EIA raised its own modeled overnight capital costs by about 20% for combined-cycle and 40% for simple-cycle gas technologies for AEO2026, citing turbine supply shortages and tight labor and materials markets. If your screening model still carries pre-2025 gas capex, you are budgeting against a market that no longer exists — and that error compounds when the project category itself is mislabeled.
Natural gas power plant conversion is not one project. It is three, with capital costs separated by an order of magnitude and heat rates separated by 56%. A boiler refuel of an existing coal steam unit, a brownfield repowering with new combustion turbines and HRSGs, and a gas-plus-carbon-capture build are different scopes, different equipment lists, different permitting paths, and different in-service dates. Conflate them in a board deck and both the capital number and the online date are wrong by multiples.
Key Takeaways
- EIA models an average coal-to-gas boiler conversion at $231/kW in 2025 dollars — a fleet average covering both boiler modification and a unit-specific gas lateral, which is why it will not match your site.
- The converted steam unit runs at 10,300 Btu/kWh against a 6,600 Btu/kWh NGCC benchmark — 56% more fuel per MWh, by calculation from EIA's published heat rates.
- AEO2026 new-build benchmarks: $1,158/kW for a 419 MW industrial-frame CT at 9,142 Btu/kWh, and $1,086/kW for a 627 MW single-shaft CCGT at 6,226 Btu/kWh.
- Gas with 90–97% capture carries $2,279–$2,366/kW TOC in NETL's H-class study, with net output falling from 649 MW to 579–572 MW.
- Get a firm pipeline lateral cost and schedule from the local gas utility before requesting turbine or boiler quotes. It usually governs the in-service date, not equipment procurement.
Fuel Conversion, Repowering, and Brownfield Replacement Are Not the Same Project#
EIA treats these as three distinct project categories with different costs and performance profiles, and the distinction is not academic — it determines which cost benchmark applies and which permitting path you are on.
Fuel conversion modifies an existing coal boiler to burn natural gas. The steam turbine, generator, condenser, cooling system, and switchyard stay. Repowering adds gas turbine generation to an existing site, typically feeding a heat recovery steam generator that drives the retained steam turbine. Brownfield replacement builds a new combined-cycle block on the existing site and retires the old unit outright.
Only the first is what EIA means by coal-to-gas conversion. It is not adding a combustion turbine, not replacing the boiler with a combined-cycle block, and not coal gasification. Comparing a $231/kW conversion benchmark against a $1,086/kW CCGT benchmark is comparing two different projects, not two bids on the same one.
What Changed: A 20–40% Turbine Cost Jump Your Spreadsheet Has Not Caught#
EIA revised AEO2026 gas costs upward using recent PJM cost-of-new-entry work — about 20% for combined-cycle and 40% for simple-cycle — because of turbine, labor, and materials constraints.
Simple-cycle took the larger hit, which narrows the historic capex advantage of peaking iron on a $/kW basis: $1,158/kW for the 419 MW frame CT against $1,086/kW for the 627 MW single-shaft CCGT, with the 1,227 MW 2x2x1 configuration lowest per kW at $1,032/kW. If you are chasing MW in a constrained turbine market, the per-kW premium for simple-cycle no longer compensates for its 9,142 Btu/kWh heat rate unless you genuinely need two-year delivery and low run-hours.
EIA cautions that gas-plant costs vary with project size, location, interconnection access, fuel-supply access, transportation logistics, and ambient conditions. The national benchmark is a starting anchor, not a bid.
The Equipment Actually Being Touched#
The three pathways touch almost entirely different equipment, which is why the capital spread is what it is.
Equipment Scope by Pathway#
| Pathway | In scope | Retained | Typical long-lead item |
|---|---|---|---|
| Boiler fuel conversion | Burners, windbox, fuel gas trains, igniters, controls, gas lateral and metering | Boiler pressure parts, steam turbine, generator, condenser, cooling, switchyard | Gas lateral and firm transport capacity |
| Brownfield repowering | Combustion turbine, HRSG, generator step-up transformer, auxiliary switchgear, controls | Steam turbine, condenser, cooling, switchyard, interconnection | Combustion turbine slot |
| Brownfield replacement | Full CCGT island, GSU, medium-voltage switchgear, balance of plant | Site, water rights, interconnection position, land | Combustion turbine slot and HRSG |
| Gas plus capture | Capture island, compression, auxiliary transformers, motor loads, water treatment | Depends on new-build versus retrofit scope | Capture island and CO2 offtake |
The boiler conversion removes scope rather than adding it. Fuel handling, ash systems, and coal yard staffing come off the books, which is where EIA's assumed O&M reduction originates. The repowering and replacement paths pull you into the same gas turbine and generator procurement queue every other developer is in.
