CITIC Construction Investment: It is expected that from 2026 to 2028, the global gas turbine supply-demand gap will be 16 GW, 16 GW, and 19 GW respectively.

CITIC Securities Research reports that the surge in AI computing power demand in North America is driving data centers to enter the “self-supplied power plant” era, raising the forecast for North American AI electricity demand.
The estimates for 2026-2028 have been increased by 4%, 35%, and 42% respectively compared to the previous version, with a three-year compound growth rate of AI electricity demand reaching 73%.
Slow grid connection is forcing data centers to build their own power sources, with gas turbines becoming the preferred base load power due to the lowest cost and highest efficiency.
However, their delivery cycle is long—1 to 3 years—and orders are already scheduled beyond 2028, creating a core supply bottleneck.
Other options include internal combustion engines, SOFC, aircraft modification engines, and ship modification engines.
It is estimated that from 2026 to 2028, the global supply-demand gap for gas turbines will be 16, 16, and 19 GW respectively.
Considering demand for other power generation methods, the total global power generation demand during the same period will reach 30, 42, and 59 GW.
This huge and persistent gap is driving a high boom cycle in the industries of gas turbines, internal combustion engines, solid oxide fuel cells, and aircraft modification engines.