In a recent study, McKinsey reported that inference operations in artificial intelligence—that is, running models to solve specific problems or analyze data—will account for the lion’s share of energy consumption in data centers in the future. As models grow in size and complexity, energy requirements become very high, driving up operational costs and increasing environmental pressure on companies. The report notes that improving the efficiency of hardware and software, and adopting renewable energy sources, will become essential to reducing the carbon footprint of data centers. With the rapid growth of artificial intelligence across various industries, energy management is becoming a critical factor in economic and technological sustainability, making investment in efficiency improvements a strategic move.
McKenzie: AI inference will consume the largest share of data center energy.
