2.2x quicker question execution. All efficiency enhancements are measured utilizing industry-standard TPC-DS and TPC-H benchmarks at 10 TB scale on rg.4xlarge situations.
When mixed with RG’s 30% decrease per-vCPU pricing in comparison with RA3, these efficiency good points translate to even larger price-performance enhancements, delivering extra analytics worth for each greenback spent.
Constructed-in information lake question engine – no extra Spectrum costs
With RA3, information lake queries had been offloaded to a separate fleet of nodes known as Amazon Redshift Spectrum, scanning information externally and returning outcomes again to the cluster. This structure launched community overhead, added latency, and imposed a $5/TB scanning cost on each question. RG situations change this basically with a custom-built vectorized information lake engine operating straight contained in the cluster, eliminating Spectrum scanning costs.
The aim-built vectorized engine features a extremely optimized scan layer that implements the newest information pruning strategies, a purpose-built I/O subsystem, and a spread of optimizations that use Graviton’s processing capabilities to make scanning Iceberg and Parquet information extremely environment friendly. Past uncooked scan efficiency, the engine introduces JIT ANALYZE, a functionality that robotically collects and makes use of statistics for information lake tables throughout question execution. This eliminates the necessity for guide statistics assortment. The system makes use of clever heuristics to determine queries that may profit from statistics, maintains light-weight sketch information constructions, and builds high-quality table-level and column-level statistics, all transparently. Having up-to-date statistics on information lake tables can ship orders-of-magnitude enhancements in question efficiency, and with JIT ANALYZE, you get this profit robotically with out operational overhead.
What prospects are saying
Sean Lynch, Vice President, Information and Structure, Southwest Airways:
“Amazon Redshift RG situations have the potential to ship significant enterprise affect for Southwest Airways. Based mostly on preliminary testing in our improvement atmosphere, our information warehouse workloads run 50-60% quicker, and information lake analytics are 45% quicker, enabling groups to get insights sooner, reply to operational circumstances quicker, and make data-driven choices with much less latency. These early outcomes are encouraging, and we’re excited to validate and scale these enhancements in manufacturing. All of this comes with out per-terabyte Spectrum scanning costs, delivering 30% decrease price than RA3 at a time when gas costs proceed to stress {industry} margins.”
Akshay Srinivasan, Information Engineer, tombola:
“The brand new Graviton-based Amazon Redshift RG situations delivered 1.8x-2x quicker write throughput and as much as 2.2x quicker learn speeds in comparison with RA3 throughout a various set of batch and analytical jobs, enabling us to course of 40% extra throughout the similar window. Compressed ETL cycles, accelerated time-to-insight, and decision-making now not bottlenecked by the pipeline. Collectively, these translated straight into brisker information reaching our analysts and enterprise groups sooner. What made this much more compelling was a concurrent 30% discount in compute spend alongside the good points. Delivering extra for much less is a uncommon consequence, and one value highlighting. In a volume-heavy gaming {industry} at tombola, the place question latency and price compound at scale, this has been one of many extra impactful platform choices we’ve made this 12 months.”
Modernizing your workloads to RG
At present, we’re launching rg.xlarge and rg.4xlarge occasion sizes, out there now so that you can modernize your current Amazon Redshift provisioned workloads. RG situations assist three migration paths, all accessible straight from the AWS Administration Console:
- Elastic Resize (advisable): The quickest path for many prospects migrating from RA3 or DC2, with solely 10-Quarter-hour of downtime.
- Snapshot & Restore: Greatest for you if that you must make configuration adjustments as a part of your migration.
- Basic Resize: Out there for workloads that require a full cluster rebuild.
Earlier than migrating your manufacturing workloads, we strongly suggest validating your queries and workloads on RG situations first. We’ve printed an Improve Information that will help you right-size your cluster and plan your migration with confidence.
Getting began
You can begin utilizing the RG situations (rg.xlarge and rg.4xlarge) immediately within the following AWS Areas: US East (N. Virginia), US East (Ohio), US West (Oregon), US West (N. California), Canada (Central), South America (São Paulo), Europe (Eire), Europe (Frankfurt), Europe (London), Europe (Paris), Europe (Stockholm), Europe (Milan), Europe (Spain), Asia Pacific (Tokyo), Asia Pacific (Seoul), Asia Pacific (Singapore), Asia Pacific (Sydney), Asia Pacific (Mumbai), Asia Pacific (Jakarta), Asia Pacific (Hong Kong), Asia Pacific (Osaka), Asia Pacific (Malaysia), Asia Pacific (Hyderabad), Asia Pacific (Taipei), and Asia Pacific (Melbourne).
You’ll be able to launch new clusters or migrate current clusters by way of the AWS Administration Console, AWS Command Line Interface (AWS CLI), or AWS API.
To create a brand new RG cluster within the Amazon Redshift console
- Assessment the Cluster and Nodes within the Amazon Redshift documentation.
- Select Amazon Redshift on the AWS Administration Console and select Create Cluster.
- Within the Create Cluster display, select the required RG node sort.
To modernize from RA3 or DC2 within the Amazon Redshift console
- Assessment the Improve Information within the Amazon Redshift documentation.
- Select your migration path. Elastic Resize is the appropriate place to begin for many prospects.
- Select the required RG node sort.
For pricing particulars, go to the Amazon Redshift pricing web page.
Clear up
If you’re evaluating RG situations in a check or improvement atmosphere and don’t want to proceed, you’ll be able to delete your RG cluster straight from the AWS Administration Console or by utilizing the AWS CLI to keep away from incurring further costs. If you happen to used Snapshot & Restore to create a check RG cluster alongside your current RA3 cluster, be sure you delete the RG cluster and any related snapshots you now not want. If you’re utilizing Information Sharing throughout migration, bear in mind to take away information shares and decommission your RA3 cluster after you might have totally validated your workloads on RG.
Conclusion
Amazon Redshift RG situations characterize a big step ahead for you when you run information warehouse and information lake workloads on AWS. By bringing AWS Graviton processors to Amazon Redshift Provisioned clusters for the primary time, paired with a purpose-built vectorized native information lake engine, RG situations ship as much as 2.4x higher efficiency on Iceberg workloads, as much as 1.5x on Parquet, and as much as 2.2x on RMS information warehouse workloads, all at 30% decrease per-vCPU price than RA3. The elimination of Amazon Redshift Spectrum scanning costs makes information lake question prices predictable for the primary time.
To get began with RG situations, go to the Amazon Redshift RG documentation to evaluate your workload and plan your migration.
Sources
Questions or suggestions? Drop a remark or be part of the dialogue on AWS re:Put up.
In regards to the authors
