Meta Platforms has recently entered a multi-year agreement with Amazon Web Services (AWS) to utilize tens of millions of AWS Graviton CPU cores, marking a significant shift in AI infrastructure strategy. While the AI industry often focuses on high-performance GPUs for training large models, this deal highlights growing demand for CPU capacity to handle agentic AI workloads—tasks that require real-time reasoning and more general-purpose processing power. Meta’s infrastructure head, Santosh Janardhan, emphasized that diversifying compute sources is a strategic imperative as the company scales its AI ambitions.
This move underlines a broader trend in AI infrastructure evolution, where CPUs are gaining prominence alongside GPUs for specific workload types. Despite Meta’s extensive investments in GPU and accelerator chips from companies like Nvidia, AMD, and Google, the AWS Graviton CPU cores are targeted to power intensive CPU tasks behind agentic AI applications. This shift signals the AI market’s expanding complexity and diverse compute requirements.
For investors, this deal offers indirect validation for Intel’s position in the CPU market. As AI workloads increasingly necessitate robust CPU capabilities in addition to GPUs, Intel’s dominance in general-purpose computing hardware remains relevant. The Amazon-Meta agreement thus reflects broader industry dynamics emphasizing heterogeneous compute architectures, with CPUs playing a vital role in innovation and operational efficiency for AI services.
