← Back to feed

Energy and data centers: AI’s physical cost

Electric demand from training and inference clusters strained grids, PPAs, and climate agendas.

Energy and data centers: AI’s physical cost

In 2024–2025, utilities and hyperscalers reported demand spikes tied to GPUs. Small nuclear, gas, and renewables projects were tied to AI campuses.

Chip and algorithm efficiency (MoE, quantization) became a climate and financial variable. Local communities debated water and land use.

AI history stopped being only software: it is also energy geography.