Artificial Intelligence Waiting for Green Electricity
Computationally intensive AI can consume up to 20 times more electricity under load than when idle. What if we were to schedule such computational tasks specifically for times when sufficient green energy is available? That is exactly what we are researching in the PRO-LAB project as part of the interdisciplinary energy research laboratory at the University of Bremen.
The goal: an intelligent control system that uses AI to perform tasks precisely when there is a surplus of renewable energy—such as that generated by the photovoltaic systems on-site. To achieve this, the system continuously aggregates energy data, information on cooling and ventilation, and weather data, and schedules computing processes accordingly. This helps avoid peak loads, lower energy demand, and reduce CO₂ emissions.
Our Green IT data center serves as the foundation for this initiative. The insights gained are intended to increase the efficiency of computing infrastructures in the future and thereby contribute to the climate goals of the State of Bremen.
The project is co-financed by the European Union and supported by the Senator for the Environment, Climate, and Science and the BAB – Bremer Aufbau-Bank.
More on this: https://brm.de/projekt-pro-lab/









