
Amazon Web Services (AWS) has introduced its next-generation Graviton5 processor, delivering up to 25 percent higher compute performance and improved energy efficiency. The new chip powers the EC2 M9g instances, now available in preview, and is built on advanced 3nm technology. With a significantly expanded L3 cache—five times larger than before—and reduced inter-core latency, Graviton5 is engineered for high-throughput workloads. It is further reinforced by AWS’s Nitro System and a newly developed Nitro Isolation Engine designed to offer mathematically proven cloud security.
Several early users have already validated the processor’s gains in real-world environments. Airbnb reported up to a 25 percent increase in performance across its production search workloads. Atlassian also observed impressive improvements, noting that Jira running on M9g instances achieved 30 percent higher performance and 20 percent lower latency. Paulo Almeida from the company stated they look “forward to AWS Graviton5 general availability.” SAP recorded OLTP query performance improvements ranging from 35 to 60 percent, while Siemens saw an additional 30 percent uplift for its Calibre platform, emphasizing the chip’s capabilities across enterprise-grade applications.
AWS highlighted that Graviton-based compute has become a major part of its infrastructure strategy. Over the past three years, more than half of AWS’s new CPU capacity has come from the Graviton family. Adoption among customers is also strong, with 98 percent of the top 1,000 EC2 users now running Graviton-powered workloads in some capacity.
The new processor is already demonstrating benefits in specialized, compute-heavy fields. Synopsys Inc reported up to 35 percent faster electronic design automation (EDA) runtimes when using Graviton5-backed instances. Arm also noted enhancements, observing up to 40 percent faster runtimes for Synopsys VCS. These results highlight Graviton5’s potential impact on industries such as semiconductor design, high-performance computing, and large-scale simulation.
With its blend of higher performance, improved latency, advanced security, and strong energy efficiency, Graviton5 marks a significant step forward in AWS’s custom silicon roadmap. As more enterprises shift to compute-intensive and latency-sensitive workloads, the processor is positioned to become a key driver in optimizing cloud efficiency and accelerating application performance across a broad spectrum of industries.




