Amazon EC2 R9g and R9gd Instances Powered by AWS Graviton5 Processors Now Generally Available

Amazon Web Services has officially announced the general availability of the Amazon EC2 R9g and R9gd instance families, marking a significant technological leap forward in cloud computing efficiency and performance. Powered by the newly introduced AWS Graviton5 processors—touted by the company as the most energy-efficient silicon ever developed by AWS—these memory-optimized instances deliver up to a 25 percent improvement in raw compute performance compared to their predecessors, the Graviton4-based R8g instances. This launch highlights the continuous evolution of custom cloud hardware, addressing the escalating demand for high-performance, cost-effective, and sustainable infrastructure capable of handling the most demanding modern workloads.

The introduction of the R9g and R9gd instances comes at a time when enterprises worldwide are grappling with the dual pressures of scaling their data-intensive operations while aggressively cutting carbon footprints and infrastructure costs. Memory-intensive applications, ranging from distributed databases and in-memory caches like Valkey, Redis, and MemCached, to real-time big data analytics and containerized microservices orchestrated via Kubernetes, Docker, Amazon EKS, and Amazon ECS, require underlying hardware that offers low latency and high throughput. By leveraging the advanced architectural capabilities of the Graviton5 processor, AWS is positioning these new instances as the definitive standard for organizations looking to maximize performance per vCPU while simultaneously driving down energy consumption.

Chronology and Evolution of AWS Graviton Silicon

To understand the significance of the Graviton5 launch, it is helpful to examine the deliberate and rapid evolution of Amazon’s custom silicon strategy over the past several years. The journey began with the introduction of the first-generation AWS Graviton processor in 2018, which was designed primarily for scale-out workloads that did not require the heavy lifting of traditional general-purpose x86 processors. While groundbreaking in concept, early iterations served as an alternative for lightweight web servers and microservices.

With the launch of Graviton2 in 2019, AWS fundamentally shifted the market perception of ARM-based processors in the cloud, offering dramatic performance leaps and cost savings that compelled major software vendors to build native ARM support. Graviton3, introduced in 2021, brought further enhancements, including significant boosts in floating-point performance, cryptographic processing, and machine learning workloads. Last year’s introduction of Graviton4 set a new benchmark for core density and energy efficiency, quickly being adopted by thousands of global enterprises.

The arrival of Graviton5 and the subsequent rollout of the R9g and R9gd instances represent the maturation of this multi-year silicon strategy. Rather than merely iterating on clock speeds, AWS has systematically re-architected the underlying chip design, cache hierarchies, and memory subsystems to address the specific bottlenecks encountered by enterprise databases and analytical engines. This methodical progression demonstrates AWS’s long-term commitment to reducing reliance on merchant silicon while delivering tailored hardware solutions that align precisely with cloud-scale demands.

Architectural Innovations: What Sets R9g and R9gd Apart

The technical advancements underpinning the R9g and R9gd instances extend well beyond raw clock frequency. The Graviton5 processors introduce several critical hardware improvements, most notably a substantially larger L3 cache, faster memory subsystems, and heightened network and Amazon EBS bandwidth. For organizations migrating workloads from R8g instances, these enhancements translate directly to more processing power per virtual CPU (vCPU), enabling applications to process larger datasets in memory with significantly reduced wait times.

A standout feature of the new instance lineup is the integration of Instance Bandwidth Configuration (IBC). This capability empowers cloud architects and system administrators to dynamically adjust the allocation of network bandwidth between Amazon EBS and Amazon VPC networking by up to 25 percent. In database management systems and distributed caching environments—where unpredictable I/O spikes or network congestion can severely degrade application responsiveness—IBC provides the fine-grained control necessary to optimize throughput and guarantee predictable performance under heavy load.

Furthermore, all R9g and R9gd instances are anchored by the AWS Nitro System. By offloading virtualization, storage, and networking functions to dedicated hardware accelerators, the Nitro System eliminates the traditional "hypervisor tax," granting customer applications near-bare-metal performance while maintaining robust security isolation.

Security and the Nitro Isolation Engine

Security remains a paramount consideration in multi-tenant cloud environments, and the R9g and R9gd releases introduce a groundbreaking security paradigm through the Nitro Isolation Engine (NIE). Initially deployed earlier this year with the C9g and M9g instances, the NIE represents a fundamental evolution in hypervisor architecture.

The Nitro Isolation Engine is a purpose-built hardware and software component responsible for enforcing absolute isolation between virtual machines. It achieves this by rigorously mediating all access to virtual machine memory, CPU register states, and I/O devices through a strictly minimized set of application programming interfaces (APIs). What distinguishes the Nitro Isolation Engine from traditional hypervisor security mechanisms is its reliance on formal verification—a rigorous mathematical technique used to prove unequivocally that the system behaves precisely as intended under all possible execution paths, rather than merely passing empirical test cases.

