The Data Center Debate Is Asking the Wrong Question

It should not be “whether.” It’s “how.” And maybe more importantly, “where.”

To describe the state of politics regarding data centers as a “fight” would be a considerable understatement. It is a political perfect storm, and it is already here.

To review the current landscape:

Data centers are at the front and center of the public’s political discourse, here in Texas and across the country. Here, the legislature and the Governor have all made the issue a top priority. San Marcos became the first Texas city to ban them outright. Hill County passed a moratorium, got hit with a $100 million lawsuit from a developer, and reversed course. In urban areas across North Texas and beyond, neighbors are showing up at council meetings in force, demanding that projects get halted - sometimes after those same projects were already approved through the proper process. And a Texas Tribune analysis found that nearly half of all planned data centers in the state are now being sited in unincorporated areas where counties have no legal zoning authority to manage them.

Elsewhere, Governor Hochul of New York signed an executive order yesterday making New York the first state in the nation to impose a full moratorium on new hyperscale data centers. Other states may follow suit.

All of this lands in Austin in 2027. The legislature will walk into a room where every one of these pressure points has been building for years, and they will have to decide what Texas’s answer is.

I have a perspective on this that is a little different from most of what you are reading. I have been hearing about data center opposition starting on the campaign trail and afterwards, for the better part of a year. And for the last year, I have been telling people in the industry - some of whom have asked for my political read on the situation - the same thing: however difficult data center development is right now, it will get more difficult. Because everyone running for office is getting an earful everywhere they go. The politics of this were going to get harder before they got easier. That is now happening.

So let me offer a framework, starting with something that might surprise you coming from someone who spent over a decade governing a growing city.

The people pushing back the hardest on data centers are not wrong. Not entirely.

What’s valid and what’s being exaggerated

There are real concerns here, and a governing conservative has to acknowledge them honestly. Data centers are not all quiet neighbors. A large hyperscale facility draws industrial-scale electricity, generates significant heat, consumes enormous amounts of water for cooling (at least initially), and runs twenty-four hours a day. A University of Texas study recently found data centers could account for 3 to 9 percent of the state’s total water use by 2040. For a state with serious aquifer pressure and a growing agricultural sector, that is not a number you wave away.

The grid pressure is real too. Hundreds of large data centers representing roughly 410 gigawatts of potential electricity demand have asked to connect to ERCOT through 2032. ERCOT’s projected peak demand this summer is around 92 gigawatts. The data center pipeline alone represents more than four times our current grid’s total capacity. If that demand lands on the grid without adequate infrastructure behind it, the cost lands on residential bills. That is unacceptable, as Governor Abbott has acknowledged.

But here is what is also true: not every data center is a hyperscale behemoth. The public debate has collapsed a very wide spectrum of facilities into a single monolith, and that collapse is producing bad policy instincts.

A hyperscale campus operated by a tech giant, drawing 100 megawatts or more, with its own water and power demands, is a categorically different thing from a regional colocation facility that uses no more infrastructure than a mid-size distribution warehouse and serves local businesses that need secure, reliable compute nearby. A policy environment that treats every data center as an existential threat to the grid ends up blocking the small, nimble facilities that Texas should actually be trying to attract, while the hyperscale players with deeper pockets find workarounds or move to another state.

Some of the claims being made about environmental impact are also being exaggerated, in part because a great deal of the opposition is not entirely organic. More on that in a moment.

The zoning process is broken. The rural workaround is not a fix.

I spent three years as Mayor of Southlake and another six on the Southlake City Council. I have sat through hundreds of hours of zoning hearings. I know exactly how this dynamic plays out, because I have watched it in different forms more times than I can count.

A developer comes in with a project. It goes through staff review, the planning and zoning commission, public hearings, and a council vote. At every step, the process is supposed to weigh the relevant facts: land use compatibility, infrastructure capacity, environmental impact, economic benefit. That is what the process is for. But before the vote ever happens, the community groups mobilize, the Facebook groups organize, the yard signs go up, and the council chamber fills with residents who are angry and organized and showing up on a Tuesday night. Decision makers face enormous pressure to listen to the torches and pitchforks rather than to the policy. Good planning gets drowned out. We are watching this play out right now in more than one Texas city where data centers have not gotten anything close to a fair hearing.

The problem is not that community members have opinions about what gets built near their homes. They should, and the process should hear them. The problem is that the process gets co-opted. Organized, motivated opposition floods the zone and makes it nearly impossible for decision makers to do the hard work of weighing costs and benefits honestly. The loudest and most coordinated voices - rarely the most reasonable ones - end up setting the outcome.

Developers, who are rational actors, draw the rational conclusion. If the municipal process is too easily captured, go somewhere the process can’t reach.

Texas counties, unlike cities, do not have general zoning authority. A developer who knows the city zoning fight is going to produce years of delay and a real possibility of reversal by a mobilized opposition has a legal alternative: find land in an unincorporated county, acquire it, and build. No zoning commission. No public hearing. No council vote. The Texas Tribune recently found that nearly half of all planned data centers in the state are now headed to unincorporated areas for exactly this reason.

I understand why developers make that choice. But the solution to a broken city zoning process is not to push industrial-scale development into rural communities that have no legal tools to manage it. Those communities didn’t choose to forgo zoning authority. Texas law made that choice for them. That is not consent. When a massive industrial facility appears in what was previously agricultural land, next to farms and homes that had no warning and no meaningful recourse, that community’s anger is just as legitimate as any urban neighborhood’s.

