A 1 GW AI data center draws roughly 8.76 TWh annually and costs about $876 million in electricity at $0.10/kWh if kept at full load, illustrating the scale of power needs for a single hyperscale facility.

Meta Prometheus is described as a confirmed 1 GW AI cluster; with typical PUE around 1.20, annual consumption sits near 8.76–10.51 TWh and costs range from about $876 million upward.

Even with commitments, consumers’ electricity bills could rise depending on how utility costs are allocated and how effectively companies cover those costs.

In Virginia, data centers accounted for nearly 40% of electricity use in 2024, underscoring regional concentration and potential pressure on local grids and prices.

Industry implications include opportunities for power generators, grid equipment suppliers, data center power management firms, and backup generation providers, with reliable, affordable electricity becoming a competitive edge for AI operators.

All figures here are illustrative calculations or hypothetical scenarios, not specific project disclosures, and depend on utilization, cooling efficiency, and deployment details.

Consumer Reports notes households paid 7.1% more for electricity in 2025, with inflation and aging grid contributing alongside rising AI demand.

Industry responses in 2026 include commitments from Microsoft and Google to finance needed energy and infrastructure under a White House–promoted voluntary agreement, with government measures aiming to prevent cost transfer to general consumers.

The IEA projects global data center electricity use doubling from about 485 TWh in 2025 to around 950 TWh by 2030, signaling surging power demand for AI infrastructure.

Cooling, infrastructure, and PUE affect total draw; for a PUE of 1.20, a 1 GW IT installation would require roughly 1.2 GW from the grid.

Utility systems must add capacity, which can raise costs for all customers as utilities invest in plants, transmission lines, and related infrastructure to support data centers.

Data centers for digital tech, including AI, consume substantial electricity and water, shaping overall energy demand and operating costs.