AI Data Center Power Costs: Who Is Really Paying the Bill?

The rapid expansion of artificial intelligence is driving major technology companies such as Amazon, Microsoft, Google and Meta to build massive data centers, but the electricity needed to power these facilities is becoming a growing economic and policy issue. As energy demand surges, regulators and utilities are debating who should ultimately pay the rising cost of power infrastructure.

A single hyperscale AI data center can consume as much electricity as roughly 100,000 households. According to estimates cited by energy researchers and international agencies, data centers could account for nearly half of the growth in U.S. electricity demand by 2030. Meeting that demand will require billions of dollars in new generation capacity, transmission lines and grid upgrades.

Utilities and policymakers are increasingly exploring ways to allocate those costs. Some regions are proposing special electricity tariffs or long‑term contracts requiring data center operators to pay for a large share of the grid infrastructure built to serve them. The goal is to prevent traditional customers from subsidizing the rapid expansion of AI infrastructure.

Technology companies argue they already pay for the electricity they consume and for their portion of grid upgrades. However, the ongoing AI investment boom means electricity demand is expected to keep rising sharply, leaving regulators, energy companies and communities grappling with how to balance economic growth with fair energy pricing.