Reducing data center energy use

Data centers in the United States currently account for nearly 5% of the nation's electricity consumption, a figure projected to more than double to 12% by 2030. A substantial portion, approximately 40%, of this energy is utilized by cooling systems essential for maintaining operational temperatures, particularly for high-density computing and artificial intelligence applications that generate significant heat. In response to this growing energy demand, researchers at Penn State are developing innovative software designed to enhance the efficiency of data center cooling. This software employs a physics-based AI learning model that analyzes real-time climate data, such as external temperatures and humidity, alongside economic factors like fluctuating electricity prices, to dynamically optimize cooling parameters. The developers anticipate that this technology could slash cooling system energy usage by 25%. While initially geared towards data centers, including the more than 100 commercial-level Bitcoin mining facilities in the U.S., the Penn State model also holds potential for adaptation to improve cooling efficiency in other commercial buildings.