US Gas Power Expansion for AI Data Centers Doubles in Six Months, Overtaking China

US Overtakes China in Gas Power Construction for AI Data Centers

The United States has surged ahead of China in building new gas-fired power generation, driven primarily by the AI boom and data center expansion. According to Global Energy Monitor, gas power capacity under development in the US jumped 50% from 252GW to 378GW between January and mid-20262

2

. This represents roughly one-third of the global total and marks a dramatic reversal from just six months earlier when China led in gas plant construction2

2

.

The surge in gas-fired power projects reflects a stunning 76% increase in under-construction facilities during the first half of this year alone2

2

. The US is now building twice as much gas-fired capacity as China and more than any other country globally4

4

. If all proposed projects are completed, the US will expand its gas fleet by approximately two-thirds at a capital cost exceeding $647 billion2

2

.

Around half of this new capacity is directly tied to AI data centers requiring massive amounts of electricity to handle surging electricity demands of AI applications2

2

. The pipeline of gas projects exclusively for AI data centers has exploded from just 4 gigawatts in early 2024 to 97 gigawatts by the end of 2025, and further to more than 189 gigawatts by mid-20261

1

.

Climate Costs and Fossil Fuel Dependence Mount

The massive buildout of gas power for AI infrastructure carries significant climate costs. Using gas rather than renewable energy to power this surge in AI data centers threatens to increase US power emissions by as much as 20%, according to estimates2

2

. Many facilities are being built with inefficient turbines, which can substantially increase greenhouse gases emissions1

1

. Some gas plants are permitted to emit more greenhouse gases annually than many small- and medium-sized countries1

1

.

Jenny Martos, a project manager at Global Energy Monitor, emphasized the long-term implications: "Building all of this gas for AI locks in decades of pollution and it is also locking in dependence on a volatile fuel cost, which will get passed down to rate payers"2

2

4

4

. This fossil fuel dependence contrasts sharply with China's approach, where data center development is oriented around renewable energy, especially solar and hydropower1

1

.

The rush to build AI infrastructure has created a backlog of efficient gas turbines, leading companies including Elon Musk's xAI to switch to smaller, less efficient turbines that are more heavily polluting2

2

.

Grid Delays and Infrastructure Bottlenecks Threaten Expansion

Despite ambitious plans, up to 50% of all planned US data center developments face significant delays or cancellation3

3

. The power grid is struggling to keep pace with energy demand from AI infrastructure. Regional utilities are overwhelmed with massive interconnect requests, creating multi-year backlogs for high-capacity power access3

3

.

Critical electrical hardware including high-voltage transformers and specialized switchgear currently carries lead times stretching past three years3

3

. Local utility networks simply lack the immediate surplus wattage needed to run gigawatt-scale data complexes alongside existing commercial and residential loads3

3

.

To bypass lengthy grid connection times, data center builders have increasingly turned to private power facilities known as behind-the-meter plants1

1

. The Trump administration has encouraged tech companies to bring their own power through a voluntary pledge signed by Microsoft, Meta, Google, OpenAI, and Amazon, along with several Republican governors and major utilities1

1

4

4

.

Local Opposition Intensifies Across the Country

Community resistance has emerged as a formidable barrier to data center expansion. A recent poll found that three-quarters of Americans do not want to live next to a data center, representing a huge increase in opposition over the past year2

2

. Local opposition centers around fears that massive industrial electricity demand will drive up utility costs for surrounding households3

3

.

Concerns also focus on water shortages, with two-thirds of more than 800 planned data centers located in drought-stricken areas2

2

. Residents across states like Pennsylvania, Texas, and Ohio are actively challenging new permits3

3

. Dozens of cities and counties have placed restrictions on new data centers, with New York in July becoming the first state to enact a temporary ban on new hyper-scale data center permitting and construction2

2

.

The National Republican Senatorial Committee has privately alerted top AI firms about potential political fallout from negative public opinion surrounding data centers, fearing backlash could discourage politicians from backing future projects4

4

.

Future Outlook and Market Implications

The projected pause in data center construction carries immediate implications for energy markets. Energy demand forecasts previously anticipated that computing infrastructure would drive significant future US natural gas consumption, estimating potential increases between 5 billion and 10 billion cubic feet per day3

3

. However, ongoing grid delays will likely pull actual consumption toward the bottom edge of those projections3

3

.

Future data center developments will likely require zero-impact operational models, designing facilities that offset their local grid burden, minimize water reliance, and prevent cost spikes for adjacent residential communities3

3

. Not all tracked projects will ultimately be built due to factors including financing challenges, local opposition, moratoriums, and turbine supply constraints1

1

.