Natural Gas Power

Natural gas generated roughly 6,900 TWh in 2025, about 22% of the world’s electricity and second only to coal. It is the swing fuel of most power systems. Combined-cycle plants supply bulk energy, fast-starting turbines and engines cover peaks, and gas sets the wholesale price in many markets for much of the year. Gas plants are flexible, cheap to build and emit roughly half the CO₂ of coal per kWh. Against that: volatile fuel prices, heavy import dependence in many countries, and methane leaking along the supply chain.

How it works

Pipeline gas is mostly methane (CH₄, typically 85–95%) with some ethane, propane, nitrogen and CO₂. Power plants burn it in three main ways.

Gas turbines (Brayton cycle). A compressor raises air to 15–25 times atmospheric pressure (up to about 30 in advanced machines). Fuel burns in the combustor at firing temperatures that now exceed 1,500 °C in H- and J-class units. The hot gas expands through the turbine, which drives both the compressor and a generator. Exhaust leaves at roughly 450–650 °C.

Combined cycle. In a combined-cycle gas turbine (CCGT) plant, that exhaust flows through a heat recovery steam generator. The steam drives a steam turbine (Rankine cycle) that adds about 50% more electricity with no extra fuel. The usual configurations are one gas turbine with one steam turbine (“1×1”) or two gas turbines feeding one steam turbine (“2×1”).

Reciprocating engines. Large spark-ignition or dual-fuel engines of 5–20 MW each, of the type made by Wärtsilä, MAN and Jenbacher, are grouped into plants of tens to hundreds of megawatts. They start in under five minutes and hold their efficiency at part load, which makes them a good fit for islands, mines and grids with large shares of variable renewables.

Plant engineers express efficiency as heat rate, the fuel energy needed per kWh delivered:

US data are quoted on a higher heating value (HHV) basis. European and manufacturer data are usually on lower heating value (LHV), which makes the same plant look about 10% more efficient because natural gas’s LHV is roughly 90% of its HHV. The fuel energy content calculator converts between cubic metres, therms, MMBtu and MWh of gas.

Main variants

The efficiency record for a CCGT plant was set by H-class units in Japan and France and stands at just over 63–64% LHV. Even then, about 37% of the fuel energy leaves as low-grade heat, which combined heat and power plants recover for district heating or industry.

Efficiency and capacity factor

The capacity factor of gas plants depends on their role and the gas price. In the US, combined-cycle units averaged roughly 55–60% in recent years (EIA). Shale gas was cheap enough that in some years they ran more than coal plants. Simple-cycle turbines typically run at 5–15%, starting only during demand peaks, outages or low-wind evenings. In Europe CCGT utilisation is lower and more volatile, rising when gas undercuts coal on carbon-inclusive cost and falling when gas is expensive or renewable output is high. The capacity factor calculator shows what those hours mean in annual output.

Gas plants are among the most flexible thermal generators. Modern CCGTs ramp at 30–80 MW per minute, and engines and aeroderivatives start from cold in minutes. That is why they supply much of the peak load capacity, operating reserves and backup for wind and solar. The catch: frequent cycling raises maintenance costs and lowers efficiency.

Costs

Gas power is cheap to build and expensive to run, the reverse of renewables and nuclear.

• Lazard LCOE+ 2024 (US, unsubsidised): gas combined cycle $45–108/MWh; gas peaking $110–228/MWh. The ranges assume US gas prices of a few dollars per MMBtu and capacity factors of 30–90% (CCGT) and 10–15% (peakers).
• IEA/NEA Projected Costs 2020: CCGT median of about $71/MWh at 7% discount rate, including an assumed $30/t carbon price. Regional fuel prices make the spread wide.
• NREL ATB and EIA AEO put new CCGT overnight capital at roughly $1,000–1,300/kW and combustion turbines somewhat lower per kW.

Fuel is what matters. At a heat rate of 6.5 MMBtu/MWh, every $1/MMBtu change in gas price moves generation cost by about $6.5/MWh. At US Henry Hub prices of $2–4/MMBtu, fuel costs $13–26/MWh. At the European TTF prices of 2022, when monthly averages briefly topped €200/MWh of gas (roughly $60/MMBtu), fuel alone cost roughly €350/MWh of electricity. Adding a carbon price of €80/t increases CCGT cost by about €28–32/MWh. Run your own numbers with the LCOE calculator. Set the capacity factor to 10% and you’ll see why peakers cost so much per MWh despite being cheap to build.

