Global gas turbine orders hit a record 38 GW in the second quarter of 2026, 29% above Q1 — a quarter that separate trade reporting puts up about 71% year over year, with the United States accounting for roughly half. That follows a 2025 in which worldwide orders reached 100.3 GW across 846 units, more than double the 58.2 GW and 399 units booked in 2024.
The queue those bookings created is now longer than most data center construction schedules. Wood Mackenzie puts global orders at 110 GW at the end of 2025 against worldwide manufacturing capacity of 60–70 GW per year — meaning the existing book absorbs between 1.6 and 1.8 years of full-rate production before a new order gets touched.
One clarification before these numbers do any work in your model. Orders are bookings, not commissioned capacity. None of that steel is on a pad. It is a claim on manufacturing slots that converts to operating megawatts over the back half of this decade — and as the next section shows, a large share of it is not even a firm claim.
Key Takeaways
- 38 GW ordered in Q2 2026 alone, a record quarter — about 52% above the 25 GW quarterly average implied by 2025's 100.3 GW full-year total. The curve is still steepening.
- 60–70 GW per year of global manufacturing capacity against 110 GW of orders. The ceiling is the factory, not the order desk — and below the factory, hot-section castings are the reported chokepoint.
- GE Vernova's headline 116 GW backlog is reportedly 53 GW firm and 63 GW slot reservations. More than half of the most-quoted number in the market is paid options, not contracted equipment.
- Turbine prices up 195% since 2019, heading to roughly $600/kW by end-2027 on Wood Mackenzie's estimate. Turbines are 20–30% of a combined-cycle project's cost and more of a simple-cycle one.
- Small frames are a different market from large ones. Large turbines quote around five years; smaller units reportedly run 18–36 months. If your date is before 2029, that distinction is your whole procurement strategy.
The Order Surge: Q2 2026 and the 2025 Baseline#
The 2025 total is the baseline everyone quotes, but Q2 2026 is the number that should drive your planning. Orders did not plateau after the 2025 surge — they accelerated into 2026, exactly as Wood Mackenzie predicted when it said turbine orders would peak in 2026 as developers scrambled to secure equipment for 63 GW of gas capacity additions planned from 2026 to 2030.
At the OEM level, the trajectory is just as steep. GE Vernova's gas turbine orders reached 20.2 GW in 2024, up from 9.5 GW in 2023, and in Q4 2025 the company booked 41 heavy-duty and 18 aeroderivative units, up 74% from 34 total a year earlier, valued at $7.7 billion. Mitsubishi Power has said its order book now extends to 2030, attributing the demand to AI and data centers alongside grid security needs — a statement of the company's current position, not a delivery guarantee to any specific buyer.
This is happening while gas holds its place in a growing supply mix rather than displacing anything: the IEA's mid-year update has gas and nuclear growing alongside renewables in the 2025 electricity picture.
Why the Three Headline Backlogs Do Not Add Up#
Add GE Vernova's 116 GW, Siemens Energy's reported 69 GW, and Mitsubishi's 35 GW and you get 220 GW. That number is meaningless, because none of the three is counting the same thing.
53 GW of 116 GW — the share of GE Vernova's headline backlog reported as firm equipment orders, with the remaining 63 GW being paid slot reservations that have not converted.
Trade reporting on the three books describes GE Vernova's 116 GW as roughly 53 GW firm equipment backlog and 63 GW slot reservations, Siemens' 69 GW as firm backlog with no reservations mixed in, and Mitsubishi's 35 GW as large-frame only — excluding its aeroderivative and mid-size lines. Three different definitions, one shared headline unit.
OEM Gas Turbine Order Snapshot#
| OEM | Recent order activity | Backlog / outlook | What the number includes |
|---|---|---|---|
| GE Vernova | 41 heavy-duty and 18 aeroderivative units in Q4 2025, up 74%, $7.7B; 20.2 GW in 2024 vs 9.5 GW in 2023 | 116 GW | Reportedly ~53 GW firm plus ~63 GW slot reservations |
| Siemens Energy | Strong heavy-duty demand reported alongside peers | ~69 GW reported | Firm backlog, no reservations mixed in |
| Mitsubishi Power | Order book extends to 2030 | ~35 GW reported | Large-frame only; excludes aeroderivative and mid-size |
| Industry total, 2024 | 399 units / 58.2 GW | — | Global bookings |
| Industry total, 2025 | 846 units / 100.3 GW | 110 GW on the books at year-end | Executed contracts plus binding reservations |
| Industry total, Q2 2026 | 38 GW, record quarter | — | Global bookings, US roughly half |
No OEM's backlog nets against another's. GE Vernova's 116 GW is the loudest number in the market and the most misread. It measures that company's commitments. It does not tell you how much slot capacity remains for you, and it does not offset demand queued at Siemens or Mitsubishi.
