AWS Graviton Processor
Enabling the best price performance in Amazon EC2
AWS Graviton processors are custom built by Amazon Web Services using 64-bit Arm Neoverse cores to deliver the best price performance for your cloud workloads running in Amazon EC2. Amazon EC2 provides the broadest and deepest portfolio of compute instances, including many that are powered by latest-generation Intel and AMD processors. AWS Graviton processors add even more choice to help customers optimize performance and cost for their workloads.
The first-generation AWS Graviton processors power Amazon EC2 A1 instances, the first ever Arm-based instances on AWS. These instances deliver significant cost savings over other general-purpose instances for scale-out applications such as web servers, containerized microservices, data/log processing, and other workloads that can run on smaller cores and fit within the available memory footprint.
AWS Graviton2 processors deliver a major leap in performance and capabilities over first-generation AWS Graviton processors. They power Amazon EC2 general purpose (M6g, M6gd, T4g), compute optimized (C6g, C6gd, C6gn), and memory optimized (R6g, R6gd, X2gd) instances, that provide up to 40% better price performance over comparable current generation x86-based instances for a wide variety of workloads, including application servers, micro-services, high-performance computing, CPU-based machine learning inference, video encoding, electronic design automation, gaming, open-source databases, and in-memory caches. They deliver 7x more performance, 4x more compute cores, 5x faster memory, and 2x larger caches.
Free Trial: Until June 30th 2021, all new and existing AWS customers can try the t4g.micro instances free for up to 750 hours per month. During the free-trial period, customers who run a t4g.micro instance will automatically get 750 free hours per month deducted from their monthly bill through June 2021. Refer to the FAQ for additional details.
Benefits
Best price performance for a broad spectrum of workloads
AWS Graviton2-based general purpose burstable (T4g), general purpose (M6g), compute-optimized (C6g), and memory-optimized (R6g) EC2 instances and their variants with NVMe-based SSD storage deliver up to 40% better price performance over comparable current generation x86-based instances1 for a broad spectrum of workloads such as application servers, microservices, video encoding, high-performance computing, electronic design automation, compression, gaming, open-source databases, in-memory caches, and CPU-based machine learning inference.
Extensive ecosystem support
AWS Graviton2 processors, based on the 64-bit Arm architecture, are supported by popular Linux operating systems including Amazon Linux 2, Red Hat, SUSE, and Ubuntu. Many popular applications and services from AWS and Independent Software Vendors also support AWS Graviton2-based instances, including Amazon ECS, Amazon EKS, Amazon ECR, Amazon CodeBuild, Amazon CodeCommit, Amazon CodePipeline, Amazon CodeDeploy, Amazon CloudWatch, Crowdstrike, Datadog, Docker, Drone, Dynatrace, GitLab, Jenkins, NGINX, Qualys, Rancher, Rapid7, Tenable, and TravisCI. Arm developers can also leverage this ecosystem to build applications natively in the cloud, thereby eliminating the need for emulation and cross-compilation, which are error prone and time consuming.
Enhanced security for cloud applications
Developers building applications for the cloud rely on cloud infrastructure for security, speed and optimal resource footprint. AWS Graviton2 processors feature key capabilities that enable developers to run cloud native applications securely, and at scale, including always-on 256-bit DRAM encryption and 50% faster per core encryption performance compared to first-generation AWS Graviton. Graviton2 powered instances are built on the Nitro System that features the Nitro security chip with dedicated hardware and software for security functions, and support for encrypted EBS storage volumes by default.
EC2 Instances Powered by AWS Graviton2
General Purpose

Best price performance for general purpose workloads with balanced compute, memory, and networking
Built for: General-purpose workloads such as application servers, mid-size data stores, microservices, and cluster computing.

Best price performance for burstable general purpose workloads
Built for: Broad range of burstable general purpose workloads such as large scale micro-services, small and medium databases, virtual desktops, and business-critical applications.
Compute Optimized

Best price performance for compute-intensive workloads
Built for: Compute-intensive applications such as high performance computing, video encoding, gaming, and CPU-based machine learning inference acceleration.
Memory Optimized

Best price performance for memory-intensive workloads
Built for: Memory-intensive workloads such as open-source databases (MySQL, MariaDB, and PostgreSQL), or in-memory caches (Redis, KeyDB, Memcached).

