Amazon Web Services (AWS) has unveiled its next generation of AWS-designed chips, Amazon Graviton4 and Amazon Graviton3E, designed specifically for machine learning and artificial intelligence (AI) workloads. These new chips are built on the Arm architecture and are expected to deliver up to 40% better performance and 34% lower cost per inference than the previous generation.

“AWS Graviton4 and Amazon Graviton3E are the result of years of investment in Arm-based chip design and the deep expertise of our engineering teams,” said Bill Vass, Vice President and Chief Technology Officer for Infrastructure Services at AWS. “We believe these new chips will revolutionize how customers build and run their machine learning and AI applications.”

Amazon Graviton4 is designed for high performance and efficiency, making it ideal for training and inference workloads. It features new Arm Cortex-A78 cores, a next-generation Neural Engine for accelerated machine learning, and PCIe Gen4 and ECC RAM support. Amazon Graviton3E, meanwhile, is designed for edge computing applications, offering high performance and efficiency at a lower cost than Amazon Graviton4. It features a new Arm Cortex-A75 cores, a redesigned Neural Engine, and PCIe Gen3 and DDR4 RAM support.

“With the launch of Amazon Graviton4 and Amazon Graviton3E, AWS is further strengthening its position as the leader in Arm-based computing,” said Raj Shah, Vice President of Compute Services at AWS. “These new chips will give customers even more choice and flexibility when running their workloads on AWS.”

The new AWS-designed chips will be available in the first half of 2024. They will be supported by a wide range of AWS services, including Amazon Elastic Compute Cloud (EC2), Amazon SageMaker, and Amazon Lambda.

The launch of Amazon Graviton4 and Amazon Graviton3E is a significant milestone for AWS and the Arm ecosystem. These new chips will help AWS expand its lead in the cloud computing market and make it even more attractive to customers running machine learning and AI workloads.

The adoption of Arm-based chips is growing rapidly as businesses seek more efficient and cost-effective solutions for their computing needs. AWS is well-positioned to capitalize on this trend with its growing Arm-based services and infrastructure portfolio.

The future of cloud computing is bright, and AWS is leading the way with its innovative new chips and its commitment to Arm-based computing.