Manufacturing Capacity Is the Binding Constraint#
Wood Mackenzie's ratio — 60–70 GW per year of capacity against 110 GW of orders — is the entire story compressed into one line. Divide it out and the book runs 1.6 to 1.8 years deep before a new order is reached — and Q2 2026 added another 38 GW on top.
OEMs are spending against it. Wood Mackenzie notes GE Vernova investing over $160 million to raise output from about 50 large-frame turbines a year to 70–80 by late 2026, with Siemens Energy moving key facilities to 24/7 operation. But factory investment is not where the constraint actually lives. Hot-section castings — the blades and vanes that survive firing temperature — come from a limited set of specialized foundries, and that capacity does not scale on a hyperscaler's schedule.
Neither does the balance of plant. A frame order is half the procurement problem; the generator step-up transformer, medium-voltage switchgear lineup, and interconnection package each sit in their own constrained queues.
One number that is not driving this: fuel. Henry Hub spot gas traded around $2.79/MMBtu in mid-August 2026. The bottleneck is castings, welders, and factory slots, not the molecule being burned.
GE Vernova has publicized an 11-day on-site installation for certain configurations. That applies to installation and commissioning after equipment arrives. It does not compress the order-to-delivery timeline that is actually killing schedules.
What Buyers Are Paying: Price and Lead-Time Trend#
Gas Turbine Lead Time and Price Trend#
| Metric | Prior period | Current | Source and strength |
|---|---|---|---|
| Global orders, GW | 58.2 GW (2024) | 100.3 GW (2025); 38 GW in Q2 2026 alone | Utility Dive, trade press citing OEM disclosures |
| Global orders, units | 399 (2024) | 846 (2025) | Utility Dive, trade press |
| Turbine price index | 2019 baseline | +195%, reaching ~$600/kW by end-2027 | Wood Mackenzie, consultancy estimate — a market-level figure, not a uniform markup |
| Large-frame lead time | — | ~5 years; up to 7 for some frames | S&P Global Platts, interview-based |
| Small-turbine lead time | — | Reportedly 18–36 months | Utility Dive, analyst-attributed |
| Manufacturing capacity vs. book | — | 60–70 GW/yr vs. 110 GW booked | Wood Mackenzie, consultancy estimate |
Do not blend the $/kW figures. The $600/kW estimate is turbine equipment only; Wood Mackenzie puts turbines at 20–30% of a combined-cycle project's cost and a higher share of a simple-cycle one. Installed project cost is a different number — and it splits by when the equipment was ordered, not by technology. A GridLab review of project filings found plants finishing in 2026–27 reporting $1,116–$1,427/kW while projects targeting 2030–31 routinely report $2,000/kW or more. If you are benchmarking against a peer's published capex, check which side of that line their order sits on.
The 195% figure is a market estimate. Treat it as directional evidence that a 2022 budget number is dead, not as a multiplier on a specific F-class or H-class quote. Prices are expected to keep climbing through 2027.
Reservations Are Not Orders — and That Distinction Decides Your Slot#
The GE Vernova split makes the point better than any argument: more than half of the market's most-cited backlog figure is paid options rather than contracted equipment. Reservation agreements occupy manufacturing slots without necessarily binding either party to a delivery date.
That mixed accounting is a procurement hazard, not trivia. A non-binding reservation gives you a place in a line the OEM can re-sequence. A binding contract with a named delivery window gives you standing.
Contract terms to secure before any deposit clears:
- Firm delivery date as a contractual milestone with liquidated damages, not a verbal slot assurance.
- Price-escalation cap with a defined index and ceiling, not open-ended pass-through.
- Deposit treatment on scope change — refundable, transferable to another site, or forfeited.
- Slot transferability if your load forecast or interconnection position moves.
- Balance-of-plant sequencing, with GSU transformer, HV switchgear, and interconnection dates tied to turbine delivery rather than floating.
- Fuel infrastructure gating — firm pipeline capacity dated ahead of first fire.
Your Options If You Need Power Before 2029#
If your required commercial operation date lands before 2029, the new heavy-duty frame market is closed to you in practical terms. Three alternatives remain, each on a different clock.
Used and refurbished heavy-duty frames. Retired or relocated F-class and industrial frames come to market periodically. Diligence is the whole job: verify fired hours and start counts against the OEM service agreement history, confirm hot-section inspection intervals, and price the combustion-system upgrade current emissions permitting will require. Available units are indexed in the gas turbines category.
Aeroderivative and smaller frames. These trade heat rate for speed and start flexibility, and they sit in the 18–36 month band rather than the five-to-seven-year one. For a buyer holding a 2028 date, that band is the difference between a viable path and a closed one.
