The E-CALC project (Energy Calculations for Optimizing Machine Learning Lifecycle) is supported by the Defense Advanced Research Projects Agency (DARPA). This project is designed to develop a software system focused on energy efficiency and improving energy consumption tracking in artificial intelligence processes.
Increasing Energy Demand Due to Artificial Intelligence
The rapid growth of artificial intelligence has become a serious challenge for energy resources. It is predicted that data center energy consumption will more than double by 2030, reaching approximately 840 terawatt-hours (TWh), while the consumption of these centers was about 420 TWh in 2024. This increase in demand puts additional pressure on global energy networks.
Read more: GE Vernova Acquires Trench Group for $9 Billion
Collaboration Between the University of Michigan and the University of Illinois
The research team from the University of Michigan is developing this new system in collaboration with the University of Illinois Urbana-Champaign and defense contractor BAE Systems. This project seeks to address the inadequate measurement and reporting issues in artificial intelligence energy consumption. Researchers estimate that between 20 to 30 percent of energy is wasted during the training of artificial intelligence models, which could be reduced through software optimization.
The development of the new system will allow programmers to estimate energy consumption at every stage of the artificial intelligence process, from data processing and model training to real-world implementation. This technology could be particularly significant in military and defense environments where field equipment and computing systems face hardware and energy constraints.
According to Mosarraf Choudhury, an assistant professor at the University of Michigan and the principal investigator of the project, "This project seeks to answer a practical question: how can we leverage the broad benefits of artificial intelligence with minimal energy consumption."
Read more: Hosseyn Afshin Announced Support for Technological Projects to Reduce Energy Consumption · Sembcorp Received Conditional License to Import 1.3 Gigawatts of Electricity from Renewable Energy to Singapore
