Image for illustartive porposes

The U.S. AI and data center boom is creating a severe shortage of electrical equipment, particularly high-power transformers, threatening to delay nearly half of planned U.S. data centers this year, according to Bloomberg Green. (Read the full article)

The Stargate AI campus in Abilene, Texas, exemplifies the scale of demand. When completed, the eight-building facility will consume 1.2 gigawatts of power—enough for almost 1 million U.S. households. Despite over $650 billion in planned tech spending this year by companies including Alphabet, Amazon, Meta, and Microsoft, a shortage of critical electrical infrastructure is slowing construction.

High-power transformers, switchgear, and batteries are essential for both AI operations and the expanding electricity grid, which is seeing rising demand from electric vehicles and heat pumps. U.S. manufacturing cannot meet current demand, forcing developers to rely on imports from China and other countries, often at premium prices.

Delivery times for large transformers have surged from the previous 24–30 months to as long as five years, while AI companies increasingly demand equipment in less than 18 months. Some developers, like Crusoe Energy Systems, have resorted to refurbishing older transformers from shuttered power plants to keep projects on track.

Industry experts warn that without a dramatic expansion of domestic production capacity, delays could limit the effectiveness of trillions in AI and data center investments, potentially slowing the U.S.’s ambitions in the global AI race.

Source: bloomberg.com