AWS has released the Amazon EC2 M9g and M9gd instances powered by the next-generation AWS Graviton5 processors, delivering up to 25% better compute performance and improved energy efficiency over prior generations. These instances offer substantial gains in database query speed, machine learning throughput, and web application responsiveness, supported by enhanced network and storage bandwidth.
- Up to 25% improved compute performance and 5x larger L3 cache over previous generation
- Supports PCIe Gen6, DDR5 memory, and enhanced network/storage bandwidth for demanding workloads
- M9gd variant offers local NVMe SSD storage for low latency and high I/O throughput
Infrastructure signal
Amazon EC2 M9g and M9gd instances represent a significant evolution in AWS's Arm-based processor lineup, powered by the Graviton5 chip. They deliver up to 25% higher compute performance than Graviton4 counterparts, backed by larger caches and faster memory technologies such as DDR5 at 8800 MT/s. These enhancements enable more efficient processing of concurrent tasks and larger in-memory data sets, particularly benefiting CPU-intensive uses like real-time AI orchestration and database operations.
Beyond raw compute, these instances introduce improved networking capabilities, with up to 15% higher bandwidth and a doubling of network throughput on the largest sizes. Storage throughput to Amazon EBS is also increased by roughly 20%. The M9gd variant includes local NVMe SSD storage for workloads that demand ultra-low latency and high IOPS, expanding options for data-intensive applications. This blend of CPU, memory, storage, and networking improvements bolsters the infrastructure foundation for scalable, performant cloud services.
Developer impact
Developers benefit from out-of-the-box performance advances that require minimal or no code changes. Tests from early adopters show up to 60% faster MySQL query durations and 36% improved throughput per core on observability workloads, enabling faster iteration and reduced time to insights. The wider cache sizes and memory bandwidth translate to better concurrency handling, significant for modern cloud-native and AI-driven applications.
The instances' compatibility with the latest PCIe Gen6 and DD5 memory standards not only supports bandwidth-intensive use cases but also future-proofs deployments for emerging workloads. Integration with AWS Nitro System simplifies deployment and enhances security, while Instance Bandwidth Configuration (IBC) allows developers to flexibly allocate bandwidth between networking and storage needs, optimizing performance based on workload characteristics. Together these factors streamline development workflow and foster innovation.
What teams should watch
Cloud infrastructure and platform teams should evaluate M9g and M9gd instances for workloads spanning databases, web applications, and ML inference. The improved price-performance and energy efficiency of Graviton5 processors enable cost savings and sustainability gains, key metrics in enterprise cloud strategies. Teams running agentic AI or multi-step task orchestration should prioritize testing these instances as their specialized design supports high core counts and low-latency intra-core communication.
Observability, data analytics, and microservices teams will also find benefits from larger cache sizes and expanded memory bandwidth, leading to lower latency and higher throughput. Teams managing deployments with local storage needs should consider M9gd for enhanced I/O performance. Monitoring tools and observability solutions must be updated to track new network and EBS bandwidth profiles as well as local storage characteristics to fully leverage these instance capabilities.