Contribution type: lightning talk and poster
Bottom-Up Simulation of AI Data Center Load Profiles for Energy Infrastructure Planning
Presenter: Roberto Vercellino
Description: The rapid growth of generative artificial intelligence (AI) has introduced unprecedented computational demands, driving significant increases in the energy footprint of data centers. However, existing power consumption data is largely proprietary and reported at varying resolutions, creating challenges for estimating whole-facility energy use and planning infrastructure. In this work, we leverage a novel simulation platform developed at the National Laboratory of the Rockies (NLR – formerly NREL), to develop data-driven representative whole-facility load profiles. These profiles will help discuss key differences in compute workloads, hardware, scale, load profile characteristics and infrastructure requirements between different AI data center archetypes such as training, colocation and inference.
Background: I (Robi Vercellino) have been a research scientist at NREL since 2022. The focus of my work has been energy modeling, optimization and techno-economic analysis of energy systems, with particular attention to behind-the-meter resources and microgrids, electrified transportation and AI data centers. I am relocating to Milano, Italy in August 2026, and am looking forward to introducing myself to OpenMod, and the open-source energy research community in Europe (and beyond).
Optional links: