As the Middle East’s AI ambitions reach a fever pitch, its scarcest resource—water—is consumed in staggering volumes. While much has been said about AI’s voracious appetite for energy, few realize that just 20–50 queries to ChatGPT can consume as much as half a liter of water. At a larger scale, this number increases dramatically.
As more companies adopt AI-driven strategies and technological advancements, and as data centers that support AI continue to be built across the region, the impact on water resources is becoming increasingly severe.
The recent announcement of building the world’s largest AI data center, the 5 gigawatt AI data center in Abu Dhabi, and Saudi Arabia’s rapid digital expansion through Humain, an AI company established under the Public Investment Fund to drive the kingdom’s AI strategy and other sovereign AI efforts, underscores the urgency of addressing AI’s water footprint.
AI is not the only industry that relies on data centers for cooling systems. Other digital technologies, such as cloud storage platforms, also utilize these data centers. Water usage represents a broader concern as the world increasingly depends on digital technology.
“In many industries, the connection between water and energy is often overlooked, especially in data center cooling,” says Stefan Umiastowski, Senior Vice President and Market Head for India, Middle East, and Africa at Ecolab. “As AI drives high-performance computing, these systems require massive amounts of energy and water to maintain reliable operations.”
Historically, energy has dominated the sustainability conversation. Often treated as a local issue, water hasn’t received the same scrutiny. But that’s quickly changing. “The scale of AI infrastructure is expanding rapidly, but so is the pressure on freshwater reserves,” Umiastowski adds. “In regions like the GCC, proactive resource planning is more important than ever.”
REIMAGINING COOLING IN THE DESERT
The core issue is cooling. AI servers run hot, far hotter than traditional cloud workloads. According to Maurizio Frizziero, Director of Cooling Innovation at Schneider Electric, power density per rack has exploded, jumping from 12kW in 2024 to over 130kW today, with projections reaching 600kW per rack by 2027.
“Air cooling simply can’t keep up,” Frizziero explains. “Beyond 30kW per rack, it becomes inefficient. That’s where liquid cooling comes in, especially direct-to-chip solutions, which deliver tighter control, better instrumentation, and up to 300x less water use.”
Adapting these technologies to Gulf climates is not straightforward. “Ambient temperatures in the region limit traditional free cooling,” Frizziero says. “However, hybrid systems that combine liquid cooling with adiabatic techniques offer scalable, water-conscious solutions. And if paired with smart reuse of heat, such as for agriculture or energy, the cooling footprint can become an asset.”
BUILDING SMARTER
Khazna, the UAE’s largest data center operator, is leading this cooling transition. The company considers water stewardship integral to, not incidental to, its infrastructure strategy.
“The UAE’s national agenda reinforces that sustainability isn’t optional, it’s foundational,” says Tinboat Arslanouk, Chief Business Officer – International at Khazna. “We align with net-zero goals by designing facilities that cut freshwater use and prioritize treated sewage effluent and adiabatic free cooling.”
Khazna operates on-site water treatment plants, creating closed-loop systems that reuse water efficiently while meeting technical cooling standards.
“We’re building AI-ready facilities, like our upcoming Ajman data center, using direct liquid cooling,” Arslanouk says.“These choices aren’t just efficient; they reflect a broader shift toward engineering optionality—cooling solutions that are climate-appropriate, scalable, and sustainable.”
Khazna was recently certified to build to Nvidia’s Blackwell specifications, supporting ultra-dense AI compute with water-saving liquid cooling — a move Arslanouk calls a “strategic milestone.”
SMART WATER METRICS
Technology alone won’t solve the water problem. Visibility and real-time data will. “Smart water management in AI data centers means taking a lifecycle approach,” says Umiastowski. “From design to operations, [companies must] focus on circular water use, advanced digital monitoring, and predictive analytics.”
Ecolab’s tech is used in hyperscale environments to deliver actionable insights, optimize cooling, and reduce water consumption. For instance, its partnership with Digital Realty aims to save 126 million gallons of potable water annually across 35 facilities.
“Ultimately, it’s about decoupling performance from water usage,” Umiastowski says. “With the right tools, GCC data centers can support AI growth without depleting water resources.”
FUTURE IS LIQUID AND ACCOUNTABLE
Governments across the region are taking proactive steps to address and reduce the risk of water scarcity. From Dubai’s solar-powered Hassyan desalination project to Saudi Arabia’s new Global Water Organization, the region is embedding water stewardship into national infrastructure plans.
“The groundwork is there,” says Umiastowski. “We’re seeing encouraging signs that water sustainability is actively factored into national AI strategies.”
Frizziero agrees, highlighting the importance of regulation and collective action. “Governments can push the industry by setting minimum efficiency standards, encouraging the use of low-GWP refrigerants, and supporting transparency in water reporting,” he says. “We welcome this evolution, it allows AI infrastructure to grow without fear of jeopardizing water security.”
However, the path forward is through ecosystem alignment. Arslanouk says water stewardship should be a collective standard, not a competitive edge. “The only way to scale responsibly in this region is by working hand-in-hand with hyperscalers, regulators, and solution providers to create a shared playbook for sustainable growth.”
AI’s infrastructure is energy-hungry, water-thirsty, and increasingly under environmental scrutiny. “Sustainability isn’t a trade-off,” Umiastowski says. “It’s a driver of long-term growth. And with the right vision and partnerships, economic expansion and environmental responsibility can and must move forward together.”
Ultimately, the true test of leadership in the AI age won’t just be about compute capacity or model performance — it will be about how wisely we manage the region’s precious resources.