LNG and the global gas market

Countries without pipeline supply import liquefied natural gas, cooled to about −162 °C and shipped in tankers. Liquefaction, shipping and regasification consume roughly 10–15% of the gas’s energy and add several dollars per MMBtu. They also add 10–40% to lifecycle emissions depending on the route. The US became the world’s largest LNG exporter in 2023, ahead of Qatar and Australia. LNG has partly tied together what used to be regional gas markets: a cold winter in Asia now lifts power prices in Europe.

Land, water and materials

A gas plant’s own footprint is small, often a few hectares for a gigawatt-scale CCGT. The wells, gathering lines and pipelines upstream cover far more ground, especially in shale basins. CCGTs need about a third of the cooling water of coal or nuclear plants per MWh, because the gas turbine part rejects its heat through the exhaust. NREL’s review (Macknick et al., 2012) gives median consumption of about 200 US gallons per MWh for tower-cooled CCGTs, versus roughly 700 for coal and nuclear. Air-cooled condensers can cut this to near zero at a small efficiency penalty. Simple-cycle turbines use almost no water.

Lifecycle CO2 emissions

The smokestack tells only part of the story. Methane leakage from wells, compressors, pipelines and LNG plants adds a lot, because methane’s 100-year global warming potential is about 30 and its 20-year GWP is about 80 (IPCC AR6).

• IPCC AR5 WGIII Annex III (2014): combined cycle median 490 gCO2e/kWh, range 410–650.
• UNECE 2021: NGCC 403–513 gCO2e/kWh without capture. With CCS the range falls to roughly 90–220 gCO2e/kWh, limited by capture rates and upstream methane.
• Open-cycle turbines emit 50–60% more per kWh than CCGTs because of their lower efficiency.

Nobody agrees on how much leaks. Alvarez et al. (Science, 2018) measured US supply-chain emissions equal to about 2.3% of gross gas production, roughly 60% above the official EPA inventory. The IEA’s Global Methane Tracker finds that around 40% of oil and gas methane could be abated at no net cost. Lifecycle studies place the point where gas loses its climate advantage over coal on a 20-year horizon at leakage in the low single-digit percent. On a 100-year basis gas keeps a clear advantage at most realistic leakage rates. Satellites that spot super-emitters, leak detection and repair programmes, and the EU methane regulation (which also covers imports) all target this gap.

Safety and health impacts

Gas burns with almost no sulfur dioxide, particulates or mercury. It still emits nitrogen oxides, controlled with dry low-NOx burners and selective catalytic reduction. Per unit of energy it does far less harm to health than coal or oil. Markandya & Wilkinson (Lancet, 2007) estimated about 4 deaths per TWh for European gas. Our World in Data’s synthesis gives about 2.8 deaths per TWh, compared with about 24.6 for coal and 0.03 for nuclear, wind and solar. Supply-chain risks include pipeline explosions, well blowouts and storage incidents such as Aliso Canyon (California, 2015–2016), and local air and water impacts near shale operations.

Decarbonisation options

Carbon capture. Post-combustion amine capture on CCGTs is harder than on coal because gas exhaust is dilute (about 4% CO₂ versus 12–14% for coal). The penalty is 7–10 percentage points of efficiency. The UK’s Net Zero Teesside Power project (about 740 MW, investment decision in late 2024) aims to be the first commercial-scale gas plant with CCS, at a capture rate of up to 95%.

Hydrogen blending. Most new turbines can burn 30–50% hydrogen by volume, and OEMs target 100% by around 2030. Hydrogen has about one-third of methane’s volumetric energy, so a 30% blend by volume cuts CO₂ by only about 10%, and the hydrogen must itself be low-carbon. The green hydrogen calculator shows the electricity required. See green hydrogen for the economics.

Biomethane from anaerobic digestion can substitute directly for fossil gas, but supply is limited compared with power-sector demand.

Outlook

Where gas power is heading depends on the region. In the US, shale gas displaced coal over 2008–2024 and gas now supplies roughly 40% of US electricity. Data-centre demand has revived orders for new turbines, and lead times for heavy-duty units stretched to several years by 2025. In Europe the 2022 price shock accelerated renewables and efficiency, and gas burn for power has fallen. It is still the main backstop for dark, windless spells and earns money through capacity markets. The Middle East, Southeast Asia and parts of Africa are still adding large CCGT fleets.

In a net-zero system, gas plants end up selling capacity rather than energy. They run fewer hours as firm backup alongside battery storage and wind power, and eventually use CCS, hydrogen or biomethane. The world energy map shows which countries depend most on gas, and technology comparison sets its costs and emissions against the alternatives.