GE Vernova gas turbines for AI data centers: massive expansion, pricing surge, and what builders need to know

AI data centers need enormous, firm power, and grid interconnection can't keep up. That is driving hyperscalers to buy their own power plants. GE Vernova's massive gas turbines are at the center of this shift, with its Greenville, South Carolina factory racing to expand capacity, pricing up 300% in three years, and order books filling through 2029 and into 2031.

What happened

GE Vernova's largest gas turbine plant in Greenville, South Carolina is expanding to meet AI-driven energy demand. The company hired 200 workers last year and expects to add 300 more by the end of the year. Hyperscalers including Amazon, Google, Microsoft, and Oracle are ordering the company's 31-foot, 280-ton HA-class turbines.

Each turbine can power roughly half a million homes. Microsoft bought seven of them for a Texas data center, delivering about 2.7 gigawatts of capacity enough for roughly 3 million homes. GE Vernova turbines are already online at Elon Musk's xAI Colossus 1 campus in Tennessee, and nearly a gigawatt more is being deployed at OpenAI's Stargate project in Texas.

Demand for these machines far outstrips supply. The order book is full through 2029, and GE Vernova is booking into 2030 and even 2031. Pablo Koziner, chief commercial and operations officer, told CNBC that about 20% of the company's gas power order book now goes to data center and AI applications.

One turbine costs more than $250 million. Analysts at Melius estimate that pricing has soared about 300% over the last three years, reflecting the intense capital expenditure growth in AI.

Why AI builders should care

AI builders rely on data center capacity that must be powered 24/7. Grid bottlenecks make independent power sources like gas turbines the leading solution for firm, scalable power at data center scale. The turbine supply constraint means that large AI projects need to lock in power equipment years in advance. Order books stretching to 2031 signal that the power bottleneck will remain tight for the foreseeable future.

GE Vernova's stock has gained nearly 60% in the past six months, a direct signal of investor conviction that the AI buildout's power needs are structurally growing. For builders planning AI infrastructure budgets, the 300% price increase over three years is a critical input when modeling total project costs.

Practical implications

Gigawatt-scale deployments are becoming standard. Microsoft's 2.7 GW Texas facility alone consumes enough power for 3 million homes. Each turbine's $250 million+ price tag means that a single AI data center campus can require billions in power equipment before any compute hardware is installed.

Turbine delivery lead times are now measured in years. Any AI infrastructure roadmap that assumes quick power availability after a site is chosen is likely underestimating the timeline. GE Vernova's own lead times extend beyond 2029. AI operators are increasingly choosing to own their power generation assets rather than waiting for grid interconnection.

GE Vernova says its current turbine is about twice as efficient as turbines from 20 years ago, which helps address some sustainability concerns, though natural gas remains a fossil fuel and public pushback on data center development could slow deployment.

Caveats

Pricing and deployment figures come from CNBC reporting, industry estimates, and GE Vernova disclosures. Actual contract terms, volumes, and turbine specifications may vary across individual deals. The 300% pricing increase is an analyst estimate from Melius, not a GE Vernova official figure. Environmental and community opposition could delay or alter planned deployments, especially for large campuses in areas with water or emissions restrictions.