Meta, a top artificial intelligence developer, has collaborated with Amazon Web Services (AWS) to launch tens of millions of AWS Graviton processor cores. This deployment is a significant step in the evolution of the two technology giants’ continuous cooperation. Also, it makes Meta one of the world’s largest consumers of Graviton’s custom-built silicon. The company is making a dramatic infrastructural change as it moves toward “Agentic AI” autonomous systems with the ability to reason and carry out multiple steps. Training of AI with the use of GPUs is still the main method; however, with the emergence of sophisticated agents, the demand for CPU-intensive tasks like real-time reasoning, search orchestration, and code generation has reached an all-time high.
The Strategic Shift to Agentic Workloads
The implementation will be based on the Graviton5 chip that boasts a remarkable number of 192 cores, as well as the largest cache five times bigger compared to the previous models. The above-mentioned specifications are created specifically to minimize latency between the cores by 33% and provide a high-bandwidth setting needed for processing and acting as a human.
“As we scale the infrastructure behind Meta’s AI ambitions, diversifying our compute sources is a strategic imperative,” said Santosh Janardhan, Meta’s Head of Infrastructure. “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.”
Also Read: OpenAI Elevates Team Productivity with the Launch of Workspace Agents in ChatGPT
Performance Meets Sustainability
With its use of Graviton5, Meta is targeting a 25% boost in performance over its previous CPU generations but at the same time, is making sure that it does not go against its firm sustainability policies. Thanks to the outstanding energy efficiency of Arm-based Graviton chips, Meta is able to run the vast energy hungry AI applications at a globe scale, without neglecting its environmental commitments.
“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. “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.”
A Multi-Faceted Infrastructure Vision
Meta’s Graviton agreement is a continuation of its silicon investment, which also includes its own in-house project MTIA and the collaboration with Arm to build next-generation CPUs for data centers. The combination of both initiatives provides the strong hardware platform, which is able to sustain billions of interactions through the Meta Family of Apps.
With the help of such flexibility, Meta will be able to scale up the use of Graviton processors in line with its growing AI competences during 2026 and beyond.