Energy Demands of AI Could Eclipse Those of Entire Countries, Warns Expert

A recent commentary by Alex de Vries, founder of Digiconomist, has raised concerns about the energy footprint of AI. De Vries predicts that the energy consumption related to AI will significantly increase in the coming years. For instance, models like OpenAI's ChatGPT require extensive data for training, resulting in high energy requirements. The energy usage doesn't stop at training, as active data generation also draws substantial power. Google's integration of generative AI into its services and the anticipated increase in AI server production could lead to a surge in electrical consumption associated with AI, reaching up to 134 TWh annually by 2027.