AI boom drives US to triple natural gas power projects as data centers fuel global surge

4 Sources

Share

The US has tripled its natural gas power projects in 2025, driven by AI energy demand from expanding data centers. Nearly a quarter of global gas capacity in development is now in the US, with a third directly powering data centers. This surge threatens climate goals, potentially locking in 12.1 billion tonnes of lifetime COâ‚‚ emissions from US projects alone.

AI Energy Demand Drives Unprecedented Natural Gas Expansion

The United States has emerged as the global leader in a dramatic surge of natural gas development, tripling its gas-fired power projects in 2025 as the AI boom creates unprecedented electricity demand. The country now accounts for nearly a quarter of all global gas capacity in development, surpassing China with 251,737 megawatts of gas or oil-fired plants under construction or planned, according to Global Energy Monitor (GEM)

1

2

. This represents a potential 50 percent increase to the existing US fleet of gas-fired power plants, with more than a third of this new capacity specifically designed for powering data centers that support generative AI operations

2

.

Source: The Verge

Source: The Verge

Gas-fired power generation in development globally rose by 31 percent in 2025, and if current trends continue, 2026 is on track to shatter the annual record for new gas power additions set in 2002 during America's shale gas revolution

1

2

. The rush to install more powerful hardware into expanding data centers has led to forecasts of skyrocketing power demand, with experts predicting electricity demand in the US could surge 60 percent by 2050 due to new data centers

4

.

Climate Goals Face Major Setback from Gas-Fired Power

The global surge in gas development poses severe implications for climate change efforts and the energy transition away from fossil fuels. If all planned US projects are completed, they will generate 12.1 billion tonnes of lifetime CO₂ emissions—double the current annual greenhouse gas emissions from all sources in the United States. Worldwide, the planned gas boom could produce 53.2 billion tonnes of emissions over the projects' lifetimes, nearly matching the 54.3 billion tonnes from the entire existing global fleet of gas and oil-fired power plants

2

.

Source: The Register

Source: The Register

"This 'petrotech' buildout risks tethering the power grid to fossil fuels for the next generation, thereby creating long-term pollution for a short-term power fix," Jenny Martos, project manager for GEM's Global Oil and Gas Plant Tracker, told The Register

2

. The expansion directly contradicts the Paris agreement goals, which require replacing fossil fuels with renewable energy and slashing carbon emissions to net zero by around 2050

1

.

Tech Companies Choose Speed Over Sustainability

Despite public commitments to clean energy, major tech companies are rapidly deploying natural gas infrastructure to meet immediate AI demands. Nearly 75 percent of power equipment planned for on-site use at data centers is natural gas, according to a report by Cleanview, a market intelligence platform

3

. The firm identified 46 data centers planning to build their own on-site power, with a combined capacity of 56 gigawatts—roughly the peak power draw of a major US city

3

.

"Press releases for these projects say one thing. But the 35+ permit documents, site plans, and equipment deals we found tell a much different story," the Cleanview report notes

3

. Microsoft has admitted openness to deploying natural gas with carbon capture technology alongside renewables, while Meta announced plans to deploy 2,262 megawatts of natural gas power for its largest-ever datacenter, including a $1.5 billion facility in El Paso, Texas

2

.

Risk of Stranded Assets and Gas Lock-In

The massive investment in gas infrastructure creates significant financial and environmental risks if the AI bubble fails to materialize. "There is a risk that this capacity could lock in future emissions and become stranded assets if anticipated electricity demand from AI never materializes," Martos warned

1

. Even if AI demand doesn't meet projections, the reality of gas lock-in means these plants will likely still feed the power grid to offset potential losses, further hampering the transition away from carbon emissions

2

.

Source: Axios

Source: Axios

The AI race has become so intense that companies are choosing power they can get now over cheaper renewable energy options later, amid years-long connection delays to the electricity generation infrastructure

3

. This represents a sharp reversal from just over a year ago, when virtually all data center developers planned to use the electric grid to power 100 percent of their projects

3

. The Trump administration has accelerated this trend through its AI Action Plan, which prioritizes speeding the build-out of new fossil fuel infrastructure for data centers while blocking clean energy projects

1

.

Regional Impact and Public Backlash

Texas leads the US expansion with 57.9 gigawatts of new gas power under development in 2025, followed by Louisiana and Pennsylvania. The proliferation has raised power bills for many Americans and sparked grassroots opposition over electricity costs and water consumption, halting some projects and creating political challenges. Steve Clemmer, director of energy research at the Union of Concerned Scientists, warns that "frenzied datacenter growth with little transparency or guardrails puts the public at risk of massive cost increases". Watch for continued tension between AI infrastructure expansion and local communities bearing the environmental and financial costs of this rapid transformation.

Today's Top Stories

TheOutpost.ai

Your Daily Dose of Curated AI News

Don’t drown in AI news. We cut through the noise - filtering, ranking and summarizing the most important AI news, breakthroughs and research daily. Spend less time searching for the latest in AI and get straight to action.

© 2026 Triveous Technologies Private Limited
Instagram logo
LinkedIn logo