Artificial Intelligence has become synonymous with progress, productivity, and the future.
Everyone wants to use it, but few want to know what’s actually behind it all.
And that’s exactly where a pretty uncomfortable paradox lives: while millions of people use AI tools every day without a second thought, the communities living next to data centers — those massive warehouses that make everything run — are giving a resounding no to this infrastructure in their backyard.
Texas has become the latest stage for this conflict.
In Hood County, where more than 80% of votes went to Trump in 2024, residents packed public meetings to block data center projects in the area.
The reason goes way beyond politics.
The fear is real and tangible: impact on water supply, strain on the energy grid, and the transformation of a rural landscape that people call home.
And this scenario raises a question that reaches far beyond the borders of Texas and challenges the current model of tech expansion.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
So, who really foots the bill for AI’s advancement?
That’s what this article is about. 👇
When local resistance is already blocking projects
Community opposition didn’t just stay on paper or in heated public meeting speeches. It has already produced concrete results. In Hood County itself, local commissioners responded to public pressure by tightening development rules, requiring far more detailed disclosures about the water and energy consumption of each proposed project. In other words, anyone looking to build there now has to open up their numbers before laying the first stone.
And in some cases, projects were stopped cold. Early in the American summer, county commissioners voted 3-to-2 to revoke a conceptual plan that had already been approved for an 862-acre data center. The project, called Fort Spunky, planned to use 100 megawatts of electricity and was projected to consume up to 20,000 gallons of water per day. Numbers that alarmed residents and ultimately proved enough to kill the initiative. This kind of reversal shows the debate has moved out of the abstract and into real infrastructure decisions.
What data centers have to do with your faucet
It sounds like an exaggeration, but it’s not. The data centers powering artificial intelligence operations consume staggering amounts of water to keep servers cool. This process, called evaporative cooling, basically works by evaporating water in cooling towers to bring down machine temperatures. The more processing, the more heat generated, the more water consumed. And with the explosion of AI usage, this cycle has only accelerated.
To get a sense of the scale, a single large-scale data center can consume millions of liters of water per day — depending on the size of the operation and the local climate. Some industry reports indicate that globally, the data center industry already consumes billions of liters of water per year, and that demand is expected to grow alongside the rise of generative AI. Companies like Google, Microsoft, and Meta publish sustainability reports acknowledging this water impact, but the numbers are still hard to track precisely because each facility operates under different conditions.
The problem gets even more serious when these projects arrive in regions already dealing with water scarcity or that depend on underground aquifers for domestic and agricultural supply. In the case of Texas, residents in rural areas fear that industrial water consumption by large tech installations will compromise what they have available to irrigate crops, raise livestock, and simply turn on the tap at home. This isn’t an irrational fear — it’s a legitimate concern about prioritizing the use of a finite resource.
Energy: the other side of the equation
If water is already a point of tension, energy enters the conversation with even more weight. Data centers are voracious electricity consumers, and demand only increases as more artificial intelligence models are trained and put into production. It’s estimated that training a single large language model — like the ones behind the most popular AI assistants — can consume the equivalent of the electricity used by hundreds of homes for entire months. And that’s just the training; inference, meaning the processing of each query made by users around the world, happens continuously, 24 hours a day, 7 days a week.
Local power grids, especially in less industrialized areas, simply weren’t designed to absorb that kind of load overnight. When a major tech company announces the construction of a server campus in a rural area, the impact on local energy infrastructure can be immediate and drastic. For residents who depend on that energy for everyday life, this isn’t abstract — it’s very real.
The tech sector has promised to move toward renewable sources, and some companies genuinely invest in solar, wind, and even nuclear energy to power their operations. But the transition isn’t happening at the speed AI’s growth demands. In practice, many data centers still rely on mixed energy grids that include fossil fuels to ensure supply stability. That means every prompt typed, every image generated, and every response delivered by a language model carries, to some degree, a carbon footprint that rarely shows up on the user’s screen. 🌍
Using AI isn’t the same as supporting everything about it
An interesting point in this debate is that liking AI and not wanting a data center in your backyard aren’t contradictory. Stephanie Justice, policy communications manager at Americans for Responsible Innovation, who closely follows the data center issue, sums up this idea well. According to her, using AI and wanting infrastructure to be built responsibly, with people having a voice, are not opposites.
