Data Centers in Texas: What’s True, What’s Not, and What’s at Stake
If you’ve scrolled through local news lately, you’ve probably noticed something strange. A year ago, most politicians in Texas were bragging about data center investment coming to their district. Now some of those same politicians are running against it. Zoning meetings that used to be sleepy are packed. National polls show opposition to data centers has jumped from about half of Americans to roughly three out of four in just twelve months.
At the same time, you’ve probably also heard the other side: these projects bring jobs, tax revenue, and investment that most towns would kill for.
So which is it? Are data centers going to drain our water, spike our power bills, and get built behind closed doors with no say from the people who live next to them? Or are they an economic opportunity North Texas would be foolish to turn away?
The honest answer is: both concerns are partly true, a lot of what’s circulating is flat-out wrong, and the real story is more interesting (and more fixable) than either side’s talking points suggest. This piece pulls together a recent local panel discussion held right here in Fort Worth, featuring a UT Arlington computer science professor, a TCU energy expert, and a longtime energy journalist, along with national reporting and polling on the broader backlash. The goal is simple: give you a clear, factual picture you can actually use.
First, What Is a Data Center? (Hint…you already use them daily)
Before we get into the arguments, let’s clear something up: most people who have opinions about data centers couldn’t tell you what actually happens inside one.
Here’s the plain-English version. A data center is a building full of computers, similar in concept to the laptop or phone in your hand, just on an industrial scale. Power comes in, usually from the electric grid or from an on-site power source. That power runs racks of servers and chips. Those servers heat up fast, so cooling systems pull the heat back out. And the data itself, meaning your search results, your Netflix stream, your work email, your car’s navigation, flows out the other end.
Here’s the part most people miss: data centers aren’t just for AI.
They’re the physical machinery behind basically everything digital you already rely on. Every time you say “Hey Siri,” or “Hey Alexa,” stream a show, check your bank balance online, or watch a doorbell camera notification pop up on your phone, a data center somewhere is doing the work. AI has added a new, much more power-hungry category of computing to that mix, but the buildings themselves have been quietly running the internet for two decades.
A lot of the current backlash comes from a simple realization: people assumed “the cloud” was some weightless, abstract thing. It isn’t. It’s real buildings, using real water and real electricity, usually built close to where people live.
Why North Texas Became Ground Zero
Texas isn’t an accident in this story. According to real estate research firm JLL, Texas has more data center capacity built or under construction than any state except possibly Virginia, more than double what’s happening in Virginia by some counts. North Texas specifically is already one of the top three or four data center markets in the country, and that was true even before the recent AI boom.
Why here? A few reasons came up repeatedly at the Fort Worth panel, where Nikki Morris, executive director of TCU’s Ralph Lowe Energy Institute, spoke directly to why the state has become such a magnet for this kind of investment:
- Land. Texas has a lot of it, and it’s easier to build here than in dense coastal markets.
- Power know-how. Decades of oil and gas experience means Texas has real expertise in building and delivering energy infrastructure fast.
- A flexible power grid. Unlike much of the country, Texas runs its own grid (ERCOT) as a competitive, energy-only market. That structure makes it easier and often cheaper to get new power projects built here than in most other states.
- Capital and entrepreneurs. Texas has investors and developers who know how to move quickly on big infrastructure projects.
None of that means every project is a good one. But it explains why this conversation is happening here first, and loudest.
The Good: What Data Centers Can Actually Deliver
Set the noise aside for a moment and look at what’s actually happened in towns that got this right.
Quincy, Washington started attracting data centers back in 2006. It’s a farming town of about 8,500 people. Today it hosts 30 data centers, supporting roughly 900 direct jobs plus several more local jobs for every one of those. Data centers now fund more than half of the town’s property tax base, which has paid for a new pool, a $120 million high school, and a sports complex. The town’s poverty rate dropped from nearly 30% in 2012 to about 6% in 2024.
Loudoun County, Virginia, home to the world’s densest cluster of data centers, tells a similar story at a bigger scale. Data centers sit on about 3% of the county’s land but generate around 40% of its budget. The county has cut property taxes every year for a decade, down to some of the lowest rates in the region, while still outspending its neighbors on schools and public safety.
These aren’t cherry-picked press releases. They’re documented outcomes: lower taxes, better-funded schools, and real jobs, not just during construction but ongoing technical and engineering work.