What It Costs: Benchmark $/kW Across Four Pathways#
Benchmark Capital Cost and Efficiency by Conversion Pathway#
| Pathway | Capital cost benchmark | Heat rate / efficiency note | Source |
|---|---|---|---|
| Coal boiler refueled to gas, fleet average | $231/kW, 2025$ — boiler modification plus lateral, averaged | 5% heat-rate penalty; no capacity penalty; 33% lower fixed O&M, 25% lower variable O&M | EIA AEO2026 |
| Benchmark 300 MW coal-to-gas steam conversion | $226/kW overnight, 2016$ basis | 10,300 Btu/kWh | EIA Jan 2024 capital cost assumptions |
| New industrial-frame CT, 419 MW | $1,158/kW overnight, 2025$ | 9,142 Btu/kWh; 2-year modeled lead time | EIA AEO2026 |
| New single-shaft CCGT, 627 MW | $1,086/kW overnight, 2025$ | 6,226 Btu/kWh; 3-year modeled lead time | EIA AEO2026 |
| New multi-shaft 2x2x1 CCGT, 1,227 MW | $1,032/kW overnight, 2025$ | 6,226 Btu/kWh | EIA AEO2026 |
| H-class 1x1 NGCC, no capture, 649 MW net | $1,125/kW TOC, 2018$; $44.6/MWh LCOE | State-of-the-art reference case | NETL, June 2024 |
| H-class 1x1 NGCC, 90–97% capture | $2,279–$2,366/kW TOC, 2018$; $69.3–$71.3/MWh LCOE incl. transport and storage | Net output drops to 579–572 MW | NETL, June 2024 |
Three different dollar years sit in that table. EIA's AEO2026 figures are 2025 dollars, the January 2024 capital cost report uses a 2016-dollar basis, and NETL's are 2018 dollars. Do not stack them in one bar chart without escalation, and do not treat any of them as installed EPC price on your site.
The boiler-side cost function EIA applies is $484.75 × (75/CAP)^0.35 per kW for pulverized-coal boilers and $346.25 × (75/CAP)^0.35 per kW for cyclone boilers, with CAP in MW. Run it and the scale effect is obvious: a 300 MW pulverized-coal unit prices at roughly $298/kW on the boiler side before the lateral, a 600 MW unit at $234/kW, and an 800 MW unit at $212/kW. A 600 MW cyclone unit comes in near $167/kW. Small units screen badly, which is most of why EIA's fleet average lands where it does.
EIA also excludes units outright: anything below 25 MW, plus fluidized-bed combustion and stoker boilers. Use those three filters as a first-pass portfolio screen before spending engineering hours.
The Efficiency Tax: Why the Low-Capex Path Burns More Gas#
$231/kW buys you the asset. It does not buy you the efficiency.
The operating trade is explicit in EIA's assumptions: a 5% heat-rate penalty, 33% lower fixed O&M, 25% lower variable O&M, and no capacity penalty. What stays is a steam cycle running at 10,300 Btu/kWh against a 6,600 Btu/kWh NGCC benchmark — 56% more fuel burned per MWh, by calculation from those two published heat rates. Measured against AEO2026's single-shaft CCGT case at 6,226 Btu/kWh, the gap widens to roughly 65%.
That is a permanent feature of the asset, not a commissioning issue. Whether it matters depends entirely on run-hours and remaining book life, and the fuel-cost consequence depends on gas prices this analysis does not attempt to forecast. The conversion is a capital-avoidance decision, not a performance decision.
What Can Fail: Pipeline Laterals, CCS Retrofits, and Schedule Risk#
The lateral, not the turbine, usually sets your date. EIA prices the gas lateral separately from boiler modification precisely because it is unit-specific — required gas volume, distance to the nearest mainline, and lateral diameter each swing the number. No national benchmark will survive due diligence on this line item.
Operationally the lateral is a right-of-way, easement, and utility-construction-queue problem, not an equipment problem. A frame CT carries a two-year modeled lead time and a single-shaft CCGT three years, but neither matters if the lateral and any required compression are not contracted and permitted on a parallel track. Boiler modification outages can be scheduled around a maintenance window. A lateral cannot. Both live Arizona conversions are pipeline-gated: SRP expects the gas infrastructure serving Coronado and the unit conversions complete by late 2029, and TEP expects its Springerville lateral finished by the end of 2029 with Units 1 and 2 converted in the first quarter of 2030.
