AWS Graviton5 is now generally available, delivering purpose-built performance for the agentic AI era

Tags

Key points

• AWS Graviton5 is now generally available, providing significant performance boosts (up to 25% faster overall compute, 35% faster ML inference) specifically optimized for agentic AI workloads.
• The new M9g and M9gd instances, powered by Graviton5, introduce advanced hardware features like 192 cores, 5x larger L3 cache, DDR5-8800, and PCIe Gen 6, alongside the formally verified Nitro Isolation Engine for enhanced security.
• Major enterprises including Meta, Uber, Snowflake, Adobe, Epic Games, Pinterest, Airbnb, Atlassian, Siemens, and SAP are already leveraging Graviton processors, with many reporting substantial performance gains and cost reductions.

Notable quotes

AWS Graviton5-based Amazon EC2 instances are some of the fastest EC2 instances we have ever tested.

— Denis Sheahan

In our performance tests, conducted using Airbnb's production search workloads, we are seeing improvements of up to 25% over other system architectures of the same generation, and up to 20% compared to prior generation Graviton4 instances.

— Denis Sheahan

In our testing of Jira on AWS Graviton5-based M9g instances, we observed 30% higher performance and 20% lower latency compared to the prior generation, and we look forward to AWS Graviton5 general availability.

— Paulo Almeida

The future of semiconductor physical verification lies in cloud-enabled, high-performance computing.

— Juan Rey

With AWS Graviton5-based Amazon EC2 M9g instances, we've observed a stunning 35% to 60% increase in the performance of our OLTP queries on SAP HANA Cloud-a phenomenal advancement in a single generation.

— Stefan Bäuerle

Structured claims — 79

1. 1
   Monitor product availability for procurement and deployment planning.

   Entities: AWS, Graviton5
2. 2
   Benchmark Graviton5's performance against other solutions for agentic AI workloads.

   Entities: Graviton5, artificial_intelligence
3. 3
   Compare Graviton5's performance metrics against previous generations for upgrade planning.

   Entities: Graviton5
4. 4
   Evaluate Graviton5's suitability for specific AI applications requiring real-time decision-making.

   Entities: Graviton5, artificial_intelligence
5. 5
   Assess Graviton5's core density for high-concurrency workload optimization.

   Entities: Graviton5
6. 6
   Model the impact of reduced inter-core latency on performance-sensitive applications.

   Entities: Graviton5
7. 7
   Benchmark general application performance on Graviton5 for potential migration benefits.

   Entities: Graviton5
8. 8
   Evaluate Graviton5 for machine learning inference workloads to improve throughput and reduce latency.

   Entities: Graviton5, inference_optimization
9. 9
   Assess Graviton5's impact on database performance for critical data management systems.

   Entities: Graviton5
10. 10
    Track the availability of new EC2 instance types for cloud infrastructure planning.

    Entities: AWS, Graviton5, Amazon EC2 M9g, Amazon EC2 M9gd
11. 11
    Note the product development timeline for future release expectations.

    Entities: Graviton5, AWS
12. 12
    Evaluate Graviton5's specialized capabilities for advanced AI development and deployment.

    Entities: Graviton5, artificial_intelligence
13. 13
    Consider Graviton for workloads requiring high concurrency and efficient accelerator utilization.

    Entities: AWS Graviton
14. 14
    Understand the core value proposition of Graviton for cloud cost and efficiency optimization.

    Entities: Amazon, AWS Graviton, cloud_computing
15. 15
    Monitor large-scale adoption of Graviton by major tech companies as an industry trend indicator.

    Entities: Meta, AWS Graviton, artificial_intelligence
16. 16
    Track Graviton adoption by key industry players for competitive analysis.

    Entities: Uber, Snowflake, AWS Graviton
17. 17
    Assess the market penetration and adoption rate of Graviton processors.

    Entities: AWS Graviton
18. 18
    Compare core count specifications for hardware selection and workload distribution.

    Entities: Graviton5
19. 19
    Evaluate the impact of increased cache size on memory-intensive applications.

    Entities: Graviton5
20. 20
    Benchmark memory performance for applications requiring high bandwidth.

    Entities: Graviton5, cloud_computing
21. 21
    Consider PCIe Gen 6 support for high-speed peripheral connectivity and data transfer.

    Entities: Graviton5
22. 22
    Compare M9g instance performance against prior generations for upgrade justification.

    Entities: Amazon EC2 M9g
23. 23
    Evaluate M9g instances for web application hosting to improve user experience and scalability.

    Entities: Amazon EC2 M9g
24. 24
    Benchmark M9g instances for ML inference workloads to optimize model deployment.

    Entities: Amazon EC2 M9g, inference_optimization
25. 25
    Assess M9g instances for database performance improvements in transactional systems.

    Entities: Amazon EC2 M9g
26. 26
    Identify M9gd instances for applications with demanding I/O requirements.

    Entities: Amazon EC2 M9gd
27. 27
    Compare storage capacity and type for instance selection based on data volume needs.

    Entities: Amazon EC2 M9gd
28. 28
    Benchmark M9gd instances for I/O intensive workloads to improve data processing speed.

    Entities: Amazon EC2 M9gd
29. 29
    Understand the underlying architecture for security and performance characteristics.

    Entities: Amazon EC2 M9g, Amazon EC2 M9gd, AWS Nitro System
30. 30
    Track new security features in cloud infrastructure for compliance and risk assessment.

    Entities: AWS Nitro System, Nitro Isolation Engine

Source

Source

amazon-blog

Record title

AWS Graviton5 is now generally available, delivering purpose-built performance for the agentic AI era

Author

Isaac Schultz

Published

Jun 10, 2026

URL

https://aboutamazon.com/news/aws/aws-graviton-5-cpu-amazon-ec2

Manifest ID

1781115440243990944

Significance

high

Sentiment

positive