Closer to home, this same pattern is starting to show up. West Texas has a Microsoft-backed project expected to reach nearly 2.7 gigawatts, using brackish (non-drinkable) water instead of freshwater, and Microsoft has published detailed local tax and job figures for similar projects. A Riot Platforms campus in Corsicana is scaling toward a full gigawatt of capacity. These are the kinds of projects that, done right, function less like a factory moving in and more like a new, well-funded institution planting long-term roots in a community.
None of this happens by accident, though. It happens when a project is well-run, transparent about its plans, and genuinely invested in the community around it. Which brings us to the concerns worth taking seriously.
The Bad: Real Concerns, Not Imaginary Ones
It would be dishonest to wave away every worry as overblown. A few concerns are legitimate and deserve real answers, not just reassurance.
Water use, especially in summer.
This is where things get genuinely nuanced. Newer data centers increasingly use “closed loop” cooling, where water carries heat away from the servers, gets cooled, and gets pumped back in, over and over, instead of being used once and discarded. That’s a real improvement. But as Dr. Mohammad Islam, a UT Arlington computer scientist who studies this exact issue, pointed out at the Fort Worth panel: even closed-loop systems aren’t zero-water, and the water use that does happen tends to be heavily concentrated in the hottest summer months, not spread evenly across the year. A company can accurately say its facility uses “10 times less water annually” while still requesting a very large daily water allowance for July and August. Both things can be true at once, and that nuance rarely makes it into either the marketing or the outrage.
Grid strain and the pace of buildout.
Texas’ grid operator, ERCOT, is trying to plan for a wave of new demand while the actual size of that wave is genuinely unclear (more on that in the misinformation section below). Multiple panelists agreed the honest answer is that the grid physically cannot be built out as fast as demand is arriving, which means delays and growing pains are likely regardless of how well-intentioned any individual project is.
The “behind the meter” loophole.
Some data centers are built to generate their own power entirely, without connecting to the public grid. This can mean faster construction, but Dr. Islam raised a pointed concern: because current regulations and moratoriums are mostly written around grid-connected facilities, a project that stays off the grid could sidestep water and environmental rules altogether, including running on gas power with fewer restrictions. That’s a real regulatory gap, not a hypothetical one.
Secrecy and NDAs.
Nationally, a lot of data center negotiations with local governments have happened under non-disclosure agreements, which makes it hard for residents to know what they’re actually getting, or giving up. Microsoft recently announced it’s ending this practice with local governments, which is a meaningful signal, but it’s the exception industry-wide rather than the rule so far.
These concerns share a common thread: they’re not really about whether data centers are good or bad in the abstract. They’re about whether a specific project is well-designed, honestly disclosed, and held to real standards. That distinction matters enormously, and it’s exactly where most of the public conversation goes wrong.
The River of Misinformation: Sorting Fact From Fear
Here’s where things get messy, and where a little fact-checking goes a long way.
The water math is often wrong, sometimes by a lot.
A widely circulated book on AI’s environmental footprint miscalculated data center water use by a factor of 1,000. The error has been acknowledged, but the original, wildly inflated number is still being quoted as fact in plenty of arguments today. If someone cites a jaw-dropping water statistic, it’s worth double-checking the source before repeating it.
The “data centers are spiking your power bill” claim doesn’t hold up well under scrutiny.
When researchers actually compared where electricity prices are rising fastest against where data centers are being built, they didn’t find a strong statistical link. That doesn’t mean electricity costs are a non-issue. It means the impact depends heavily on how a specific project is built and who pays for the grid upgrades it needs, not on some universal rule that data centers always raise your bill.
The eye-popping “queue” numbers are mostly noise.
You’ll often hear that hundreds of gigawatts of data center projects are “in the pipeline” nationally or in Texas specifically. In reality, a lot of that number comes from the same projects being counted two or three times as developers apply in multiple places to hedge their bets. At the Fort Worth panel, veteran energy journalist Jennifer Warren estimated that the realistic buildout in Texas through 2030 is closer to 12 to 15 gigawatts, a small fraction of the inflated headline figures. Big numbers make for scary headlines, but they’re not the same as reality.
A lot of the anger isn’t really about data centers at all.
National polling found people are actually more comfortable living near a nuclear power plant than near a data center, even though nuclear plants carry real, well-documented safety and waste considerations that data centers simply don’t. That gap is a strong hint that the backlash isn’t purely a rational risk calculation. Much of it is a stand-in for broader frustrations: worry about AI replacing jobs, years of accumulated distrust toward big tech companies, and a general feeling that communities don’t get a say in what gets built around them. Those are legitimate feelings. They’re just a different conversation than “is this specific project safe.”