Hill County tried to act. The commissioners voted for a moratorium. A developer filed a $100 million lawsuit and the county reversed course. Counties are fighting with no weapons. They cannot zone. They cannot moratorium without legal exposure. Into that void, the industry is building.

Both failure modes are real. The urban process gets captured by organized opposition. The rural workaround sidesteps the legitimate planning questions entirely. Neither is a policy. Both are symptoms of the fact that Texas does not yet have a coherent answer to where this infrastructure belongs and what obligations it carries to the communities around it.

One more complicating layer

The organic community opposition is real. I want to be clear about that, because what I am about to say should not be used to dismiss legitimate neighbors with legitimate concerns.

But some of what is driving this debate is not organic at all. Some reporting has pointed to more than $39 million in documented foreign funding flowing through activist organizations that have signed onto the congressional moratorium push. I am not in a position to verify every thread of that, and I want to be careful not to overstate it. What I can say is this: China’s interest in this debate is not hard to understand. Its interest is not simply to slow American AI infrastructure. It is to own AI compute 100%. If American development stalls long enough, the global capacity to run and train frontier AI models tilts toward an adversary that is not holding public hearings about whether to build, and does not have our best interests in mind.

Real community anger and foreign-funded amplification are not mutually exclusive. Both can be true at once. Texas lawmakers sitting across from a packed council chamber should at least be aware that not everything in that room grew there on its own.

Do we really want to follow New York’s lead?

Which brings me to the most important political data point of this whole debate.

This week, Governor Kathy Hochul of New York signed an executive order imposing the first statewide moratorium on new hyperscale data centers in the country. The New York legislature had passed the underlying bill 44 to 16 in the Senate and 102 to 39 in the Assembly.

New York is now officially closed for this business.

And here in Texas - Texas, which has built its entire economic identity on being the alternative to New York’s regulatory instincts - some of the voices calling for the same thing are Republicans.

I want to sit with that for a moment. When the policy reflex of Texas Republicans on economic development starts to rhyme with Kathy Hochul’s, something has gone sideways. I am not saying the underlying concerns are wrong. I have already said they are not. But the answer to those concerns should not be the same answer that a Democratic governor in the most regulated large state in the country just reached. Texas has competed with New York for decades by being a place where businesses can operate without endless regulatory barriers. A Texas moratorium, or the functional equivalent of one produced by a hostile 2027 session, hands New York its best economic development pitch in a generation.

The compute that gets blocked in Texas does not disappear. It goes to Virginia, to Georgia, to Arizona. Or it goes to a country that is not holding public hearings about whether to build it. DeepSeek has been pricing its AI models more than thirty times cheaper than American competitors on a per-token basis. That is a deliberate strategic bet that cheap, abundant compute is a geopolitical advantage. Slowing our own build-out is not a neutral act. It is a concession to the adversary running the influence campaign.

Governor Abbott has the right instinct. The market can do the rest.

To his credit, the governor has found the correct framing on the how question. His directive to the PUC and ERCOT in June was clear: data centers pay for their own electric infrastructure and interconnection costs. They do not pass those costs to residential ratepayers. They bring water-efficient technology. They report their usage.

That is exactly right. Texas has spent the last five years hardening and strengthening the ERCOT grid after the failures of Winter Storm Uri. That work cannot be undone by an unconstrained industrial build-out that treats the grid as a free resource. Our infrastructure has to support our population first. The “bring your own power and water” standard protects that priority without slamming the door on development.

SB 6 from last session moved in the same direction: interconnection cost-sharing for large loads, disclosure requirements, and the obligation to disconnect from the grid during emergencies rather than drawing capacity away from homes and hospitals. That is good governing. It does not ask whether. It asks how.

And now the market can do what markets do. With those guardrails in place, developers will look for stranded power - generation capacity that exists but is not fully utilized. West Texas has it in the oilfield, where associated gas is flared and power generation goes underused. Rural areas with excess transmission capacity become attractive for the right-sized project, not the hyperscale campus that overwhelms local water and roads, but the colocation facility that fits what the community’s infrastructure can actually support. Small, nimble data centers that bring their own power and manage their own water have a cleaner path. That is not a coincidence. That is the “bring your own” standard working as intended.

Texas should be actively trying to attract those facilities. They are the development that fits. The ones that don’t fit should be required to make the infrastructure whole before they build - not blocked, but not subsidized either.

The 2027 risk

The legislature’s interim charges already signal where the session is headed: reviewing the data center sales tax exemption that grew from $14 million in 2013 to a projected $3 billion in foregone revenue, scrutinizing water usage, monitoring SB 6 implementation. Some of that review is overdue. Texans deserve a clear return on billions in tax expenditures.

What is not overdue is a moratorium. What is not overdue is treating this as a culture war rather than a planning problem. What is not overdue is giving China and Russia the outcome they have been paying to manufacture.

The goal for 2027 is a framework - on how and where - that gives communities a legitimate voice without handing veto power to the loudest Facebook group, gives rural counties something other than a choice between helplessness and a lawsuit, gives developers a predictable process so the rural workaround loses its appeal, and gives Texas the infrastructure it needs to compete in the AI economy for the next twenty years.

That is a governing problem. It is exactly the kind of problem the current Republican Party has been struggling to prioritize over the performance of outrage.

I have been saying for a year that this gets harder before it gets easier. It is now harder. The question is whether Austin can get it right before the storm passes and we find out the hard way what we lost.

The compute is coming. The only question is whether it lands here, on terms that work for Texans, or somewhere else, on terms that work for everyone else.