Capture changes the asset class. NETL's June 2024 H-class study prices a 649 MW net 1x1 NGCC without capture at $1,125/kW TOC and $44.6/MWh LCOE. Add 90–97% capture and total overnight cost roughly doubles, LCOE rises to $69.3–$71.3/MWh including transport and storage, and net output falls to 579–572 MW.
That derate — 70 to 77 MW of net capacity consumed by the capture island's parasitic load, about 11% to 12% of the uncaptured net output — is the number procurement teams miss. You are buying auxiliary transformers, medium-voltage switchgear, and motor loads for a plant whose salable output shrank. Size the auxiliary electrical scope against gross generation plus capture-island load, not net export. NETL's figures describe new-build H-class economics; nothing in this data set supports extrapolating them to bolting capture onto an existing coal boiler.
Who Pays and Who Approves: Regulatory Status of Active Conversions#
Approval status is where screening decks most often overstate certainty. Two Arizona conversions cleared a real regulatory gate this year; others on the list have only a company announcement behind them.
On March 4, 2026 the Arizona Corporation Commission voted unanimously to approve converting the Springerville and Coronado coal generating plants to natural gas. The Commission's own meeting record documents the unanimous vote on TEP's Springerville request. Springerville Units 1 and 2 represent 700–800 MW of combined baseload; TEP put the boiler conversion at roughly $170 million against more than $450 million in coal upgrades and retrofits required to keep operating under the prior compliance path. At 700–800 MW, $170 million implies $213–$243/kW.
Compare that to the right benchmark, not the headline one. TEP's figure covers boiler conversion; the Springerville pipeline is being built by a third-party gas supplier, so the lateral is not inside it. The apples-to-apples comparison is EIA's boiler-side function, which at 700–800 MW returns $212–$222/kW — putting TEP slightly above the modeled curve, which is what you would expect from a real project. Set against EIA's $231/kW fleet average, which bundles an averaged lateral, the alignment looks tighter than it is. This is the classification error the rest of this article is about, running live in a single comparison.
SRP's Coronado estimate is structured differently and should not be compared to it directly: $1.1 billion through 2045, covering conversion and operation, which the Commission noted is about $300 million less than a new gas facility over the same period. That is a lifecycle number, not an overnight capital number.
Active Coal-to-Gas Conversion and Repowering Projects#
| Utility | Plant | Project type | Region | Status |
|---|---|---|---|---|
| Tucson Electric Power | Springerville Units 1–2 | Coal-to-gas boiler conversion | Arizona | ACC approved unanimously March 4, 2026; ~$170M; lateral late 2029, conversion Q1 2030 |
| Salt River Project | Coronado | Coal-to-gas boiler conversion | Arizona | ACC approved March 4, 2026; $1.1B through 2045 incl. operation; target late 2029 |
| Xcel Energy | Harrington Station, 1,018 MW | Coal-to-gas boiler conversion | Texas | Completion targeted May 2025, per Xcel; plant was originally dual-fuel capable, which Xcel says lowered conversion cost |
| Duke Energy | Cayuga Energy Complex | Brownfield gas repowering | Indiana | Broke ground, per Duke Energy |
| DTE Energy | Belle River | Coal-to-gas boiler conversion | Michigan | Unit 1 in 2025, Unit 2 in fall 2026, per DTE |
| NRG | Joliet Generating Station | Coal-to-gas conversion, three units | Illinois | Reported complete, 1,326 MW, per NRG |
DTE described Belle River as a roughly $154 million project extending plant life by at least 13 years. NRG has reported converting four coal plants totaling nearly 3,000 MW, including Joliet at 1,326 MW, Shawville at 590 MW, Big Cajun II at 540 MW, and New Castle at 325 MW. These are company figures, not regulator-verified costs or independent estimates. Treat them as screening anchors and nothing more — and note that at the ACC hearing, a TEP witness put the Springerville conversion near $200 million against the $170 million in the filed application, which is the normal drift between an application number and a live estimate.
