Data center load flexibility for grid frequency regulation and emissions

Ali Jahanshahi and co-authors have posted an arXiv preprint introducing a new metric and framework for coordinating GPU data center load with grid frequency regulation (arXiv:2601.22487, submitted 30 Jan 2026).

• Main announcement: The paper introduces the Exogenous Carbon metric and the EcoCenter framework to quantify and maximize how GPU data centers can provide frequency regulation, with the stated result that such participation can yield Exogenous carbon savings that oftentimes outweigh Operational carbon emissions.
• Background and details: The submission (arXiv v1, 30 Jan 2026) frames the problem as due to AI/ML data center growth increasing energy use and destabilizing grids that rely on fossil-fueled frequency regulation reserves; it provides methodological contributions (metric + EcoCenter) and demonstrates coordination strategies to reduce the need for fossil-fueled regulation reserves.