Gas turbine shortage creates bottleneck for AI data center power

A global shortage of gas turbines is emerging as a major bottleneck for the expansion of artificial intelligence infrastructure. As hyperscale data centers require massive amounts of baseload power, the manufacturing queue for heavy-duty turbines has extended into 2031.

In the United States, data center power demand is projected to rise from 31 GW in 2025 to approximately 66 GW by 2027, according to Goldman Sachs. While global manufacturing capacity is estimated at 60 to 70 GW per year, total industry orders have already surpassed 110 GW. This supply-demand imbalance has caused capital costs for combined-cycle power plants to roughly double in recent years.

Major manufacturers are reporting unprecedented backlogs. GE Vernova reported a gas power equipment backlog of 116 GW as of Q2 2026, with delivery slots being offered as far out as 2031. Siemens Energy has seen its turbine sales nearly double between 2024 and 2025, with approximately 60% of recent orders linked to data center projects. Mitsubishi Heavy Industries also maintains a significant large-frame turbine backlog.

This shortage affects both the public grid and private ‘behind-the-meter’ generation projects, as tech firms attempting to bypass grid congestion still rely on the same constrained supply of turbines.