Half of US Data Center Builds Delayed or Cancelled as AI Power Demand Flips the Breakers

According to Tom's Hardware, half of all planned US data center builds have been delayed or cancelled, constrained by shortages of power infrastructure and components from China.

The Crisis

What's Causing the Bottleneck?

Power Grid Limits

• AI data centers consume 5-10x more power than traditional data centers
• Grid capacity takes years to expand (transmission lines, substations)
• Permitting delays of 3-5 years for new power infrastructure
• Renewable intermittency -- AI needs 24/7 baseload, solar/wind can't always provide it

Chinese Component Restrictions

• GPU supply chain -- some components sourced from Chinese manufacturers
• Trade restrictions -- export controls create supply chain disruptions
• Cost increases -- alternative suppliers charge premium prices

The AI Build-Out Paradox

The very technology driving economic growth is now constrained by physical infrastructure:

• More AI = More Data Centers = More Power Needed
• But: Power infrastructure takes years to build
• Result: AI growth slows while waiting for the grid to catch up

Who's Affected?

• Hyperscalers: Google, Microsoft, Meta, Amazon all facing delays
• Colocation providers: Equinix, Digital Realty adjusting plans
• AI startups: May lose competitive advantage due to compute delays
• Utilities: Scrambling to meet unprecedented demand

Solutions

1. Nuclear power -- TMI restart, SMR deployment for baseload
2. On-site generation -- data centers building their own power plants
3. Efficiency gains -- better chip architectures, liquid cooling
4. Demand management -- shifting compute loads to off-peak hours