In the race to develop cutting-edge AI infrastructure, we often celebrate technological advancement while overlooking a critical environmental cost: water consumption.
While many discussions focus on AI’s energy demands, the water required to power and cool these systems represents a growing concern, especially in water-scarce regions like Saudi Arabia, where digital expansion is accelerating rapidly.
The Unseen Thirst of Modern Computing
The water usage behind AI infrastructure is multifaceted and often invisible to consumers and businesses alike.
According to Salah Al-Kafrawi, senior consultant at EY for data and AI, “We have no universal approach to assess how much water is consumed while using or training AI,” which leaves the true scale of the problem poorly understood.
This water consumption occurs through both direct and indirect channels:
- Direct usage through cooling systems for servers
- Indirect consumption via electricity generation
- Treatment and transportation of water itself
Most concerning is that available data likely underestimates actual consumption by a factor of 10 or more. While some tech companies publish water usage estimates, most businesses remain unaware of their total water footprint.
The Water-Energy Nexus in Computing
Abdullah Al-Otaibi, Saudi Arabia’s managing director at Ecolab, frames the issue through what scientists call the “water-energy nexus,” the interconnected relationship between water and power consumption.
“Water serves a dual purpose in data centers,” Al-Otaibi explains. “It cools infrastructure directly and gets consumed indirectly when generating the electricity that powers high-performance computing.”
This creates a concerning cycle: AI requires substantial power, that power generates heat, and cooling that heat demands even more water. Data centers rely heavily on cooling towers to dissipate the intense heat produced by servers, creating another significant source of water usage.
The Saudi Context: Where Technology Meets Scarcity
In Saudi Arabia, where water is precious and AI infrastructure is rapidly expanding, managing this balance becomes increasingly urgent. Abdulelah Al-Shehri, assistant professor of chemical engineering at King Saud University, notes that most global data centers rely on potable water despite its scarcity.
The numbers are revealing:
- Saudi data centers currently use hybrid cooling systems that consume between 2.3 and 2.8 liters of water per kilowatt-hour
- The current 300 MW of operational data centers could use approximately 6.7 million cubic meters of water annually, equivalent to the needs of 160,000 Saudi households
- Capacity could more than quadruple to 1,300 MW within five years, potentially matching the water needs of 700,000 households
These figures only capture direct water usage for cooling. The indirect water cost tied to energy production remains even greater.
The Energy Source Factor
The choice of energy source significantly impacts water consumption.
Research shows that solar and wind are the most water-efficient energy sources, while biomass, often promoted as a renewable option, can consume “up to 100 times more water than natural gas,” according to Al-Shehri.
This challenges developers to consider not just the carbon footprint but also the water footprint of different energy options when powering AI infrastructure.
Promising Solutions on the Horizon
Despite these challenges, experts point to several promising approaches that could significantly reduce water usage:
1. Rethinking Temperature Standards
Traditional data centers operate at around 21 degrees Celsius, but companies like Microsoft and Google have piloted “high-temperature data centers” operating at up to 35 degrees Celsius.
This shift allows systems to rely more on air cooling than water cooling, potentially driving the direct water footprint to near zero.
2. Heat Recovery Systems
Absorption chillers can reclaim up to 40 percent of waste heat and repurpose it for cooling.
In Switzerland, company Infomaniak now uses server heat to warm 6,000 homes, demonstrating how waste heat can become a resource rather than a burden.
3. Smart Water Management
Better data and AI tools can help companies understand and reduce their water footprints. Ecolab’s audits have helped clients cut water usage by 44 percent, energy by 22 percent, and emissions by 12 percent, all while improving system reliability.
Their partnership with global data center operator Digital Realty has implemented an AI-driven system expected to reduce water use by up to 15 percent and prevent the withdrawal of 126 million gallons of potable water annually.
Expert Editorial Comment
For regions facing water scarcity while pursuing digital transformation, integrating water efficiency from the beginning is essential. “The key is starting early – smart water strategies begin before construction,” Al-Otaibi emphasizes.
By addressing these challenges now, the tech industry can ensure that the expansion of computing capability doesn’t come at the expense of increasingly precious water resources.
For Saudi Arabia and similar regions, where water is scarce but technological ambitions are vast, water-efficient design isn’t just environmentally responsible, it’s a business imperative that will determine the sustainability of digital growth for decades to come.
