AllianceBernstein: Water Risk From Data Center Boom Threatens $350B AI Build-Out

4 min read
Sara Rosner, Director of Responsible Investing Research at AllianceBernstein, co-authored the April 9, 2026 analysis warning that rising data center water demand could threaten the projected $350 billion AI infrastructure build-out.
Sara Rosner, Director of Responsible Investing Research at AllianceBernstein, co-authored the April 9, 2026 analysis warning that rising data center water demand could threaten the projected $350 billion AI infrastructure build-out.| Photo: Alliancebernstein

Water scarcity is emerging as a critical constraint on the artificial intelligence infrastructure boom, with analysts at AllianceBernstein warning that the $350 billion projected for AI-focused business investment in 2025 alone is placing unprecedented strain on freshwater supplies — a risk they say belongs on every investor's radar.

The analysis, published April 9, 2026, was authored by Sara Rosner, Director of Responsible Investing Research at AllianceBernstein; James Russo, Portfolio Manager for US Strategic Core Equities and Senior Research Analyst; and Henna Nordqvist, CFA, Corporate Credit Analyst on the Fixed Income team. The research was conducted in collaboration with the Columbia Climate School.

Data Center Expansion Drives Surging Water Demand

Global capacity for data centers — the physical infrastructure housing AI servers — is expected to rise 23% annually over each of the next five years, with the United States leading the expansion, according to the AllianceBernstein report. That build-out is projected to spike data center electricity demand by 160% in the US, contributing to a 25% increase in overall power demand through 2030, per the International Energy Agency.

But electricity is only part of the resource equation. AI's powerful data centers tend to run at high temperatures, and many require large volumes of ultrapure water for continuous cooling. The AllianceBernstein analysts note that a data center's long-term viability will likely depend as much on access to clean, ample water as on reliable power.

Data centers are not the only participants in the AI boom competing for water resources. Semiconductor manufacturers and utilities — particularly those relying on coal and nuclear power generation, which account for approximately 70% of all US freshwater withdrawals — are also intensive water users expected to require significantly more supply going forward. The report notes that most water withdrawn for power generation is returned to its source after cooling.

Water Stress Threatens AI Infrastructure Execution

Research conducted jointly by AllianceBernstein and the Columbia Climate School found that data centers, semiconductor manufacturers, and utilities combined are expected to grow demand for clean water by 33% through 2030. The report also highlights that many data centers are either situated in or planned for regions already experiencing water stress — a geographic concentration that amplifies risk.

The analysts point to real-world examples of companies already adjusting their strategies in response to water constraints. Constellation Brands abandoned a planned beer plant in Mexico, and Google scrapped plans for a data center in Chile, both citing water-related concerns. A Bloomberg study cited in the report found that approximately $70 trillion in global GDP could be directly exposed to high water stress by 2050.

"Water risk is emerging as a key constraint for these and other industries — potentially more disruptive than climate change itself," the AllianceBernstein team wrote. The firm argues that water stewardship must be integral to active investment selection, and that company engagement is essential to determining a business's water-risk exposure.

AllianceBernstein notes that many companies outside the tech industry, whether household names or unfamiliar start-ups, are participating in the AI boom in some fashion, but that progress will more readily favor innovators offering lasting solutions to water scarcity.

Investment Opportunities Across the AI Supply Chain

AllianceBernstein identifies several technology categories positioned to benefit from rising water risk across the AI infrastructure supply chain.

The firm highlights ocean-sourced cooling as a significant emerging opportunity. Though costlier than purifying water from municipal sources, desalinating seawater is described as very effective in semiconductor fabrication, using either evaporation or membrane filters to remove minerals that can harm equipment and microchips. The global desalination market is expected to exceed $50 billion by 2032. Companies cited as key enablers in this space include DuPont Water Solutions, LG Chem, and Flowserve.

Direct-to-chip liquid cooling represents another avenue, placing metal cooling plates in direct contact with processors to dissipate heat while using minimal water. Asetek is identified as one of the leading public companies in this segment, which currently includes numerous private firms.

Closed-cycle cooling — described as highly water efficient and environmentally friendly — circulates liquid coolant through systems ranging from evaporative towers to condensers. SPX Technologies is highlighted as a global pioneer in the public markets, with privately held companies such as Kelvion and Hamon Group also advancing the technology.

Active Investing in a Water-Stressed World

The AllianceBernstein report frames water risk not only as a threat to data center development, but also as a source of investable opportunity. The firm argues that with a significant share of today's market value tied to a handful of AI leaders, responsible investing requires recognizing water-risk exposure at every level of the supply chain — as well as the opportunities created by it.

The analysis was co-authored with contributions from Maxwell Lulavy, Responsible Investing Research Analyst at AllianceBernstein.

The views expressed in the AllianceBernstein report do not constitute research, investment advice, or trade recommendations and are subject to change over time. References to specific securities are for illustrative purposes only.