GE Vernova's gas turbine plant in Greenville, South Carolina is running at capacity to meet demand from hyperscalers building AI data centers. Amazon, Google, Microsoft, and Oracle are lining up to buy turbines; the company hired 200 workers last year and expects 300 more by year-end. The order book is now full through 2029 with orders extending into 2030 and 2031.
One 31-foot, 280-ton turbine generates roughly 2.7 gigawatts—enough to power about 3 million homes. Microsoft purchased seven for its Texas data center; xAI Colossus 1 and OpenAI's Stargate project are already deploying them. A single turbine costs more than $250 million, with prices up 300% in the last 3 years, according to Melius Research. About 20% of GE Vernova's gas power order book is now AI/data center applications.
For architects: this is a capex bottleneck. Supply of standalone power infrastructure for AI is constrained through 2031. Grid-independent data center deployment is no longer optional for hyperscalers but materially supply-constrained. GE Vernova stock is up 60% in six months.