Arm-based custom chips drive growth for Microsoft and Amazon cloud AI infrastructure

Major cloud providers Microsoft and Amazon are increasingly relying on Arm-based custom silicon to support AI infrastructure. Microsoft CEO Satya Nadella noted that CPUs are becoming as critical as GPUs for running agentic AI workloads, which require complex tasks like data preprocessing and tool orchestration. Microsoft is deploying its Azure Cobalt 200 chips across more than 25 data centers globally.

Amazon's chip business, including Graviton, Trainium, and Nitro, has reached an annualized revenue exceeding $25 billion. The company reported that 98% of its top 1,000 EC2 customers use Graviton, with the latest Graviton5 offering significant performance improvements for machine learning inference and real-time reasoning.

In a related development in the AI services market, Meta has emerged as one of Microsoft's largest AI customers. Meta spends hundreds of millions of dollars annually on Microsoft Azure to power its AI needs, including software development, consuming trillions of tokens per week. This highlights a growing trend where major tech firms utilize diverse cloud and model platforms to optimize for cost and availability while developing their own proprietary AI capabilities.