Google plans gas-fired power for AI data centers in Texas with power from a 933-megawatt natural gas plant, based on filings reviewed by Cleanview and permit documents submitted by Crusoe Energy. The site sits in Armstrong County, close to the “Goodnight” campus, and it could pump out up to 4.5 million tons of CO₂ every year. That’s almost as much as all of San Francisco puts out annually, about 4 million tons, and satellite photos show they’ve already started building.
This move fits a growing pattern: Big tech companies like Google, Amazon, Microsoft, and Meta are leaning more on fossil fuels to keep up with the crazy energy appetite of AI. Google once promised to run completely on carbon-free energy 24/7 by 2030, but its latest sustainability report from 2024 tells a different story. Since 2019, the company’s emissions have shot up by 48%, mostly because of its rapid data center growth.
Why AI Is Increasing Fossil Fuel Demand
The rapid advancement in artificial intelligence has led to an increase in electricity demands. Advanced AI data centers use more electricity compared to conventional cloud computing data centers, owing to high-density GPU installations and ongoing training cycles.
Major drivers include:
- Generative AI workloads are growing at a fast pace
- Requirements for cooling GPU-based infrastructure
- Constraints in grid networks for renewable-rich areas
- “Always-on” reliability required for AI services
Based on projections from the International Energy Agency (IEA), electricity demands in data centers may amount to about 1,000 TWh by 2026, comparable to the overall electricity usage of Japan.
Also Read: Energy Shock From Iran War Shows Need To Move Beyond Fossil Fuels
Explanation of the Texas Gas Plant Project
Google’s new gas plant in Texas will be an off-grid solution for power production. In other words, it will provide electricity directly to the campus without adding any power to the broader grid.
Key technical details:
- Capacity: 933 MW natural gas plant
- Site: Goodnight AI data center campus
- Estimated emissions: 4.5 million tons CO₂ per year
- Type of energy generation: Dedicated energy production
According to Crusoe Energy’s permit filing, the Texas gas plant will fit into a vertically integrated compute-and-energy approach, where compute demand equals energy infrastructure.
Also Read: Geothermal Energy’s Potential To Replace Nearly Half Of Europe’s Fossil Fuel Power
How Big Tech Is Reshaping Energy Markets
Google plans gas-fired power for AI data centers, and is part of a larger trend throughout the technology industry where large companies are looking for fossil fuels to power their ever-expanding AI.
According to sources, Meta has developed a huge AI facility in Louisiana, which runs on natural gas, and Amazon is also building large multi-gigawatt data center complexes powered by gas-supported energy systems. Additionally, Microsoft has publicly announced new projects in West Virginia and Texas that will be linked to the natural gas supply chain.
| Source | Annual CO₂ Emissions |
|---|---|
| Proposed Texas gas plant (Google-linked) | ~4.5 million tons |
| San Francisco city emissions | ~4 million tons |
| Large coal plant (average) | ~5-6 million tons |
| 1 million U.S. homes (approx.) | ~6 million tons |
Also Read: Canada Set To Boost Investments In Ukraine’s Energy Sector
Climate Commitments Vs Operational Reality
Google’s climate policy has changed substantially during the last five years. The company stated its commitment to wind, solar, geothermal, and nuclear energy sources. Nevertheless, current developments show a different picture:
- 2023: Stopped claiming “carbon neutral”
- 2024: Increased emissions by 48% compared to 2019
- 2025: Rethought goals as “climate moonshots,” not fixed ones
Google plans gas-fired power for AI data centers, which shows an industrial paradox: AI development exceeds the speed of energy transition.
Grids with renewable energy sources cannot provide stable baseload power, forcing businesses to use gas as a transitional energy source.
Also Read: Saudi Arabia Plans To Build $2 Billion Solar Farms In Turkey, Turkish Energy Minister Says
What This Means Going Forward
By confirming Google’s plans for gas-fired power generation to support AI data centers, we see that national energy policy is now being directly influenced by AI infrastructure. Instead of relying on centralized grids, large companies are now developing their own private energy ecosystems built specifically to meet their computing requirements.
Natural gas is a reliable and scalable energy source. However, there are also long-term climate-related risks associated with natural gas as a power source, which may impede technology companies from achieving net-zero emissions.
Also Read: Greater Noida Microsoft Project Pulled Up For Illegal Groundwater Use
Energy Shift Within the Industry Sector
Some believe that this is now developing into a systemic trend rather than an experimental one because of the following reasons:
- Quicker timeline of the construction of gas power stations versus renewable projects
- Growing AI compute density needs
- Congested grids in locations with renewable installations, such as California and the Midwest
As stated by Cleanview researchers, this can be seen as one of the first significant investments into fossil-based energy facilities linked to an AI compute hub, particularly Google’s.
Also Read: India’s Power Demand Peaks At Five-Year High, Data Shows
A Turning Point for AI Energy Infrastructure
The introduction of gas-powered data centers for AI is changing how digital infrastructure is built and powered. As companies like Google continue to expand on their use of AI systems, the demand for energy will force businesses to make hard decisions between corporate sustainability commitments and their need to operate effectively.
Google plans gas-fired power for AI data centers, and this shows us that the future of artificial intelligence will depend as much on fossil fuels as it does on algorithms – at least in the short term.
Also Read: Clean Power Momentum Builds In India, Though Structural Gaps Continue
FAQs
1. Why is Google resorting to natural gas at its AI data centers?
The high power consumption demands for AI computing necessitate a steady power supply, and natural gas has an advantage over other renewable energy sources because it can be installed more rapidly and delivers consistent baseload power generation.
2. What’s the maximum annual CO₂ emissions from the Texas site?
The facility could produce 4.5 million tons of CO₂ annually, which is about equivalent to San Francisco’s annual emissions.
3. Is this an indication that Google has forsaken its green agenda?
It’s not yet clear. While Google maintains its goal of achieving carbon-free energy, recent events suggest a move towards mixed energy systems.
Also Read: Experts Say Cold Winter, Natural Gas Prices, And Data Centers Fueled Rise In U.S. Carbon Pollution

0 Comments