Meta just signed an agreement to deploy AWS Graviton processors across its operations. This deal is a major step in the long-standing partnership between Meta and Amazon Web Services. Meta now stands as one of the world’s largest customers for these chips. The initial rollout includes tens of millions of cores. The contract also allows Meta to increase this capacity as its AI needs grow.

Powering Next-Generation Agentic Workflows

GPUs are still the standard for training big AI models. However, agentic AI creates new technical challenges. These systems need serious processing power for tasks like real-time reasoning and code generation. They also require high performance to search for information and manage multi-step tasks.

AWS built the Graviton5 chip to handle exactly these CPU-heavy workloads. This chip gives Meta the efficiency it needs to handle billions of interactions. It also helps Meta support its next generation of AI tools. This deal shows a clear change in how AI infrastructure is built. Meta is moving beyond basic hardware setups to use processors designed for complex, multi-step workflows. Meta’s shift to Graviton5 directly addresses the CPU-intensive demands of reasoning, searching, and task orchestration. 

By integrating this hardware, Meta obtains the necessary processing power needed for their agent processes. This will enable Meta to have the necessary infrastructure capacity to meet its changing needs for generative AI technology. Both organizations are leveraging this technological partnership to scale Meta’s AI engagements.

Explore IT Tech News for the latest advancements in Information Technology & insightful updates from industry experts!

News Source: Businesswire.com 

Information Technology

View All

The New CIO Imperative explores how CIOs are redefining their role in the agentic AI era—moving beyond experimentation to driving measurable business outcomes. Based on research from over 1,000 global

Streamline the customer experience from start to finish with generative AI. GenAI-powered Now Assist makes it easier than ever for customers to self-solve, empowers your agents to quickly resolve complex

As a leader in tech, how can you spend less time dealing with everyday operational issues, and more time thinking about how to deliver productivity, efficiency, and better experiences for