Google’s rapid expansion of AI is reshaping how the company sources energy, revealing reliance on natural gas even as it continues to promote long-term clean energy ambitions.
Prompting the shift is a large data center campus under development in Armstrong County, Texas. Built in partnership with Crusoe Energy, the project includes plans for an on-site gas-fired power facility capable of generating hundreds of megawatts of electricity. Research from the organization Cleanview indicates that the plant could release millions of tons of carbon emissions annually, placing it among the more emission-intensive power sources tied to a single data center development.
The proposed gas plant would operate independently of the broader grid, supplying power to portions of the campus. While Google has confirmed its involvement in the data center, it has not finalized an agreement to purchase energy from the gas facility, leaving open the possibility of adjustments to the project’s energy mix.
Microsoft, Meta, Amazon
Clearly, the escalating energy demands of AI systems are prompting facility developers to scramble. Training and running advanced models requires vast computational resources, and each new data center can consume as much electricity as a small city. Google alone used more than 30 million megawatt-hours of electricity in 2024, a figure that has climbed sharply in recent years.
To meet this demand, companies are moving beyond traditional grid connections. Increasingly, they are building dedicated energy infrastructure directly alongside their facilities, a model known as behind-the-meter power. This approach reduces wait times for grid access and offers greater control over supply, but often relies on fossil fuels due to their availability and scalability.
Microsoft, Meta, and Amazon are all pursuing similar strategies, incorporating natural gas into their energy portfolios as they rapidly build out AI capacity. In some cases, companies are pairing gas generation with renewable sources such as wind and solar, though fossil fuels often provide the baseline power needed for continuous operation.
This shift departs from earlier approaches, when major tech firms emphasized matching electricity use with renewable energy purchases. Google was widely regarded as a leader in that model, investing heavily in wind and solar projects and setting ambitious targets for carbon-free operations.
But in recent years, those commitments have become more difficult to maintain. The company has reported rising emissions linked to its growing data center footprint, even as it continues to fund clean energy initiatives.
Natural gas, while carbon-intensive, offers a reliable and readily deployable source of power at a time when grid infrastructure is struggling to keep pace with demand. Renewable energy projects, by contrast, can take years to permit and build, and often require storage solutions to ensure consistent output.
An Active Debate
Supporters of the approach argue that gas can serve as a transitional resource while cleaner technologies mature. Companies including Google are investing in emerging options such as advanced nuclear, geothermal energy, and long-duration battery storage, though these remain in earlier stages of deployment.
Critics, including some lawmakers, question whether expanding fossil fuel use undermines climate commitments. Recent inquiries from US senators have pressed tech companies to explain how their energy strategies align with global emissions targets, particularly as the scale of AI infrastructure continues to grow.
The tension highlights an ongoing dilemma. AI is transformative technology with economic and strategic importance, attracting massive investment. Yet the infrastructure required to support it carries significant environmental costs. Whether cleaner energy solutions can catch up remains an open question, but in the interim, natural gas is playing a larger role than many had anticipated.