Media coverage skews negative on both data centers and the industries that came before them.
This was something the panelists compared to how oil and gas coverage has worked for decades: the bad actors and worst-case stories get the headlines, while the routine, well-run projects that quietly fund schools and employ locals don’t make the news at all. That doesn’t mean the bad stories are fake. It means they’re not the whole picture.
What Actually Works: Rules, Not Bans
Here’s the most encouraging finding in all of this, and it rarely makes headlines: most people don’t actually want data centers banned outright. They want them held to clear, enforceable standards.
Recent national polling found 57% of people prefer clear rules that allow responsible projects to move forward, compared to just 23% who want a blanket ban on new projects. When people were shown a hypothetical project that met real disclosure, cost, and public review requirements, most said they’d support it moving forward. And when asked whether companies should be required to pay for the grid upgrades their own facilities need, rather than passing that cost to everyone else’s electric bill, support jumped to nearly 80%.
That lines up closely with what came out of the Fort Worth panel. Instead of trying to rank companies as simply “good” or “bad,” the more useful approach is a checklist: who owns the project, how is it financed, what’s the plan for decommissioning it decades from now, what water and power commitments is it making, and is any of that information actually verifiable by an independent source, not just the company’s own press release.
Some developers are already moving this direction voluntarily. Meta has launched a billion-dollar fund for community initiatives near its sites. One of its Louisiana projects generated enough local tax revenue to fund $50,000 teacher bonuses. OpenAI paired a recent Ohio project with tens of millions in community grants and technical education credits. Microsoft has ended NDAs with local governments. These aren’t universal industry practices yet, but they show what “good operator” behavior looks like in practice, and they give communities and regulators something concrete to ask for from everyone else.
What You Can Do
If you want to form your own informed opinion instead of just absorbing whatever shows up in your feed, a few things help:
- Research the specific project details, not the industry-wide headline.
Ownership, financing, water source, and community commitments tell you far more than a scary or glowing national statistic. - Treat both extremes with skepticism.
If a claim sounds like “this will cause a drought” or “this will definitely lower your taxes,” it’s probably oversimplified in one direction or the other. - Show up locally.
Zoning meetings, panels like the one held recently in Fort Worth, and city council sessions are where the actual, project-specific decisions get made, not on social media. - Look and ask for independent verification, not just company-published numbers.
Universities and independent researchers are increasingly stepping into this role, and that kind of nonpartisan information is exactly what’s been missing from the national conversation.
My Prediction on Data Centers in Texas
Data centers aren’t going away, and pretending North Texas can simply opt out of this economic shift isn’t realistic, or necessarily in our own best interest. The opportunity here is genuinely large: good-paying jobs, major tax revenue, and long-term infrastructure investment, the kind of thing that funded new schools and lower taxes in towns that got it right.
But “getting it right” is the whole game. The difference between a data center that strengthens a community for decades and one that just leaves behind a substation and a higher electric bill comes down to a few unglamorous things: real transparency, enforceable community commitments, and a willingness from all of us to ask sharper questions than the headlines are asking.
North Texas has the land, the power expertise, and the economic momentum to do this well. Whether we actually do comes down to demanding good rules now, not blanket bans, and not blind acceptance either.
P.S. Let’s not forget China
Chinese AI labs (Kimi K3, DeepSeek, and Qwen) are closing the gap with America’s best models shockingly fast this year. What analysts used to call a six to eight month U.S. lead has shrunk to something closer to weeks. Frontier AI runs on compute, and compute runs on data centers. Every project we slow down here doesn’t just cost North Texas jobs and tax revenue, it’s ground we’re not covering in a race that has real national security stakes: whose AI systems the rest of the world ends up building on, and whose rules get baked into them.
I guarantee you China is not waiting around for us to sort out our zoning meetings.
Wouldn’t you rather see this capacity built here, and built responsibly, than watch it get built somewhere else, or by someone else?
Sources: National reporting, polling, and industry data referenced in this piece draw on “Why Everyone Suddenly Hates Data Centers,” The AI Daily Brief, August 21, 2026 (also available via the AI Daily Brief episode page), along with the Fort Worth panel discussion linked above. And my personal opinion.