In the Texas Panhandle, where the Ogallala Aquifer has long sustained cattle ranching and irrigated crops like wheat and cotton, a new kind of neighbor is arriving. Projects such as Fermi America’s Project Matador near Amarillo, a proposed 11-gigawatt AI data center campus spanning thousands of acres, promise economic transformation. Yet they also spotlight intensifying competition for scarce resources. As hyperscale facilities powering artificial intelligence proliferate, their demands for electricity and cooling water are colliding with the needs of agriculture and local communities. This tension is not unique to Texas, but the Panhandle’s arid conditions and aquifer dependence make it a vivid case study in the trade-offs ahead.
Data centers have always required power and cooling, but the AI surge has dramatically escalated both. According to the Lawrence Berkeley National Laboratory, U.S. data centers consumed about 176 terawatt-hours of electricity in 2023, roughly 4.4% of national generation. Projections for 2028 range from 325 to 580 TWh, or 6.7% to 12% of total U.S. electricity. Globally, the International Energy Agency forecasts data center demand more than doubling to around 945 TWh by 2030, with AI-optimized facilities driving much of the growth. A single hyperscale AI campus can draw as much continuous power as a mid-sized city.
It’s worth noting that in his recent State of the Union address, President Donald Trump announced a “ratepayer protection pledge” aimed at addressing rising electricity costs driven by the explosive growth of AI data centers. He stated that major technology companies have an obligation to provide for their own power needs and explicitly directed them to build plants to produce their own electricity. The goal, according to Trump, is to prevent these energy-intensive facilities from straining the existing grid or increasing utility bills for American households and businesses.
Water use follows a similar upward trajectory. Evaporative cooling systems, essential for dissipating heat from dense GPU racks, account for the bulk of direct consumption. A large facility can require up to 5 million gallons per day, equivalent to the domestic use of a town of 10,000–50,000 people. U.S. direct water consumption by data centers reached tens of billions of gallons annually by 2023, with hyperscale AI sites projected to push that figure even higher. Indirect use, tied to electricity generation, especially from thermal plants, multiplies the footprint.
In Texas, the Houston Advanced Research Center provides some of the clearest statewide estimates. Existing data centers were projected to consume about 25 billion gallons of water in 2025, split between direct cooling and power-plant demands. By 2030, depending on growth and grid mix, that could rise to 29–161 billion gallons of water use annually. AI facilities typically use roughly twice as much water per megawatt-hour as conventional facilities. In the Panhandle, where groundwater from the depleting Ogallala Aquifer serves farms and ranches, even modest additional draws raise alarms. Project Matador, for instance, has secured agreements for up to several million gallons daily from municipal sources while exploring hybrid dry-wet cooling and recycled water to mitigate impact.
These concentrated demands create direct friction. Agriculture remains the dominant water user nationwide, often accounting for 70–80% in Western states, but data centers introduce high-density, year-round industrial loads in places already facing drought. In the Texas Panhandle, residents near proposed sites have voiced concerns at public meetings about aquifer drawdown affecting irrigation and livestock. Similar stories emerge elsewhere: Meta’s facility in Georgia has faced well-water complaints from neighbors, Northern Virginia data centers consumed nearly 2 billion gallons in 2023, a +63% jump since 2019, and Arizona communities weigh new builds against groundwater limits.
Yet the story is not one-sided. Proponents highlight substantial benefits. Data centers bring construction jobs, ongoing operations employment, and billions in investment and tax revenue. Texas has attracted tens of billions in recent commitments, including Google’s Panhandle expansions and elements of the Stargate project. AI infrastructure underpins productivity gains across sectors, from precision agriculture to manufacturing and healthcare. Supporters argue that targeted incentives and innovation can align growth with sustainability. Many operators are shifting to air or hybrid cooling, wastewater reuse, and on-site renewables. Project Matador touts a mix of nuclear, solar, gas, and battery storage to deliver behind-the-meter power with reduced grid strain. Industry-wide, water-use effectiveness is improving as some facilities now achieve near-zero evaporative loss.
Critics counter that without stronger guardrails, short-term economic wins risk long-term resource depletion. In water-stressed basins, even a small percentage of state totals can overwhelm local systems during peaks. Farmland conversion to data centers often seeks large, flat parcels, reducing agricultural output in an era of food-security concerns. Rising electricity prices can squeeze farm margins, while backup generators add local air-quality burdens.
Regulatory responses are accelerating, reflecting these tensions. Starting in spring 2026, Texas regulators will require data centers and crypto facilities to report direct water use, cooling technologies, and power sources starting in spring 2026, a step toward better planning. Virginia, the nation’s data-center capital, debates tightening tax incentives and mandating local water-impact reviews. Other states consider efficiency standards, higher industrial rates, or moratoriums on new builds in stressed areas. Policymakers face the classic dilemma, foster innovation that drives GDP and competitiveness, or safeguard traditional economies and community resilience. (Source: texastribune, iea.org, eesi.org)
