Slot Machine

Key implications

On 28 October, GE Vernova reported 122 gigawatts (GW) of gas turbines on firm order or reserved, up from 50 GW in March 2025. In times past, an order book of that size would have meant a wave of power plants: turbines ordered, permits filed, and steel rising within two or three years. Yet the field is lagging the factory. Of the 64 GW of US gas capacity planned for 2027–30, 15% is under construction; 58% has not yet sought regulatory approval; and of the gas planned for 2025–27 at the end of 2023, almost a third has since been canceled or postponed. KPA analysis shows why. Much of the turbine shortage is a shortage of reservations—paid options on factory time—and in the third quarter those options began hardening into firm orders faster than new ones were signed. The question is no longer only whether turbines can be found, but who holds the risk once a reservation becomes an obligation.

• The binding queue for new gas plants has moved from the grid to the factory. GE's 122 GW is about six years of its current output; on its own expansion plan, the book would not be fully delivered until around mid-2031. In the third quarter, GE sold several slots for delivery in 2032. Siemens Energy holds a further 69 GW, more than four years of its deliveries.
• The shortage is real at the factory but partly paper in the field. Planned US gas capacity has risen 3.5 times in two years, to 71 GW, while capacity under construction rose from 5 GW to 13 GW. Of the gas plans on record for 2025–27 at the end of 2023, about a third is operating today.
• As reservations harden into orders, the risk moves to the buyers and their customers. In the third quarter, GE converted 13 GW of reservations into orders and signed 8 GW of new ones. About 80% of its gas book is with utilities and other traditional customers. At 2025 costs, the combined-cycle plants now planned would cost at least $31 billion more than at 2023 costs.

—October 2026