By employing formal verification, AWS has established the Nitro System as the first formally verified cloud hypervisor in the industry, pioneering a new standard for mathematically proven cloud security. This level of verification provides enterprise compliance officers, financial institutions, and government agencies with unprecedented cryptographic and mathematical assurances regarding tenant isolation and data protection.

Detailed Specifications and Workload Suitability

The R9g family is designed specifically for workloads requiring balanced high-performance compute and expansive memory capacity. Available in 11 distinct sizes ranging from a modest single vCPU (r9g.medium) to a massive bare-metal configuration with 192 vCPUs and 1,536 GiB of memory (r9g.metal-48xl), the instances cater to a wide spectrum of operational scales.

The standard R9g instances are EBS-optimized, making them exceptionally well-suited for distributed applications where persistent storage is managed via Amazon Elastic Block Store. Network bandwidth scales linearly with instance size, starting at up to 15 Gbps for smaller configurations and scaling up to 100 Gbps for the largest 48xlarge and metal variants. Similarly, EBS bandwidth scales from up to 12 Gbps to a dedicated 72 Gbps.

For applications requiring high-speed, low-latency temporary storage or scratch space, the R9gd variant incorporates local NVMe-based SSD block-level storage directly attached to the physical host. These instances are ideal for open-source databases, distributed real-time big data analytics, large in-memory transactional processing, and caching tiers that benefit from local, high-IOPS storage layers. The local NVMe capacity scales from 59 GB on the entry-level r9gd.medium up to an impressive 11.4 TB (configured as three 3.8 TB drives) on the r9gd.48xlarge and metal instances.

Broad Software Ecosystem and Seamless Migration

A common hurdle in the adoption of novel processor architectures has historically been software compatibility. However, the maturation of the ARM ecosystem—largely catalyzed by the widespread success of earlier Graviton generations—has ensured that the transition to Graviton5 is remarkably frictionless.

R9g and R9gd instances support a comprehensive array of major Linux distributions, including Amazon Linux 2023, Amazon Linux 2, Ubuntu 22.04 and newer versions, Red Hat Enterprise Linux (RHEL) 8.4 and above, SUSE Linux Enterprise Server 15 SP3+, and Debian 12+. For the vast majority of applications written in modern programming languages such as C, C++, Rust, Go, Java, Python, .NET Core, Node.js, Ruby, and PHP, no code modifications are required. Developers simply select the appropriate ARM-based Amazon Machine Image (AMI) or deploy multi-architecture container images built for the arm64 architecture.

To further accelerate adoption, AWS provides a suite of migration tools and resources. The AWS Graviton Getting Started Guide offers comprehensive documentation on building, running, and optimizing workloads. For enterprises managing large Java codebases, AWS Transform automates the complex refactoring processes typically associated with migrating legacy applications from x86 architectures to Graviton. Additionally, the Graviton Savings Dashboard enables financial operations (FinOps) teams to continuously monitor and quantify cost efficiencies realized across their cloud estates.

Pricing, Availability, and Economic Implications

At launch, Amazon EC2 R9g and R9gd instances are immediately available in select AWS Regions, including US East (N. Virginia), US East (Ohio), US West (Oregon), and Europe (Frankfurt). Organizations can procure these instances through multiple purchasing models designed to accommodate diverse financial strategies, including On-Demand instances, Savings Plans, Spot Instances, Dedicated Instances, and Dedicated Hosts.

From an economic perspective, the arrival of R9g and R9gd instances addresses the pressing need for enterprise cost optimization. As cloud budgets face increased scrutiny, the ability to extract up to 25 percent better compute performance while consuming less energy allows businesses to lower their total cost of ownership (TCO) without sacrificing performance or scalability. Furthermore, the energy efficiency of the Graviton5 processor directly contributes to corporate sustainability targets, helping enterprises reduce the carbon footprint associated with their cloud computing operations.

Looking ahead, industry analysts project that the general availability of R9g and R9gd instances will accelerate the broader migration of enterprise databases and analytical engines away from legacy x86 infrastructure. As software vendors increasingly optimize their codebases for ARM architectures, the performance delta between custom silicon and merchant hardware is expected to widen further. By combining mathematically proven security via the Nitro Isolation Engine, flexible bandwidth configuration, and generational performance leaps, AWS has reinforced its leadership in custom cloud silicon, setting a formidable benchmark for the entire cloud computing industry.