Reciprocating engines and standby gensets. For campus-scale bridge and backup power, standby and prime-rated packages remain the fastest route to firm capacity, at the cost of fuel logistics, emissions permitting, and footprint. Current classes and availability sit in the generators category, and lead-time context is in the data center backup power analysis.
Buyer Decision Framework: Reserve, Buy Used, or Bridge#
| Your situation | Recommended path | Primary risk to manage |
|---|---|---|
| Required COD before 2029 | Used or refurbished frames, aeroderivative units, or reciprocating bridge capacity | Condition risk on used equipment; heat rate and fuel cost on bridge units; emissions permitting |
| Site, permitting, and financing locked for 2029–2031 | A new-build heavy-duty reservation is defensible — with a contractual delivery date and escalation cap | Multi-year deposit exposure; balance of plant slipping behind turbine delivery |
| Load forecast, site, or financing still in flux | Hold turbine deposits; deploy bridge capacity and preserve optionality | Losing slot position — accept it rather than fund a frame for a project that may change scope |
| Interconnection is the gating item, not generation | Solve interconnection first; size behind-the-meter bridge capacity to the gap | Queue position and large-load rule changes ahead of energization |
| Backup and resilience duty only | Standby gensets and UPS, not turbines | Runtime hours, fuel storage, emissions compliance |
Model the mix against your own in-service date and site constraints in the Power Intelligence Tool before committing to any path.
What This Means for PPA Structuring and Investors#
Backlog growth reads as revenue visibility for turbine manufacturers. Read the same data from the buyer side and it says something less comfortable: gas-fired project timelines across the entire pipeline are lengthening at once.
That pushes PPA start dates right, extends bridge-power obligations, and raises the probability that offtake contracts signed against a 2028 COD get renegotiated. Underwriters should stress-test any gas-backed data center deal for a two-year turbine slip and price the bridge generation required to cover it.
The verdict: if your in-service date is before 2029, stop pricing new heavy-duty frames and start diligencing used units, aeroderivative packages, and reciprocating bridge capacity this quarter. If you are genuinely committed to 2029–2031, place the reservation now — but do not wire a deposit against a verbal slot assurance, a floating delivery window, or an uncapped escalation clause.
FAQ: Gas Turbine Orders, Lead Times, and Pricing#
What is the current status of the gas turbine order backlog?#
Wood Mackenzie put global orders at 110 GW at the end of 2025 against 60–70 GW of annual manufacturing capacity, and Q2 2026 added a record 38 GW. Individual OEM backlogs are not comparable — GE Vernova's 116 GW reportedly mixes firm orders with slot reservations, Siemens' figure is firm-only, and Mitsubishi's covers large frames only.
How long are gas turbine lead times right now?#
US gas-fired turbine wait times have reached as much as seven years for some frames, with large turbines commonly quoted around five years. Smaller turbines are reported in an 18–36 month window, which is a materially different market.
How much have gas turbine prices actually gone up?#
Wood Mackenzie estimates turbine prices up 195% since 2019, reaching about $600/kW by end-2027. That is a market-level estimate, not a uniform markup by class, region, or contract. At the project level, filings reviewed by GridLab show 2026–27 completions at $1,116–$1,427/kW against $2,000/kW-plus for 2030–31 targets — the same technology, two pricing regimes, split by order date.
Which OEMs still have open gas turbine capacity?#
No public source answers that reliably, because backlog disclosures use inconsistent definitions and none of them state remaining sellable slots. GE Vernova, Siemens Energy, and Mitsubishi Power all report multi-year books. The practical answer is to solicit named delivery windows from each OEM directly rather than inferring availability from published backlog totals.
Why did gas turbine orders double in a single year?#
Orders went from 58.2 GW across 399 units in 2024 to 100.3 GW across 846 units in 2025, driven largely by AI data center load, industrial demand, and utility capacity additions. A meaningful share came from buyers reserving forward slots years ahead of commissioning rather than from traditional utility procurement cycles.
What should a buyer do if they need gas-fired power before 2029?#
Treat the new heavy-duty frame market as closed and work the three shorter-cycle paths: used and refurbished frames, aeroderivative or smaller units in the reported 18–36 month band, and reciprocating or standby capacity for bridge duty. Each trades heat rate, emissions headroom, or footprint for speed, and each has to be evaluated against your interconnection date — a turbine that arrives before the substation does not produce revenue.
Before You Wire the Deposit#
Benchmark your required commercial operation date against what is actually deliverable, not against a new-build quote. Check secondary-market availability for gas turbines and generators, size the bridge generation you would need if a new-build slot slips two years, and read how each price figure in this article is classified in our methodology. If the used or bridge path closes your gap inside 18 months, that is the procurement decision — not the reservation.