Amazon EC2 R8i and R8i-flex Instances: Highest Performance and Fastest Memory on AWS

AWS has announced the general availability of its Amazon EC2 R8i and R8i-flex instances, powered by custom Intel Xeon 6 processors. These memory-optimized EC2 instances deliver the highest performance and fastest memory bandwidth of any comparable Intel processors in the cloud.

With up to 20% higher performance, 15% better price performance, and 2.5x greater memory throughput compared to the previous R7i generation, these new instances are designed for memory-intensive workloads such as SQL/NoSQL databases, in-memory analytics, and enterprise SAP applications.

‍

Overview of Amazon EC2 R8i and R8i-flex

Both R8i and R8i-flex instances are designed for workloads that demand high memory bandwidth and low latency.

• R8i instances: Ideal for demanding, mission-critical workloads that need sustained high CPU, network, and storage throughput.
• R8i-flex instances: More affordable versions of R8i, offering 5% lower pricing and better price performance for workloads that don’t always require maximum compute utilisation.

Learn more about EC2 instance families and their use cases.

‍

R8i vs R8i-flex: Choosing the Right Instance

When to use R8i:

• Enterprise applications needing scalability up to 384 vCPUs and 3TB memory.
• SAP HANA workloads, certified with the highest aSAPS benchmark in the cloud.
• Analytics, AI models, and databases requiring consistent high throughput.

When to use R8i-flex:

• Workloads that don’t fully utilise compute resources 100% of the time.
• In-memory databases (Redis, Memcached), real-time analytics, and mid-sized enterprise apps.
• A first choice for businesses balancing cost and performance.

Explore more about AWS EC2 pricing models to optimise cost efficiency.

‍

The specs for the R8i instances are as follows.

The specs for the R8i-flex instances are as follows.

‍

Key Benefits of R8i and R8i-flex

• 20% faster performance compared to R7i
• 30% faster PostgreSQL, 60% faster NGINX, and 40% faster AI recommendation models
• Scales up to 384 vCPUs & 3TB RAM for database applications
• SAP-certified with best-in-class aSAPS benchmarks
• Built on AWS Nitro System, delivering 2x more network & storage bandwidth
• Configurable bandwidth allocation between EBS and network for database tuning

‍

Real-World Use Cases

High-Performance Databases

Run relational and non-relational databases such as PostgreSQL, MySQL, and MongoDB with improved query performance and scalability.

SAP HANA Workloads

R8i instances are certified for SAP HANA, offering unmatched resilience for large-scale enterprise ERP systems. Learn more about running SAP on AWS.

In-Memory Analytics

Ideal for Redis, Memcached, and Apache Spark clusters that demand low-latency memory access.

AI and ML Applications

With Intel AMX FP16 support, R8i delivers accelerated performance for deep learning training, inference, and recommendation models.

‍

How D55 Helps with AWS EC2 R8i Migration

At D55, we specialise in helping SMEs and mid-market organisations migrate, modernise, and optimise workloads on AWS.

With the new R8i generation, we can help you:

• Run an AWS-funded migration assessment to see if R8i is right for your workloads.
• Optimise costs with AWS Savings Plans and FinOps best practices.
• Modernise applications by moving from EC2 to serverless (AWS Lambda) or Aurora/DynamoDB for long-term efficiency.

Learn more about our Cloud Migration Services and Application Modernisation.

Final Thoughts

The new Amazon EC2 R8i and R8i-flex instances deliver a step change in performance and scalability for memory-intensive workloads. With up to 20% better performance and 2.5x memory throughput, they enable enterprises to handle more demanding applications while keeping costs under control.

If your organisation is considering a move to AWS or looking to modernise existing workloads, the R8i generation provides the foundation for future innovation.

Book a free discovery session with D55 to explore how R8i and R8i-flex can support your cloud journey.

‍