Lowest cost per GiB of memory in Amazon EC2
Built for: Memory-intensive workloads such as open-source databases (MySQL, MariaDB, and PostgreSQL), in-memory caches (Redis, KeyDB, Memcached), electronic design automation (EDA) workloads, real-time analytics, and real-time caching servers.
Local NVMe-based SSD storage options are also available in general purpose (M6gd), compute-optimized (C6gd), and memory-optimized (R6gd) instances. Additionally, compute-optimized C6gn instances with 100 Gbps networking and support for Elastic Fabric adapter (EFA) are also available.
EC2 Instances Powered by First Generation AWS Graviton

Amazon EC2 A1 Instances, powered by first-generation AWS Graviton processors, deliver significant cost savings over other general-purpose instances scale-out applications such as web servers, containerized microservices, data/log processing, and other workloads that can run on smaller cores and fit within the available memory footprint. These instances will also appeal to developers, enthusiasts, and educators across the Arm community.
Customer Testimonials

Snap Inc, known for its popular social media services such as Snapchat and Bitmoji, adopted AWS Graviton2 based instances to optimize their price performance on Amazon EC2. Snap uses Amazon EC2 C6g and M6g instances with Amazon Elastic Kubernetes Service (EKS) to run its Messaging Core Service, which is the heart of messaging on the platform, as well as its backend context service. Snap is also looking forward to adopting R6g instances for running its ElastiCache workloads.
"Using Graviton2-based instances, we were able to reduce the Snapchat messaging fleet size and significantly lower cost compared to C5 instances. The move to C6g instances reduced CPU utilization by roughly 10% because of better performance on Graviton2. We are planning to move additional workloads to Graviton2 in the future and have found similar performance gains on other large workloads."
- Aaron Sheldon, Software Engineer at Snap

Honeycomb.io provides an observability tool for introspecting and interrogating production systems, designed and evolved to meet the needs of platforms, microservices, serverless apps, and complex systems. Honeycomb.io is running Amazon EC2 M6g instances for their ingest worker test fleet. The entire transition process took about 15 hours by an experienced generalist systems engineer/SRE. Comparing to existing C5 instances, Honeycomb.io observed M6g instances to be superior in every aspect—they cost less, have more RAM, exhibit lower median and significantly narrower tail latency, and run cooler with the same proportional workload per host. Converting the entire ingest worker fleet to M6g instances has allowed Honeycomb.io to run 30% fewer instances, and each instance costs 10% less.

“We are always looking for ways to streamline our cloud costs, and we tried Graviton2 based instances as they represent a big opportunity to improve the price performance profile for our workloads on Amazon EC2. We have benchmarked similar performance with Graviton2-based R6g instances for our Kafka service, at 20% lower cost compared to R5 instances. With these results, we look forward to exploring the benefits of Graviton2 based instances on additional workloads.”

"Formula One is renowned for pushing the limits of efficiency and for this reason we are always exploring ways to optimise our cloud costs. We tried the new Graviton2 based instances as they offer significant opportunity to improve the price performance profile of some of the CFD workloads we currently run on Amazon EC2. Having benchmarked the new Graviton2-based C6g and C6gn instances we have seen the potential to lower costs by up to 40% compared to similar workloads run on C5n instances. Based on these results we plan to include Graviton2-based workflows for a range of our CFD simulations moving forward."
- Pat Symonds, Chief Technical Officer, Formula One

NextRoll provides an industry-leading marketing technology and data stack that fuels growth for businesses of all kinds via AdRoll, RollWorks, and NextRoll Platform Services.
"At NextRoll we are focused on providing the innovative solutions companies need to keep growing, wherever their ambition takes them next. To do that we are always seeking new ways to expand and optimize our service offerings. We are excited to adopt AWS Graviton2 based M6g and C6g instances for our web server and stream processing workloads built on Erlang, Go, and Java. We have observed up to 50% total cost savings over our previous generation EC2 instances for these workloads. Much more savings and fewer instances needed with higher performance. All software packages that we needed to migrate were easily available and we plan to switch everything we can toGraviton2 based instances."
- Valentino Volonghi, CTO at NextRoll

