Bloomberg Pegs Abilene 1GW AI Data Center at $30B, $17B on Chips
TL;DR
- Bloomberg estimates that building and equipping a 1-gigawatt AI data center in Abilene, Texas would cost roughly $30 billion in total.
- Chips account for $17 billion of that bill, with $3 billion for power infrastructure and $2.4 billion for construction, per Cushman & Wakefield's breakdown.
- Short-term memory used in data centers now costs more than five times what it did a year ago; power-equipment backlogs stretch beyond 2030.
A 1-gigawatt AI data center in Abilene, Texas would cost roughly $30 billion to build and equip, and $17 billion of that goes to chips. The breakdown comes from a Bloomberg analysis by Lynn Doan, drawing on an exclusive cost breakdown prepared with Cushman & Wakefield.
Power infrastructure runs another $3 billion. Shell and construction cost $2.4 billion.
Bloomberg attributes the sticker shock to compounding pressures. "Construction material costs such as steel and cement have skyrocketed because of trade wars, fuel expenses and geopolitical conflicts, while the AI boom itself has driven up expenses for everything from chips to copper," the report says. "The price of short-term memory used in data centers is more than five times what it was just a year ago."
And the queue for power gear is long. Demand for electrical equipment, Bloomberg reports, is so high that "suppliers have backlogs stretching beyond 2030."
Abilene is not a hypothetical venue. The OpenAI campus there was expected to consume about 1.2 gigawatts when completed, which places the $30B figure in the same order of magnitude as spend already in motion on the site. The alert sits inside a run of infrastructure coverage on our tracker this week, including banks still working through Oracle's $38B AI data-center loan.
Originally reported by bloomberg.com
Read the original article →Original headline: Bloomberg Analysis: Equipping a 1GW Abilene AI Data Center Would Cost ~$30B, $17B of It on Chips