Login
Why Water Risk Is the Missing Variable in AI Infrastructure Planning
While power dominates the headlines in AI infrastructure, water is the silent arbiter of project viability. Investors and developers obsess over megawatts and grid capacity, but the reality is that cooling systems are tethered to a resource that is often less predictable and more politically charged. When water or wastewater capacity hits a ceiling, the fallout moves beyond engineering. It triggers permitting stalls, operational interruptions, and structural impairment of asset value.
Across the U.S., municipalities are no longer just providing service; they are becoming the ultimate ‘gatekeepers’ for high-volume users. For instance, Tucson now requires any new or expanding large water user expecting more than 7.4 million gallons per month to submit a conservation plan, undergo public review, and secure City Council approval before accessing Tucson Water.
Marana’s policy further states that Marana Water will not supply potable water to data centers for cooling and requires documentation of an alternate source. In Chandler, the city council unanimously rejected a proposal to rezone land for a 422,000-square-foot AI data center campus after public opposition emphasized water use, noise, and limited local benefit.
Strategically positioned between engineering and financial close, these water policies represent a major ‘blind spot’ for developers. Late-stage discovery of water limitations results in stranded capital and protracted entitlement delays. For modern investors, such water risk is now a primary underwriting variable that can dictate the viability of an entire transaction.
Why Power Is Only Half the Constraint
Power determines how much IT load can be energized, but cooling determines whether that load can operate within temperature limits on peak summer days. Cooling design also determines whether the site depends on local water, meaning the true constraint is rarely singular.
Data centers typically rely on one of two primary heat rejection approaches.
Evaporative systems, such as cooling towers, remove heat through water evaporation. This requires continuous makeup water to replace evaporative loss and generates blowdown to control mineral concentration. Blowdown becomes a wastewater stream, tying the facility to sewer capacity, discharge regulations, and pretreatment requirements.
Dry systems, such as air-cooled chillers and dry coolers, reduce direct on-site water consumption but increase electrical demand as outdoor temperatures rise, particularly during summer peaks. That shift moves the constraint toward grid capacity and power pricing during the very hours when electricity is most expensive and constrained. In both configurations, the constraint does not disappear but shifts, and each approach carries a distinct exposure profile that must be evaluated at the basin and grid level.
Inside the Water Footprint of AI Data Centers
Water exposure extends beyond the visible intake line and is often more complex than initial site reviews suggest.
In tower-based systems, make-up water demand rises as ambient temperatures increase because more heat must be rejected during peak hours. Blowdown volumes also rise, increasing steady wastewater discharge. In many jurisdictions, wastewater capacity determines viability before raw water supply does. Dissolved solids and treatment chemistry can trigger pretreatment mandates or exceed plant acceptance thresholds, creating operational bottlenecks that were not modeled at the outset.
The true water footprint of an asset is often obscured by ‘siloed’ diligence. While a facility might minimize on-site usage, it remains tethered to the water intensity of the local energy mix—a dependency that creates a hidden risk during peak demand. Because most models consider water, power, and wastewater as isolated variables, the full scale of the water-energy nexus is rarely consolidated. This leaves the project exposed to systemic failure points that only become visible late in the development cycle.
Why Water Risk Is Frequently Mispriced
The assumption that water is a stable, predictable utility is a significant blind spot in traditional underwriting. Standard diligence often stops at a letter of intent from a provider, ignoring regulatory contingencies—such as recycled water mandates or peak-heat restrictions—that govern high-intensity facilities. Failing to account for these municipal requirements leads to Capex volatility and structural delays, turning a simple utility expense into a primary threat to projected returns.
At a portfolio level, aggregated corporate reporting can obscure localized exposure. Average water intensity metrics do not reveal whether specific assets sit in basins facing physical scarcity or wastewater systems operating near capacity. Valuations that assume perpetual expansion can fail at the local level when additional allocation is unavailable, undermining long-term growth assumptions embedded in underwriting models.
From Environmental Constraint to Financial Exposure
Water risk tends to accumulate over time, moving through operations, regulation, and local politics until it becomes a real constraint on performance.
For operators, the first pressure points are often summer peaks, when supply limits tighten and water quality can swing at the exact moment cooling systems are working hardest. This dilemma then leads to emergency operational changes that pull maintenance forward, or take short outages. Ultimately, the revenue impact of those decisions is usually disproportionate to the duration of the disruption.
For developers, on the other hand, regulatory shifts can trigger midstream redesigns. A project engineered around potable water may be required to transition to reclaimed supply, adding infrastructure, storage, and treatment complexity after capital has already been committed.
Public opposition at the local level introduces political friction that stalls approvals and compounds reputational risk. Contentious infrastructure upgrades can derail project schedules and force unfavorable cost-sharing renegotiations. Collectively, these municipal factors feed into underwriting through increased delay risk, Capex volatility, and a diminished capacity for long-term expansion.
What Needs to Change in Infrastructure Planning
Water must be evaluated at the same stage as power during site screening and early design.
A simple confirmation of water availability is no longer sufficient. Basin-level allocation rules, drought contingency plans, wastewater capacity, discharge quality requirements, and embedded grid water intensity must be assessed before engineering assumptions are finalized.
Every investment memo and design review should include a transparent water balance that identifies source type, volume requirements, discharge pathways, and regulatory triggers under peak conditions. This allows engineering and underwriting teams to evaluate exposure in parallel rather than sequentially.
Water limits are now shaping asset values in a direct, measurable way. Resilience starts with expansion plans that can hold up under tighter supply caps, and with capital that funds backup sourcing options and protection against shifting rules. Financing and insurance need to move to basin-by-basin risk models, because water availability is already the deciding factor in approvals and the constraint that most reliably dictates whether an asset can keep performing over time.
# # #
About the Author
Dr. Vian Sharif is the Founder and President of NatureAlpha, an AI-first fintech platform delivering science-based environmental risk insights across nearly $3 trillion in assets under management. With 20 years of experience at the intersection of finance, technology, and sustainability, she also serves as Head of Sustainability at FNZ Group and is a global advisor on nature-aligned investing. She holds a PhD in Environmental Behavior Change and was recognized with a 2025 Fin-Earth Award for Natural Capital and Biodiversity.
The post Why Water Risk Is the Missing Variable in AI Infrastructure Planning appeared first on Data Center POST.
