Updated
Updated · HPCwire · Jun 12
PNNL Unveils Guide for AI Data Centers as U.S. Power Use Could Hit 12% by 2028
Updated
Updated · HPCwire · Jun 12

PNNL Unveils Guide for AI Data Centers as U.S. Power Use Could Hit 12% by 2028

1 articles · Updated · HPCwire · Jun 12

Summary

  • PNNL and industry partners released a new guide for siting, building and operating AI data centers to cut energy costs and help keep the grid stable as electricity demand climbs.
  • U.S. data centers used about 4.4% of national electricity in 2023, and PNNL says that share could reach 12% by 2028 as companies expand AI-driven computing.
  • The guide urges grid-interactive designs, including microgrids, on-site batteries and smart controls that can shift or trim power use when grid demand is high.
  • Cooling alone consumes roughly 20% to 40% of a data center's energy, so the guide highlights tactics such as higher temperature set points and pre-cooling before peak demand.
  • More than 50 partners—including NVIDIA, IBM, Carrier and Siemens—helped build what PNNL says is a dynamic online resource for developers, utilities and policymakers.

Insights

With AI's energy use soaring, can voluntary guides prevent government bans on new data centers?
Can data centers become grid partners before they trigger widespread energy and water shortages?
Is making AI more energy-efficient simply fueling an even greater demand for its use?

AI Data Centers’ Energy Surge: U.S. Electricity Demand to Triple by 2028—Risks, Innovations, and the New PNNL Performance Framework

Overview

The rapid growth of artificial intelligence and high-performance computing is putting huge pressure on data centers and energy grids in the United States, with thousands of new facilities being built. To address this, the Pacific Northwest National Laboratory, together with NEMA and ASHRAE, released the AI Data Center Energy Performance Framework in June 2026. This framework gives stakeholders practical strategies to improve energy efficiency, control costs, and boost reliability throughout the data center lifecycle. It provides clear guidance for every stage of development and operation, helping the industry manage rising energy demands and infrastructure challenges.

...