The growth driver is AI compute demand, which is pulling capital and capacity into Texas at a pace that has outrun the state’s planning infrastructure

Decision Lens

Texas data centers currently consume less than 1% of the state’s water. A May 2026 UT Austin white paper projects that share could reach 3% to 9% by 2040 — a range wide enough to signal genuine uncertainty, but narrow enough to confirm material risk. The critical complication: this estimate captures both on-site cooling consumption and the indirect water embedded in generating the electricity that powers these facilities. For operators with large Texas footprints, water is no longer a facilities management variable. It is becoming an infrastructure planning constraint with regulatory and reputational dimensions that sit squarely in this function’s lane.

90-Second Brief

Now, a UT Austin white paper published in May 2026 projects Texas data centers could account for 3% to 9% of the state’s total water use by 2040, up from under 1% today. The estimate includes both direct cooling water and the water intensity of the state’s power generation mix, which relies heavily on natural gas, coal, and nuclear, sources that already consume roughly 5% of Texas water. Texas currently has more than 400 data centers operating or under construction. Fragmented governance across dozens of water authorities makes coordinated planning difficult and increases regulatory exposure for operators.

What’s Actually Happening

The growth driver is AI compute demand, which is pulling capital and capacity into Texas at a pace that has outrun the state’s planning infrastructure. The water exposure embedded in that growth is two-layered. The first layer is direct: evaporative and process cooling at hyperscale facilities consumes water on-site at volumes that scale with power density. The second layer is indirect and less commonly modeled: Texas’s generation mix — dominated by thermal sources including natural gas, coal, and nuclear — is itself water-intensive. Power production already accounts for approximately 5% of Texas water consumption, and data center load growth directly amplifies that draw.

What makes this operationally distinctive is the fragmentation of water governance in Texas. Responsibility is distributed across municipalities, groundwater conservation districts, river authorities, private suppliers, and multiple state agencies. There is no single counterparty for operators to engage. The COMPASS consortium at UT’s Bureau of Economic Geology has identified the absence of shared definitions and consistent metrics as a first-order problem — one that currently prevents operators, planners, and regulators from even agreeing on what counts as data center water use, let alone managing it.

Why It Matters for Global Heads of Data Center Energy?

The water-energy nexus in Texas creates a planning interdependency that cannot be siloed into a facilities team. Energy procurement decisions — particularly the sourcing mix used to power Texas campuses — directly influence the indirect water footprint that will increasingly draw regulatory attention. An operator running large volumes of power from thermal grid sources in a water-stressed Texas region is not just exposed to carbon accounting risk; it is accumulating water liability that may trigger future permitting friction, municipal opposition, or utility coordination requirements.

For heads of energy with Texas interconnection positions or active PPA negotiations in the state, the composition of the offtake matters beyond price and basis risk. Renewable power from sources with low water intensity — wind, solar PV — reduces the indirect water footprint tied to each megawatt-hour consumed. That distinction may become material in water-stressed planning regions where regulators or local authorities gain standing to challenge data center expansion on water grounds. The practical implication: water impact should enter PPA evaluation criteria and site selection criteria now, not when a regulatory framework forces it.

The Forward View

The UT Austin white paper’s policy recommendations point toward tighter disclosure requirements and coordinated planning frameworks that would pull data center operators into regional water governance conversations they have largely avoided. An integrated planning framework combining hydrologic projections, grid capacity models, land-use constraints, and permitting processes — as recommended by COMPASS — would formalize water as a co-equal variable alongside power in Texas data center development approvals.

If adopted, even partially, this changes the operator engagement model. Energy teams would need to coordinate with water authorities at the same stage as utility interconnection discussions. Site selection workflows would require hydrologic screening. Long-duration permitting risk would expand to include water availability alongside grid capacity. Operators who build proactive relationships with water planning bodies before frameworks are mandated will have a structural advantage in queue position and community approval timelines.

What We’re Uncertain About?

  • Actual cooling technology mix across Texas facilities. The 3%–9% range is wide precisely because water intensity varies dramatically by cooling approach — air-side economization, evaporative towers, and direct liquid cooling have very different water profiles. No public inventory of cooling technologies deployed across Texas’s 400+ facilities exists, and operators are not currently required to disclose this. Resolution would require either voluntary industry reporting or a state disclosure mandate.

  • How water governance fragmentation affects permitting timelines. Whether distributed authority across Texas water bodies translates into formal permitting friction — or remains an informal stakeholder risk — is not yet established. Whether the COMPASS framework gets adopted into regional water plans or state agency guidance would be the leading indicator to watch.

  • Pace of regulatory uptake. The white paper’s recommendations are advisory. Whether Texas legislative or agency action moves to formalize data center water reporting requirements within the current planning cycle is unknown. Federal interest in data center resource use could accelerate state action or remain decoupled.

  • Indirect water intensity under a changing generation mix. As Texas adds wind and solar capacity, the indirect water draw per MWh consumed by data centers will decline — but the rate depends on grid mix evolution and data center load growth simultaneously. These are not independent variables, and the interaction effect is not modeled with precision in the current white paper.

One Question to Bring to Your Team

For each active Texas site and interconnection position, do we know the split between direct cooling water consumption and the indirect water intensity embedded in our grid-sourced power — and have we mapped either figure against the water planning region’s current stress classification and projected constraints by 2035?

Sources

  • Utexas — Data Centers Are Growing in Texas, But Big Questions Remain About Water Use – UT Austin News – The University (Link)

Tags