AI's Massive Energy Appetite Raises Climate Concerns

Artificial intelligence is extremely energy intensive, with current AI using 1-1.5% of global electricity and future growth potentially increasing demand dramatically.

Both the training and inference phases of AI are very energy-hungry due to the large datasets and billions of parameters involved.

Cooling for AI systems can add significantly to energy costs, up to 50% on top of computing needs.

More transparency and data are needed on where AI systems are located and their energy sources to accurately assess environmental impact.

Without changes, unchecked AI growth could drive up electricity demand 10x in a worst-case scenario, so sustainability should be considered an AI risk factor.