AI is the new cause of power outages

(Dan Tri) - In addition to the usual causes of power outages, a new factor has recently been discovered that is also causing more power outages. That is artificial intelligence.

Sudden, intermittent power outages or skyrocketing electricity bills are no longer just the result of storms, infrastructure failures or spikes in demand.

A new cause is emerging, putting huge pressure on the grid and directly affecting consumers: the explosion of Artificial Intelligence (AI).

According to data from the US Energy Information Administration, the average price of electricity in the US has increased by 6.5% in one year, from 16.41 cents/kilowatt-hour in May 2024 to 17.47 cents/kilowatt-hour in May 2025.

This is a significant increase in the history of the US electricity industry, reflecting not only rising costs but also showing that energy consumption is at a record high.

In addition to people using more energy, a Bloomberg report points out that AI is a contributing factor, not only leading to power outages but also worsening the quality of residential electricity.

How does AI put pressure on the power grid?

AI data centers could increase pressure on the power grid (photo: Imaginima).

AI-enabled data centers are placing a significant strain on the power grid, experts say.

Thanks to Ting devices – smart sensors that monitor electricity quality in homes – researchers discovered this problem.

The Ting device, originally designed to prevent electrical fires, also measures “bad harmonics” (total harmonic distortion), which reduce energy efficiency and can cause damage to household electronics.

Bad harmonics occur when the normal current is interrupted, creating spikes (voltage surges) and dips. In contrast, good harmonics are a continuous and balanced voltage wave.

The worst power outages typically occur near large data centers, with more than 75% of these outages occurring within about 50 miles, affecting nearly 3.7 million Americans.

Areas like Northern Virginia and Chicago, home to a large concentration of data centers, are seeing the most severe distortions.

A data center in Sterling, Virginia, USA (photo: Gerville).

Unlike the steady and predictable increase in electricity demand due to population growth, the construction of new data centers can put almost immediate pressure on local infrastructure, causing sudden, exponential increases in electricity demand.

However, experts also suggest that optimizing grid structure using AI could be an economical and effective solution to improve efficiency.

The Importance of Power Quality and Solutions

Poor power quality can not only damage electronic equipment, but can also lead to malfunctions, overheating, and the risk of electrical fires. This is especially concerning given the increasing number of forest fires, factory and house fires in many places.

To address this challenge, a number of measures are being considered and implemented. Data centers are being asked to build dedicated substations and transformers to handle the increased power and protect household circuits.

Several players in the power industry are also working to develop dedicated modular systems to produce clean energy that directly serves data centers.

However, depending on how the situation develops, other long-term solutions may be needed. Increasing the exploitation of clean, renewable energy sources for electricity production is seen as a positive and sustainable solution for the future.

Source: https://dantri.com.vn/khoa-hoc/ai-la-nguyen-nhan-moi-gay-ra-mat-dien-20250930004139127.htm