Can a building anticipate its next cooling peak? Markham Hislop talks with PhD researcher Guaraang Malik about using machine learning to predict building energy consumption and inform heating and cooling controls.

Malik explains how simulations, weather and operating schedules support forecasting, and how cooling empty rooms earlier could reduce peak demand. They also explore running AI on servers inside buildings to retain greater control over data.

Malik distinguishes that proposal from his untested hypothesis about reducing the energy impact of cloud computing. His research remains at the simulation stage, with access to operational building data an important next step ahead.