Untangling GPU Power Consumption: Job-Level Inference in Cloud Shared Settings

Pierre Jacquet, Maxime Agusti, Eddy Caron, Camille Coti, Marcos Dias de Assunção, Laurent Lefèvre, Anne‐Cécile Orgerie · 2026

As the demand for AI-driven workloads increases, the energy consumption of Graphics Processing Units (GPUs) devices has come under intense scrutiny, particularly in hyperscale data centers where large numbers of accelerators are centralized and leased to diverse clients.

Read the paper · More papers on PaperTik