Key Points
- Meta to deploy tens of millions of AWS Graviton processor cores for AI workloads
- Graviton5 chip features 192 cores and cache five times larger than previous generation
- Deal focuses on CPU-intensive agentic AI tasks including reasoning and code generation
Meta has signed an agreement with Amazon Web Services to deploy tens of millions of Graviton processor cores, marking a significant expansion of its artificial intelligence infrastructure.
The deal, announced on Friday, will see Meta use AWS‘s custom-designed Graviton5 chips to power workloads related to agentic AI. These are autonomous systems capable of reasoning, planning and completing complex tasks without continuous human input.
EVENT
Infosec Reimagined
Infosec Reimagined 2026 is the premier information security summit where top leaders—CISOs, CROs, CIOs, CTOs and risk executives—converge to redefine cyber resilience.
EVENT
CIO Prism
CIO Prism unites forward-thinking technology leaders to exchange transformative insights, shape digital strategies, and foster innovation, empowering enterprises to excel in an era of rapid technological change.
EVENT
Cyber Surakshit Uttar Pradesh
Find out strategies, frameworks and solutions for building a resilient and secure digital ecosystem across Uttar Pradesh.
The partnership reflects a shift in how large-scale AI systems are being built. While graphics processing units remain essential for training large language models, running AI agents that can execute multi-step tasks in real time requires substantial central processing unit capacity.
Tasks such as code generation, real-time reasoning, search queries and coordinating complex workflows are CPU-intensive. Graviton processors are designed specifically for these workloads, offering the processing power needed to handle billions of user interactions.
“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,” said Nafea Bshara, Vice President and Distinguished Engineer at Amazon.
Technical specifications of Graviton5
The Graviton5 chip features 192 processing cores and a cache, the high-speed memory that processors use to store frequently accessed data, that is five times larger than the previous generation. According to AWS, this reduces communication delays between cores by up to 33 per cent.
The chips are built on 3-nanometre technology, a manufacturing process that produces smaller and more power-efficient processors. AWS designs these chips in-house and controls the full process from chip design through server architecture.
The processors run on the AWS Nitro System, a combination of dedicated hardware and software that provides direct access to the underlying hardware while maintaining security. This allows Meta to run its own virtual machines, software that simulates a complete computer system, without performance compromises.
Advertisement
The Graviton5 instances also support the Elastic Fabric Adapter, which enables low-latency communication between multiple processors. This is critical for AI workloads where large tasks must be distributed across many processors working simultaneously.
Meta’s infrastructure strategy
Santosh Janardhan, Head of Infrastructure at Meta, said diversifying compute sources was a strategic priority as the company scales its AI ambitions.
“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,” Janardhan said.
Meta has been a longstanding AWS customer, using the cloud provider’s infrastructure to support its global services. The Graviton deployment starts with tens of millions of cores, with the flexibility to expand as Meta’s AI capabilities grow.
Your Questions, Answered
What is the Meta and AWS Graviton deal about?
Meta has signed an agreement to deploy tens of millions of AWS Graviton processor cores to power its AI workloads, particularly for agentic AI systems that handle tasks like reasoning, code generation and search.
What is agentic AI?
Agentic AI refers to autonomous systems that can reason, plan and complete complex tasks without continuous human input. These systems require significant CPU processing power to coordinate multi-step workflows.
What are the specifications of Graviton5 chips?
Graviton5 features 192 processing cores and a cache five times larger than the previous generation. It is built on 3-nanometre technology and reduces core communication delays by up to 33 per cent.
Why is Meta using Graviton instead of GPUs for AI?
While GPUs remain essential for training AI models, running agentic AI workloads like real-time reasoning and code generation is CPU-intensive. Graviton processors are purpose-built for these specific tasks.