AI's Energy Appetite Set to Devour Power Grid

Nvidia CEO Jensen Huang has expressed concern about the power needs of artificial intelligence (AI), warning that computing's energy requirements could be as much as 1,000 times higher than current levels.

Speaking at Stanford University in May, Huang estimated that AI agents running continuously would consume significantly more power than today's search engines. He noted that this could lead to a bottleneck in the power grid, which is already showing signs of strain.

Goldman Sachs has projected that U.S. data center power demand will climb from 31 gigawatts in 2025 to 41 gigawatts in 2026 and 66 gigawatts in 2027, more than doubling in just two years.

Nvidia expects global annual data center capital expenditure to scale toward a $3 trillion to $4 trillion AI infrastructure opportunity over the next five years. Huang's estimate is directional, but Goldman's modeling suggests that the power crunch is real and imminent.