Hyperscaler AI capex nears $750B; a 1GW data center costs about $30B
A Bloomberg analysis by Lynn Doan estimates that building and equipping a 1GW AI data center in Abilene, Texas, could cost roughly $30 billion, including about $17 billion for chips, $3 billion for power and $2.4 billion for construction. TechNews, citing aggregated data, reports that Meta, Amazon, Microsoft and Alphabet are expected to spend nearly $750 billion in combined capital expenditure this year, more than the annual GDP of Ireland or Belgium. It also reports that the hyperscalers' bond issuance rose from about $20 billion in 2024 to $121 billion in 2025, with roughly $65 billion of off-balance-sheet financing routed through special purpose vehicles.
TechNews argues the data center buildout has grown beyond ordinary tech business expansion into a capital race involving the chip supply chain, building materials, power grids and Wall Street financing.
This source provides an excerpt. Continue reading on the original site.