Meta has entered into an agreement with Amazon Web Services (AWS) to deploy AWS Graviton processors at scale, aiming to support its next generation of AI workloads.
The company will initially use tens of millions of Graviton cores, with the ability to increase deployment as AI demands grow.
Go deeper with GlobalData
Access deeper industry intelligence
Experience unmatched clarity with a single platform that combines unique data, AI, and human expertise.
This move expands the longstanding partnership between Meta and AWS. According to Amazon, it reflects the growing importance of CPU-intensive infrastructure for agentic AI, which requires real-time reasoning, code generation, complex search processes, and task orchestration.
Meta will rely on Graviton5 chips to power its agentic AI operations, which necessitate infrastructure capable of handling billions of interactions and coordinating multi-step tasks.
By adopting Graviton5, the Facebook owner aims to meet the demands of code generation, reasoning, and model training with a platform designed for high compute output.
Meta infrastructure head Santosh Janardhan said: “AWS has been a trusted cloud partner for years, and expanding to Graviton allows us to run the CPU-intensive workloads behind agentic AI with the performance and efficiency we need at our scale.”
The Graviton5 processor includes 192 cores and a cache five times larger than its predecessor, which AWS claims reduces core communication delays by up to 33%.
The chips operate on the AWS Nitro System, which enables direct access to hardware while maintaining standard AWS services such as Elastic Network Adapter and Elastic Block Store.
Graviton5 instances also use the Elastic Fabric Adapter, which allows for low-latency, high-bandwidth communication between distributed AI workloads, a technical requirement for complex, large-scale coordination.
The chips are developed entirely in-house by AWS and are said to deliver up to 25% better performance than previous models and maintain a focus on energy efficiency. They are produced using 3-nanometre chip technology, a process expected to yield improvements in computing efficiency and lower energy use.
Amazon vice president and distinguished engineer Nafea Bshara said: “Meta’s expanded partnership, deploying tens of millions of Graviton cores, shows what happens when you combine purpose-built silicon with the full AWS AI stack to power the next generation of agentic AI.”
The partnership builds on Meta’s existing use of AWS services, including Amazon Bedrock, and forms part of its broader strategy to equip its infrastructure for emerging AI scenarios and agentic systems.
