Amazon Web Services is telling its own engineers to shut down idle servers after finding most of its virtual machines barely get used. The push comes as AI workloads strain the cloud giant's data centers and drive up prices for machine-learning compute.
Amazon Web Services is asking its own engineers to turn off computers they aren't using. Roughly 65% of EC2 instances run at average CPU utilization below 20% over 30-day measurement windows, the company found, even as AI workloads strain its data centers.
A tool built to catch ghost servers
AWS upgraded its Compute Optimizer tool to catch the waste. It now analyzes CPU utilization and network I/O over 14-day lookback periods, flagging instances where peak CPU stays below 5% and network traffic is negligible. The directive targets idle EC2 instances, the virtual servers that form the backbone of AWS's cloud infrastructure.
AI demand breaks the pricing model
GPU-accelerated instances, the machines that train and run large language models, have become the scarcest resource in cloud computing. EC2 Capacity Blocks for machine learning rose roughly 15% in January, then climbed about 20% more by July, two price increases in six months for the same service. The persistent GPU shortage has reportedly pushed some AWS customers to consider shifting operations to competing cloud providers.
Amazon builds capacity even as it cuts waste
Even as it squeezes more use out of existing servers, Amazon keeps expanding. Amazon CEO Andy Jassy disclosed that AWS added over 3.8 gigawatts of power capacity in the previous year — enough electricity to power roughly 2.8 million homes. The company plans to double that capacity by 2027.
The 65% underutilization figure isn't only an AWS problem, however. It reflects industry-wide habits of over-provisioning, because spinning up a new instance has always been easier than right-sizing an existing one.
Source: Crypto Briefing
Trading involves risk.
Risk Disclaimer