Buyer Decision Framework: Which Pathway Fits Your Asset#
| Decision vector | Favors boiler refueling | Favors new CCGT | Favors deferring or hybridizing |
|---|---|---|---|
| Run-hours and remaining life | High run-hours on a depreciated asset with 10-plus years of steam-cycle life | High run-hours where the fuel gap compounds over a long book life | Volatile or declining run-hours; merchant exposure |
| Fuel exposure and policy risk | Bounded operating horizon with a known retirement date | Long horizon where 6,226 Btu/kWh materially changes lifetime fuel burn | Policy uncertainty; capture at $2,279–$2,366/kW is not a casual hedge |
| Schedule and siting | Switchyard, cooling, and steam island intact; lateral route short and uncontested | Two- to three-year turbine lead times acceptable; site absorbs HRSG and civil scope | Lateral route contested or firm gas capacity unavailable |
| Boiler eligibility | Over 25 MW, not fluidized-bed, not stoker | Any site | Unit fails the EIA screen |
| Balance sheet | Regulated utility with rate-base recovery | Owner able to carry $1,000-plus/kW and a three-year build | Short holding period; model storage or hybrid alongside |
A regulated utility with sustained high run-hours, secure firm gas transport, and a depreciated steam island should model the conversion first. A merchant owner with a five-year hold and uncertain dispatch should model new CCGT or a hybrid configuration, because the efficiency penalty never goes away.
Separating Benchmark From Bid#
Most budget failures in this category are classification failures. A $231/kW modeled retrofit average, a $1,086/kW EIA overnight cost, a NETL total-overnight-cost estimate in 2018 dollars, and a utility press-release project total are four different kinds of number, and only one of them is a transaction. Pressure-test turbine and balance-of-plant assumptions against the four-tier pricing framework in the equipment marketplace, compare generation technologies on cost and time-to-power in the Power Intelligence Tool, and see how each figure is classified in our methodology.
FAQ: Coal-to-Gas Conversion Cost, Timeline, and Eligibility#
How much does it cost to convert a coal plant to natural gas?#
EIA models an average of $231/kW in 2025 dollars across its eligible coal fleet, split between boiler modification and a site-specific gas lateral. Boiler-side cost follows $484.75 × (75/CAP)^0.35 per kW for pulverized-coal units and $346.25 × (75/CAP)^0.35 for cyclone units, so a 300 MW pulverized-coal unit prices near $298/kW on the boiler alone while an 800 MW unit prices near $212/kW. Real projects vary: TEP's Springerville conversion is roughly $170 million for 700–800 MW, or $213–$243/kW.
Is a boiler conversion lower cost than building a new combined-cycle plant?#
On capital, decisively: $231/kW modeled against $1,086/kW for a 627 MW single-shaft CCGT and $1,032/kW for a 1,227 MW 2x2x1. On efficiency, the converted steam unit runs at 10,300 Btu/kWh against a 6,600 Btu/kWh NGCC benchmark — 56% more fuel per MWh. Which wins depends on run-hours and remaining asset life, not on the capital number alone.
Which coal units cannot be converted under EIA's modeled pathway?#
Units below 25 MW, fluidized-bed combustion units, and stoker boilers are excluded from EIA's conversion option. Those three filters are the fastest first-pass screen on a portfolio.
Does a coal-to-gas conversion reduce plant capacity?#
EIA assumes no capacity penalty for boiler refueling, but applies a 5% heat-rate penalty from lower stack temperature and higher moisture loss. Fixed O&M falls 33% and variable O&M falls 25% in the same assumption set, reflecting reduced staffing, maintenance materials, and waste disposal.
How long does a gas conversion or repowering take?#
EIA models a two-year lead time for a 419 MW industrial-frame CT and three years for a 627 MW single-shaft CCGT. For boiler conversions the gating item is usually the gas lateral rather than the outage: TEP expects its Springerville pipeline complete by the end of 2029 with conversion in the first quarter of 2030, and SRP targets late 2029 at Coronado. DTE's published Belle River sequence converts the first unit in 2025 and the second in fall 2026.
What does gas with carbon capture cost?#
NETL's June 2024 H-class study puts a 649 MW net 1x1 NGCC at $1,125/kW TOC in 2018 dollars and $44.6/MWh LCOE without capture, rising to $2,279–$2,366/kW and $69.3–$71.3/MWh with 90–97% capture including transport and storage, with net output falling to 579–572 MW. Those figures describe new-build H-class plants, not retrofits of existing coal boilers.
What to Do Next#
Request a firm lateral cost and schedule letter from your local gas utility — including compression, metering, and firm transport capacity — before you solicit a single turbine or burner quote. Run your unit capacity through EIA's boiler cost function and screen it against the 25 MW, fluidized-bed, and stoker exclusions. Then confirm that the switchyard, steam island, and auxiliary electrical equipment you plan to reuse can actually carry the converted plant's operating profile, and price the alternative honestly: compare conversion against new gas turbine capacity on both capital and lifetime fuel burn before the recommendation reaches a board deck.