SmugMug operates two very large online photo platforms, SmugMug and Flickr, enabling more than 100M members to safely store, search, share, and sell tens of billions of photos.
"We were an early adopter of Graviton instances, starting with Amazon EC2 A1 instances. Today, we have our PHP-based photo-serving app tier running on AWS Graviton2 based Amazon EC2 M6g and C6g instances using Spot. It's not very often that a piece of technology comes around that enables you to get 40% better price performance just by recompiling. Moving to AWS Graviton2 based instances required minimal code changes; it just worked! We intend to migrate the vast majority of our workloads to AWS Graviton2 within the year and one of our next steps will be to move our caching servers to r6gd.”
Shane Meyers, Principal Operations Engineer, SmugMug
Partner Testimonials

“Aqua’s customers choose to deploy their container workloads on AWS Graviton2-based EKS clusters for improved workload density and cost-effectiveness. We partnered with AWS to extend our market-leading, full lifecycle protection of container workloads to include support for Arm64-based Graviton2 infrastructure, ensuring our mutual customers can enjoy its benefits with no compromise on security across all AWS container services.”
– Amir Jerbi, CTO and Co-founder of Aqua Security

As the largest continuous integration and delivery (CI/CD) platform, CircleCI is the central hub where developers seamlessly take ideas to execution, at scale. CircleCI is one of the few services to offer fully cloud-based Arm compute for CI/CD, allowing customers to build and test their applications for Arm without managing their own machines.
"The AWS Graviton2 instances make Arm-based CI/CD possible for the first time at scale. We needed a partner who could provide Arm-based compute with the capacity and reliability we require. With CircleCI’s Arm build fleet running on Graviton2, developers can build, test and deploy Arm-based applications in a truly serverless experience, without managing their own machines. By partnering with AWS, we’re able to bring that scale to the Arm developer community at a level not previously possible."
Tom Trahan, VP Business Development, CircleCI

CrowdStrike is a global cybersecurity leader in cloud-delivered endpoint protection built from the ground up to stop breaches. CrowdStrike Falcon delivers full spectrum runtime protection and visibility for hosts and hosted containers leveraging a cloud-native platform and a single lightweight sensor, without requiring any additional deployments or integrations. AWS Graviton2 Arm64-based instances offer compelling performance and cost benefits which will help drive adoption across a much wider variety of workloads. Under the shared responsibility model, AWS customers require consistent and seamless protection across their Amazon EC2 compute fleet, regardless of the underlying processor architectures. That is why CrowdStrike is delighted to extend CrowdStrike Falcon protection and visibility for all Graviton-based EC2 instances.

Datadog is the monitoring and analytics platform for developers, operations, and business users in the cloud age. "We're excited the Graviton2 instances are now generally available. We've seen our own developer productivity increase as these instances have boosted the performance of key continuous integration pipelines by 50%," said Jimmy Caputo, Product Manager at Datadog. "With the Datadog Agent for Graviton / Arm, all AWS customers can use Datadog to monitor the Graviton2 instance performance alongside the rest of their infrastructure.“

Dynatrace provides software intelligence to simplify cloud complexity and accelerate digital transformation. As a leader in cloud-native and hybrid observability, we see a massive growth in the popularity of AWS Graviton2. Our mutual customers are seeking to take advantage of the price/performance benefits of Graviton2 and are excited to do so with the exact same experience from Dynatrace full-stack OneAgent, which is now readily available for the Arm architecture and AWS Graviton2.
Software and Services
Operating Systems









Container Services






Developer Tools and Software
















Agents









And many others...
Partner Blogs
Genymotion
Nov 5th, 2020
Articles and announcements
Jeff Barr
Mar 16, 2021
Mar 12, 2021
Kirk Davis
Dec 4, 2020
Resources
1 20% lower cost and up to 40% higher performance for M6g, C6g, and R6g instances over M5, C5, and R5 instances respectively, based on internal testing of workloads with varying characteristics of compute and memory requirements.