AI Data Centers Chase 945 TWh Power Demand as Cheap Electricity Overtakes Talent

3 articles · Updated · Silicon Canals · Jul 4

Global data-center developers are picking new AI campuses for cheap, abundant electricity first, as the IEA sees sector power use rising from 415 terawatt-hours in 2024 to about 945 by 2030.
AI is driving the shift because AI-focused server electricity demand is projected to grow about 30% a year, versus roughly 9% for ordinary servers, turning grid capacity into the main site constraint.
Iowa shows the new logic most clearly: wind supplies about 63% of its electricity, and Google, Microsoft, Meta and Apple have helped bring roughly 27 facilities there with land, tax breaks and low-cost power.
Northern Virginia and Ireland show power is necessary but not sufficient. Virginia still handles about 70% of global internet traffic despite shortages, while Ireland's data centers used 22% of electricity in 2024 after a connection freeze.
Nearly two-thirds of new US capacity is now being built outside established hubs, suggesting the next data-center map will follow where new generation comes online fastest.

Sources

Silicon Canals14h ago

AI Boom Shifts Data Center Location Priority from Talent to Cheap Power

windowsforum.com14h ago

AI Data Center Memory Squeeze: Why PC and Phone Upgrades May Cost More | Windows Forum

cyberpeace.org14h ago

Data Centres, Ecological Limits, and the Cost of AI Sovereignty

4 Sources

Will the internet's future be dictated not by silicon, but by access to cheap power and abundant water?

Is AI's power demand a crisis, or the ultimate catalyst for a global clean energy revolution?

U.S. Data Center Power Crisis: AI-Driven Demand to Double Electricity Use by 2030 and Reshape the Grid

Overview

The United States is experiencing a historic surge in electricity demand, mainly due to the explosive growth of data centers and the rapid adoption of AI technologies. While each AI task is becoming more energy efficient, the overall electricity use is rising because of the sheer volume of AI applications and the increasing number of energy-intensive AI agents. As a result, data center electricity consumption is expected to double by 2030, reaching levels comparable to all U.S. residential use. This trend highlights the urgent need for reliable and sustainable power solutions to support the expanding digital economy.

...