Meta Expands Compute Portfolio with Tens of Millions of AWS Graviton Cores as Agentic AI Emerges as Major CPU and GPU Development

Meta has recently established a significant collaboration with Amazon Web Services (AWS) aimed at integrating tens of millions of Graviton CPU cores into its AI computing infrastructure dedicated to Agentic AI.

Revolutionizing AI Computing: Meta’s Gigantic Step with AWS Graviton CPUs

As the landscape of Agentic AI evolves, businesses are swiftly enhancing their AI infrastructures to tackle the surging demands for computational power. Notably, we’ve observed multi-gigawatt agreements being finalized, signaling a pronounced increase in CPU utilization.

A close-up of an Arm processor labeled 'ALC14C00-AL-A0-8-C-ES' on a blue background.

Amid this technological advancement, Meta has made headlines by announcing a groundbreaking partnership with AWS, focusing on the utilization of AWS’s Graviton CPUs. According to the details revealed, Meta plans to incorporate “tens of millions”of these Graviton CPUs into their computing portfolio. This is a remarkable development considering that each Graviton5 chip comprises 192 Arm Neoverse cores, marking a substantial enhancement in processing capability.

Key Highlights of the Meta-AWS Partnership

Meta is teaming up with AWS to integrate millions of Graviton cores into its computing strategy.
This partnership reflects the evolving computational needs driven by advances in Agentic AI.
The Graviton5 cores are specifically designed to meet these demanding requirements, illustrating the concept that no singular chip architecture can efficiently handle all workloads.

This collaboration positions Meta as one of the foremost Graviton consumers globally, underscoring the crucial role of CPUs in the realm of Agentic AI. Demand for CPUs has surged, prompting inquiries from AI companies to major manufacturers like Intel, AMD, NVIDIA, and Amazon for expanded chip availability.

“This isn’t just about chips; it’s about giving customers the infrastructure foundation, as well as data and inference services, to build AI that understands, anticipates, and scales efficiently to billions of people worldwide. Meta’s expanded partnership, deploying tens of millions of Graviton cores, shows what happens when you combine purposebuilt silicon with the full AWS AI stack to power the next generation of agentic AI.”

– Nafea Bshara, Vice President and Distinguished Engineer, Amazon

Simultaneously, AI companies are developing their own custom silicon tailored for AI applications. Meta has joined forces with Broadcom to create proprietary AI silicon intended to support a “Multi-Gigawatt”ecosystem. Although Meta is already producing multiple MITA series accelerators, supply constraints among chip manufacturers like TSMC and Samsung have led to a strategic pivot towards collaborating with firms that are currently producing chips or engaging with cloud AI service providers.

Another exciting development in our chips business as Meta has decided to bet big on Graviton, our leading CPU chip—committing to tens of millions of Graviton cores. Agentic AI is becoming almost as big a CPU story as a GPU story. Complex multi-step orchestration, real-time… pic.twitter.com/nccGQ6Gw0n

— Andy Jassy (@ajassy) April 24, 2026

Furthermore, the commitment to “tens of millions of cores”represents merely the initial phase of deployment. Meta intends to aggressively expand its AI capabilities in the upcoming years, promising an influx of additional chips as the AI ecosystem grows and matures.

Source & Images