AWS Unveils Next-Generation Amazon EC2 R9g and R9gd Instances Powered by Graviton5 Processors

SEATTLE — Amazon Web Services (AWS), an Amazon.com company, announced the general availability of the new Amazon EC2 R9g and R9gd instance families. Powered by the cutting-edge AWS Graviton5 processors—touted as the most energy-efficient silicon ever designed by the cloud giant—these memory-optimized instances deliver a staggering 25% performance uplift over their predecessor, the Graviton4-based R8g instances.

The launch marks a significant milestone in custom cloud hardware development. As enterprises face escalating demand for high-performance computing alongside mounting pressure to curb energy consumption and infrastructure costs, the R9g and R9gd series arrive as a powerful antidote. Offering superior compute efficiency, elevated network bandwidth, and rigorous security features rooted in formal verification, these instances are engineered to tackle the most demanding memory-bound workloads in the modern data ecosystem.

Main Facts: Architecture and Capabilities

The newly minted R9g and R9gd instances are purpose-built for memory-intensive applications. While both instance types share foundational compute and networking architectures, the R9gd variants distinguish themselves by incorporating high-speed, local Non-Volatile Memory Express (NVMe)-based SSD block-level storage. This local scratch space makes the R9gd line exceptionally well-suited for distributed real-time big data analytics, open-source databases, and high-throughput caching engines that demand ultra-low-latency disk I/O.

At the heart of these instances lies the AWS Graviton5 processor. Building upon the iterative successes of Graviton3 and Graviton4, the fifth-generation chip introduces several architectural enhancements:

• Higher vCPU Performance: Delivers up to 25% better compute performance per vCPU compared to the R8g line.
• Expanded L3 Cache: A significantly larger L3 cache accelerates data retrieval and reduces memory latency.
• Faster Memory and Bandwidth: Enhanced memory throughput couples with higher network and Amazon Elastic Block Store (EBS) bandwidth limits.
• Energy Efficiency: Delivers all performance improvements while operating at a reduced energy envelope, aligning with corporate sustainability initiatives.

Furthermore, both instance families feature Instance Bandwidth Configuration (IBC). This capability empowers cloud architects to dynamically adjust the allocation of network bandwidth between Amazon EBS and Amazon Virtual Private Cloud (Amazon VPC) traffic by up to 25%. Such fine-grained control ensures that applications with unique throughput profiles—such as heavily transacted relational databases or massive distributed caches—can optimize their data pathways to prevent bottlenecks.

Chronology: The Evolution to Graviton5

The journey toward the R9g and R9gd release represents years of iterative semiconductor engineering and cloud architecture refinement by AWS’s internal silicon teams.

• The Early Foundation: AWS disrupted the cloud infrastructure market years ago by introducing its custom Arm-based Graviton processors, challenging the historical duopoly of x86 architecture in enterprise data centers.
• Graviton4 and the R8g Era: The introduction of Graviton4 and the subsequent R8g instances established a high bar for memory-optimized workloads, offering strong price-to-performance ratios and early integrations with the AWS Nitro System.
• Early 2025 (The Nitro Isolation Leap): Earlier in the year, AWS rolled out advanced compute (C9g) and memory (M9g) instances integrated with the Nitro Isolation Engine (NIE). This introduced formally verified hypervisor security, creating a cryptographic and mathematical baseline for tenant isolation.
• Today (General Availability of R9g/R9gd): AWS officially commercializes the R9g and R9gd lineup, bringing Graviton5 performance paired with the Nitro Isolation Engine to memory-optimized workloads globally, initially launching across major US and European cloud regions.

Supporting Data: Instance Specifications

To accommodate diverse enterprise scaling requirements, both the R9g (EBS-Only) and R9gd (NVMe SSD storage) lines are available in 11 distinct sizes, ranging from agile medium instances up to bare-metal metal-48xl configurations sporting 192 vCPUs and 1,536 GiB of memory.

Amazon EC2 R9g Specifications (EBS-Only)

Amazon EC2 R9gd Specifications (With Local NVMe SSD Storage)

Official Responses and Security Focus: The Nitro Isolation Engine

In technical statements accompanying the launch, AWS emphasized that hardware performance is only part of the value equation; security isolation remains paramount, especially in multi-tenant cloud environments.

All R9g and R9gd instances are anchored by the AWS Nitro System, which offloads virtualization, storage, and networking overhead to dedicated hardware accelerators. This architecture provides applications with near-bare-metal performance while enforcing strict security boundaries.

Crucially, the new instances leverage the Nitro Isolation Engine (NIE). Daniel Abib, speaking on behalf of the AWS engineering team, highlighted how NIE fundamentally shifts cloud hypervisor security:

"The Nitro Isolation Engine enforces isolation between virtual machines by mediating all access to virtual machine memory, CPU register state, and I/O devices through a minimal set of APIs. By leveraging formal verification—a mathematical technique to prove that hardware and software behave precisely as intended under all possible states, rather than just in targeted test cases—NIE establishes the AWS Nitro System as the first formally verified cloud hypervisor."

This mathematical certainty minimizes attack surfaces, offering enterprise compliance officers and security auditors unprecedented peace of mind when migrating sensitive, mission-critical databases and microservices to the cloud.

Implications: Workload Optimization, Migration, and Market Impact

The deployment of R9g and R9gd instances carries broad implications for software engineering teams, system administrators, and corporate IT budgets.

1. Seamless Migration Paths

For organizations currently utilizing R8g instances, transitioning to R9g is frictionless. Most applications require zero code modifications. Because the underlying architecture remains Arm-based, multi-architecture container images built for arm64 run natively. Teams managing containerized applications via Amazon Elastic Kubernetes Service (EKS), Amazon Elastic Container Service (ECS), or standard Kubernetes deployments can drop in R9g instances to instantly capture performance gains. Furthermore, AWS supports a vast spectrum of operating systems out-of-the-box, including Amazon Linux 2023, Ubuntu 22.04+, RHEL 8.4+, SUSE Linux Enterprise Server 15 SP3+, and Debian 12+.

2. Broad Workload Suitability

The memory-optimized nature of the instances makes them ideally suited for:

• In-Memory Caching & Databases: High-performance deployments of Redis, Valkey, and Memcached.
• Relational and Open-Source Databases: Heavy transactional workloads requiring rapid buffer pool management and low disk-latency scratch space.
• Big Data Analytics: Distributed real-time analytics frameworks processing massive streaming datasets.
• Modern Programming Ecosystems: Optimized compilation and runtime execution for applications written in C/C++, Rust, Go, Java, Python, .NET Core, Node.js, Ruby, and PHP.

3. Tooling and Automation

To ease the transition from legacy x86 architectures or older Graviton iterations, AWS provides robust supporting tools:

• AWS Graviton Getting Started Guide: Comprehensive documentation for building and profiling workloads.
• Graviton Savings Dashboard: A dedicated utility for monitoring fiscal optimization.
• AWS Transform: An automated migration tool designed to streamline the porting of Java applications from x86 environments to Graviton silicon.

Availability and Pricing

At launch, Amazon EC2 R9g and R9gd instances are immediately accessible in select AWS Regions, including US East (N. Virginia, Ohio), US West (Oregon), and Europe (Frankfurt).

Customers can procure the instances through flexible purchasing models tailored to operational needs, including Savings Plans, On-Demand, Spot Instances, Dedicated Instances, and Dedicated Hosts. Developers can begin provisioning instances directly via the Amazon EC2 Console or interact with programmatic deployment tools, supported by the newly introduced AWS MCP Server and plugin integrations for AI-assisted cloud management.