Amazon’s Texas Data Center Power Plant Could Be the US’s Biggest CO2 Emitter

Amazon is planning a data center in Pecos County, Texas that includes an on-site natural gas power plant permitted to release 33 million tons of carbon dioxide per year, according to the New York Times. That would make it the largest single source of climate pollution in the United States. The move comes as Amazon's overall carbon emissions rose 16% last year, a trend the company partly attributes to the energy demands of AI, and while it still maintains a pledge to reach net-zero emissions by 2040.

What happened

Detail	Fact
Location	Pecos County, Texas
Power source	On-site natural gas plant
Permitted CO2 emissions	33 million tons per year
Comparison	More than any other power plant in the US
Amazon emissions trend	Up 16% last year
Amazon net-zero pledge deadline	2040

The New York Times reported that Amazon’s planned Pecos County facility would include a natural gas plant with a permitted ceiling of 33 million tons of CO2 annually. No other US power plant currently emits at that scale, which means this single facility could reshape national emissions statistics on its own.

An Amazon spokesperson confirmed the data center will “be powered by new on-site generation that won’t raise electricity costs for Texas families.” That framing is deliberate: data centers have drawn growing political pushback partly because of their effect on local electricity prices, so Amazon is positioning this as a consumer-friendly move even as the environmental cost is substantial.

How did Amazon’s carbon footprint get here?

Amazon reported a 16% increase in carbon emissions last year, which the company has linked in part to AI infrastructure demands. The direction matters: the company co-founded the Climate Pledge, a commitment to eliminate its carbon emissions by 2040. Moving the opposite way in 2025 and then greenlighting one of the most carbon-intensive power plants in US history is a significant contradiction.

The spokesperson acknowledged the gap, saying “the world looks different now than when we co-founded the climate pledge,” while also insisting “our commitment hasn’t changed.” That’s a hard claim to hold alongside a 33-million-ton annual emissions permit.

Amazon is not alone. Other large tech companies are also backing natural gas plant development to keep pace with data center power demand driven by AI workloads. The pattern across the industry points to a structural problem: the energy required to train and serve AI models is growing faster than the capacity of renewable sources to supply it.

Why it matters

For businesses watching AI infrastructure costs and sustainability commitments, this story has two practical angles. First, the energy cost of AI is real and escalating, and it is now showing up in ways that affect national policy and utility markets. Second, corporate climate pledges are under genuine pressure when energy demand from AI grows at this pace.

For anyone building AI-dependent products or services, the energy footprint of the models you call through an API ultimately traces back to decisions like this one. That is not a reason to stop using AI tools, but it is a reason to track how your AI provider handles its energy sourcing as regulation tightens. You can read our broader coverage of how AI labs are navigating risk thresholds for context on how these infrastructure decisions intersect with policy.

The political dimension is also worth watching. Data centers already face opposition over electricity costs. Adding a facility that doubles as the country’s largest CO2 source gives regulators and local governments a much sharper target. Permitting fights, state-level emissions rules, and potential federal scrutiny could all affect the timeline and viability of this and similar projects.

Our take

The “our commitment hasn’t changed” line from Amazon is doing a lot of work. Committing to net-zero by 2040 while permitting the largest CO2-emitting plant in the country is not a contradiction that communication can paper over for long. At some point the math has to work.

For our clients thinking about AI adoption and sustainability, the honest answer is that the energy cost of AI is currently being absorbed by infrastructure providers in ways that are only partially visible. That may not stay invisible. Carbon taxes, emissions reporting requirements, and supply chain disclosure rules are all moving in the direction of making that cost explicit. Businesses that start tracking their AI-related energy consumption now will be better positioned when reporting becomes mandatory.

If you are evaluating how to integrate AI tools into your operations responsibly, our AI integration work includes a look at provider choices and their downstream implications, not just raw capability.

What to do about it

1. Ask your AI vendors directly about their energy sourcing and emissions reporting policies.
2. Audit how much of your cloud and AI spend maps to high-carbon infrastructure regions.
3. Track proposed legislation in your operating markets around data center emissions and carbon disclosure.
4. Factor carbon cost into your AI vendor evaluation, not just price and performance.

The cost of AI’s energy appetite is becoming harder to ignore, and the companies that plan for it now will face fewer surprises later.

Source: TechCrunch · AI