Skip to content
Submit incident
AII-1000-28393
ConfirmedA court, a regulator or the organization itself has stated the core facts, or the record cites a court or regulator source.

AI Data Centers Strain US Power Grid During Heat Wave

Rapid expansion of AI data centers in the U.S. mid-Atlantic region has created severe strain on the electrical grid, forcing authorities to approve emergency backup diesel power during extreme heat. Residents face a dilemma between blackout risks and air pollution from generators that mostly lack environmental controls.

1 July 2026 · PJM Interconnection · Governance, accountability and labour
1 July 2026
AII-1000-283932 sources ↓
Quick facts
Category
Governance, accountability and labour
Country
United States
Source
Yahoo!奇摩股市
Date
1 July 2026
Drafted by
Raidu Incident Detection

What happened

The U.S. Department of Energy approved PJM Interconnection, the region's grid operator, to require data centers to activate backup diesel generators during emergencies to prevent blackouts, described as a 'last resort' measure during high-temperature weather in July 2026.

PJM's region, covering Washington to Chicago, faced peak power demand estimated at 166 gigawatts, exceeding the previous 2006 record of 165 gigawatts, driven by both extreme heat and rapid AI data center expansion.

Reported impact

Virginia residents reported being trapped between two risks: potential blackouts from grid strain or air pollution exposure from diesel generators, with approximately one-third of Virginia data centers located within 500 feet of residential areas or schools.

Environmental advocates noted that many backup diesel generators lack adequate air pollution control equipment, raising health concerns as regulators had already approved over 8,000 diesel generators for Virginia data centers in recent years.

Index analysis

The incident reflects a structural shift in U.S. power systems, with AI and cloud computing creating electricity demand growth far exceeding traditional industrial loads and outpacing infrastructure development in generation, transmission lines, and transformers.

Grid reliability experts noted that high temperatures, rapid data center expansion, and regulatory system delays converged to create new risks, with future data center load increases projected to far exceed current levels.

Governance controls

PJM initiated a 'backstop procurement' process to secure additional power supply when existing capacity market mechanisms proved insufficient, and planned to encourage consumer demand reduction before requiring diesel backup activation.

Regulators acknowledged the need to balance reliable power supply, electricity price impacts, and fair cost distribution, though local organizations expressed concern that communities bore the health costs of data center development.

Classification

Category
Governance, accountability and labour
Industry
IT infrastructure and hosting
Country
United States
Harm status
Potential harm
Intent
Unintentional
Timing
Unknown
Harm types
Environmental, Public interest
Harmed
General public, Civil society
Failure modes
Overreliance
AI systems
AI data centers

Sources and evidence

This record was researched and written by the Index, based on the following original reporting.

Yahoo!奇摩股市
PJM推後備採購機制,因應AI數據中心用電暴增壓力
1 July 2026
凤凰网(凤凰新媒体)
AI数据中心"抢电",美居民怒了
2 July 2026