Justice points out that AI is happening, and people need to know how to use it and integrate it into their careers and daily lives. But that doesn’t cancel out their right to have a say in how and where data centers get built. It’s a simple and powerful line of reasoning: adopting the tool doesn’t mean blindly accepting all the side effects it brings to certain communities.
There’s also an important warning against the easy assumption that mass usage equals mass support. As AI gets embedded in search engines, workplaces, and other products people already use, it’s become practically impossible not to participate. Not everyone who uses the technology regularly is necessarily a fan of it. Many workers end up pressured to become fluent in AI because it’s the tool that’s available, even if, for them, that doesn’t justify the costs data centers impose on neighboring communities.
Where the support for these projects comes from
It’s not all opposition, and it would be unfair to paint the picture with only one side. These projects find support among organized labor unions, particularly in construction trades, whose workers directly benefit from the work needed to build and maintain data centers. This, by the way, presents a particular political challenge for Democrats, who see unions as a core part of their voter base.
On the industry side, Dan Diorio, executive vice president of state policy and government affairs at the Data Center Coalition, suggests that the economic benefits of a data center go well beyond the initial construction jobs. According to him, these projects generate substantial tax revenue that funds schools and supports millions of jobs spread across construction, manufacturing, technology, and local services. It’s the classic regional economic development argument — one that doesn’t always convince the people living next door to the project.
Sustainable development: promise or contradiction?
The term sustainable development shows up in virtually every annual report from major tech companies. Net-zero carbon goals, clean energy commitments, water offset programs — all of it exists, and in some cases represents real progress. But there’s a fundamental tension that no corporate report fully resolves: the speed of AI infrastructure expansion is currently outpacing the ability to make that expansion genuinely sustainable. It’s like trying to build the road while the car is already going full speed.
What’s happening in Texas isn’t an isolated case. Communities in various parts of the world have already gone through similar conflicts involving data centers. In some places, local governments even imposed temporary moratoriums on new projects precisely because public infrastructure — water, energy, and transportation — couldn’t keep up. The public pushback doesn’t come from an aversion to technology itself, but from a very rational perception that the benefits stay concentrated among a few, while the costs get spread broadly across surrounding communities.
The path toward real sustainable development in this sector runs through some structural changes that are being discussed, even if cautiously. Among them are the large-scale adoption of cooling technologies that don’t rely on potable water, direct investment in upgrading local power grids before installing major infrastructure, and more transparent policies around resource consumption disclosure from companies. On top of that, the efficiency of AI models themselves is a factor: smaller, more efficient, and specialized models can deliver comparable results at a fraction of the energy cost of large-scale generalist models.
What’s at stake beyond Texas
The discussion that erupted in Hood County is really a symptom of something much bigger. The rapid growth of artificial intelligence is creating demand for physical infrastructure at an unprecedented scale, and the regions receiving this infrastructure aren’t always the ones benefiting most from the technologies it supports. There’s a clear asymmetry between who uses AI and who bears the side effects of its existence. That asymmetry isn’t neutral — it has geographic, economic, and social dimensions that deserve far more attention than they get in the public conversation about technology.
Thinking about the future of AI responsibly means bringing into the conversation not just the engineers who develop the models and the companies that sell them, but also the communities living near the infrastructure that makes it all work. It means honestly asking how to ensure that technological progress doesn’t come at the expense of access to basic resources like clean water and reliable energy for people who are already vulnerable. And it means recognizing that sustainable development can’t just be a nice section in an annual report — it needs to show up in the decisions about where, how, and at what pace this infrastructure grows.
Artificial intelligence has enormous potential to solve complex problems, including environmental challenges. But that potential only truly materializes if AI’s own expansion is guided by the same standards of responsibility it promises to apply to the world. Texas is reminding everyone of that right now — but the question will keep coming up long after the cameras move on. 💡
