San Marcos Rejects Data Center: Drought Reality
San Marcos rejects a major data center due to drought, raising questions about infrastructure in climate-sensitive regions. #DataCenter #Sustainability
Water is the new zoning issue.
When San Marcos, Texas, rejected a rezoning application for a proposed 200-acre data center, the decision sent a signal that goes well beyond one city council vote. It confirmed what infrastructure developers in the Sun Belt have been quietly dreading: drought isn't a background risk factor anymore. It's a deal-killer.
The rejection highlights a tension that's been building for years — the explosive growth of data center development colliding head-on with the hard limits of freshwater supply. And with AI workloads driving power densities (and cooling demands) to levels unimaginable a decade ago, that collision is only going to get louder.
What Was Actually Proposed — and Why San Marcos Said No
The site in question was a 200-acre parcel slated for rezoning to accommodate a large-scale data center. Details on the specific developer or end tenant weren't disclosed in available reporting, but the scale alone tells you something: 200 acres in Central Texas is a serious commitment. Projects at that footprint typically represent hundreds of megawatts of capacity and billions in capital investment.
San Marcos sits in the Hill Country region, an area that has faced persistent drought conditions over the past several years. The Edwards Aquifer — the primary water source for much of Central Texas — has been under increasing stress, with mandatory pumping restrictions triggered repeatedly as aquifer levels drop. The city's rejection wasn't a rejection of economic development. It was a recognition that approving a massive water consumer during an active drought crisis would be fiscally and environmentally reckless.
Data centers aren't always the water hogs they're assumed to be — but the ones that rely on evaporative cooling absolutely are. A large hyperscale facility using evaporative cooling towers can consume anywhere from 1 to 5 million gallons of water per day. For a municipality already watching its reservoir levels, that's not an abstraction. That's a constraint with teeth.
Drought Is Rewriting the Site Selection Playbook
The San Marcos case isn't an outlier. It's a preview.
Across the American West and Southwest, water availability is becoming as critical a site selection variable as land cost, power access, and fiber connectivity. Arizona's Maricopa County — for decades one of the hottest data center markets in the country — has seen multiple municipalities pump the brakes on new developments citing water concerns. Mesa, Queen Creek, and Rio Verde have all made headlines in recent years for restricting new water connections or rejecting large developments outright.
What's changed isn't just the drought conditions — it's the political and regulatory willingness to say no to projects that would have sailed through approval a decade ago.
The data center industry grew up in an era when land and water were treated as abundant inputs. Phoenix was a data center darling because of cheap land, no state income tax, and favorable power rates — the water question was an afterthought. That calculus is being forcibly updated.
In the Pacific Northwest, the city of The Dalles, Oregon — home to a massive Google campus — has faced growing scrutiny over the billions of gallons of Columbia River water consumed annually by those facilities. Oregon state legislators have introduced bills requiring greater transparency around data center water use. The direction of travel is clear: public tolerance for opaque, high-volume water consumption by private tech infrastructure is declining.
What the Rejection Costs San Marcos
It would be dishonest to frame this as purely a win for the environment without acknowledging the economic trade-offs.
A 200-acre data center development — if it had proceeded — would have represented significant construction employment, permanent operations jobs, and a substantial addition to the local tax base. Data centers are notoriously light on permanent headcount (a large facility might employ 50-200 full-time staff), but they're extraordinarily heavy on property tax revenue. In some Texas counties, data center facilities have become among the largest single contributors to local tax rolls.
San Marcos is home to Texas State University and has a diversified local economy, but it's not immune to the lure of that kind of tax revenue. The city essentially chose long-term resource security over short-term fiscal upside — a decision that will look either wise or costly depending on how the next decade of drought plays out.
Local businesses that would have supplied construction materials, catering, facilities services, and other ancillary services won't see that contract pipeline materialize. That's a real cost, even if it's the right call.
Building Data Centers in Water-Stressed Regions — Is There a Path Forward?
The San Marcos rejection doesn't mean data centers can't be built in drought-prone areas. It means they can't be built the same way.
A handful of approaches are gaining traction among operators serious about deploying in constrained water markets:
Air-side economization — using outside ambient air for cooling rather than evaporative water cooling — can dramatically reduce or eliminate water consumption. Facilities in cooler climates have used this for years. The challenge in Texas is that summer wet-bulb temperatures make purely air-side cooling economically punishing during peak heat months.
Closed-loop liquid cooling — direct-to-chip or immersion cooling systems — recirculate water in a closed loop with dramatically lower consumption than open-cycle evaporative systems. Several hyperscalers are deploying this at scale, and it's not niche technology anymore. Microsoft, Google, and a growing list of colocation providers are treating liquid cooling as a standard engineering decision, not an exception.
Recycled municipal water is another lever. Some facilities are negotiating agreements to use reclaimed wastewater for cooling rather than drawing from potable water supplies. Austin has explored this approach with industrial users. It's not a silver bullet — reclaimed water still consumes a resource — but it relieves pressure on drinking water supplies.
The developer that comes back to San Marcos — or any similar market — with a credible, independently verified water consumption plan will have a fundamentally different conversation with city council than the one that just happened.
What Infrastructure Professionals Should Take From This
The San Marcos case is a case study, not an anomaly. Here's what it actually implies for people doing site selection, permitting, and development work in this space:
Water risk is now a first-order variable. It doesn't belong in the environmental appendix of a feasibility study. It belongs on page one, alongside power and fiber. If your site selection process isn't incorporating aquifer stress data, municipal water availability projections, and drought trend analysis, it's incomplete.
Community opposition is water-literate now. Local residents and councils in drought-stressed regions have lived through mandatory water restrictions. They know what a dry reservoir looks like. Developers who show up with boilerplate "we're committed to sustainability" language are going to get eaten alive in public hearings. You need specifics — gallons per day, comparison benchmarks, mitigation commitments with teeth.
Regulatory risk is accelerating. What's a voluntary best practice today becomes a permit condition tomorrow and a compliance requirement the year after. Operators who get ahead of water efficiency requirements — rather than fighting them — will have a competitive advantage in siting.
The deeper lesson is this: the infrastructure industry built its site selection models during a period of relative climatic stability. That period is over. The projects that get built in the next decade will be the ones that treat environmental constraints not as obstacles to work around, but as engineering parameters to design within.
San Marcos didn't kill data center development. It killed a particular kind of data center development — the kind that assumes water is free and unlimited. The industry has the technology to do better. The question is whether the economics will catch up fast enough to matter.
[INTERNAL LINK: water consumption trends]
[INTERNAL LINK: data center development challenges]
[INTERNAL LINK: sustainable cooling solutions]
Ready to explore how your infrastructure projects can adapt to these new realities? Visit InfraSale Marketplace to discover innovative solutions and resources.