In a world increasingly run by data, the physical infrastructure powering our digital lives is expanding at an explosive rate. But this growth comes with a hidden cost: an immense and growing thirst for water. We’re sitting down with Dominic Jainy, a leading expert on technology infrastructure, to unpack a startling new report on the collision course between the data center boom and global water security. We’ll explore the staggering water footprint of a single facility, the high-stakes risks in hotspots like the U.S. and China, and the innovative solutions and policy shifts needed to build a more sustainable digital future.
A single 1MW data center can use the annual drinking water of over 400 people. Could you walk us through the step-by-step process of how a new facility’s water needs are assessed and share an anecdote illustrating the tangible impact on a local community’s resources?
It’s a statistic that really stops you in your tracks, isn’t it? Over 25 million liters of water per year for a relatively small facility. The truth is, the assessment process is often far too simplistic. Ideally, it should begin with a deep analysis of the local watershed, understanding seasonal fluctuations, and mapping out competing demands from agriculture, industry, and the local population. But in reality, for many projects, the process is transactional: can we secure a water permit for X million liters? Yes or no. It doesn’t always account for long-term climate projections or the cumulative impact of multiple facilities in one region. While I can’t share specific project details, you can easily picture the impact. Imagine a community in a drought-prone area that has been promoting water conservation for years. Then, a single new facility moves in and its cooling systems start consuming the drinking water equivalent of a whole new neighborhood. It creates a direct and palpable tension between economic growth and the community’s most essential resource.
The U.S. and China show a major disconnect between massive data center growth plans and high water insecurity. What are the top on-the-ground challenges operators face in these regions, and what are the first three practical steps a company should take before breaking ground there?
The disconnect is alarming. Both the U.S. and China have water security scores over 50 out of 85, which signals significant insecurity, yet their data center markets are set to explode. On the ground, the primary challenge for operators is navigating a landscape of immense uncertainty. In the U.S., water oversight is highly decentralized, so you’re dealing with a patchwork of local regulations that can change. You’re also facing the physical reality of climate change; a region that seems water-rich today might face severe drought in five years, threatening your cooling operations. The first three steps for any company should be strategic, not just logistical. First, conduct a genuine, long-term water resilience assessment, not just a permitting check. Look 20 years into the future. Second, engage with all local and regional authorities from day one. Don’t just talk to the economic development board; talk to the water utility, environmental agencies, and community groups to build a shared plan. Third, design for resilience from the very beginning. Make using non-potable water and advanced, low-water cooling systems the default plan, not a backup you might consider later.
You mentioned solutions like immersion cooling and using non-potable water. Could you break down the process for implementing one of these solutions in an existing facility and share some key metrics on the water savings and potential operational trade-offs, such as increased energy use?
Let’s take retrofitting a facility to use non-potable water, like treated sewage effluent. The process begins with sourcing; you have to identify a reliable local supply and then negotiate access with the municipality. Next comes the engineering challenge: you need to build on-site infrastructure to further treat that water to prevent corrosion and buildup in your cooling systems. This involves filtration, chemical treatment, and new plumbing to connect to your cooling towers. The water savings can be massive, potentially reducing your reliance on precious drinking water by over 90% for cooling. However, this is a perfect example of the critical trade-off the report highlights. All that extra pumping and on-site water treatment requires energy. You might dramatically improve your Water Usage Effectiveness, but your Power Usage Effectiveness could worsen. You’ve solved one environmental problem but potentially exacerbated another, which is why a holistic view is absolutely essential.
The report contrasts France’s stronger water governance with the decentralized approach in the U.S. What specific policies from France could be adapted for a U.S. state, and what would be the step-by-step process for a local regulator to implement them effectively?
France’s success, reflected in its lower water insecurity score of 37, stems from a more cohesive and robust governance framework. The U.S. could learn a lot from that. A state looking to get ahead of this issue could adapt France’s approach by first centralizing its strategy. The first step for a regulator would be to implement mandatory, location-specific reporting of water use for all data centers. You simply cannot manage a resource you aren’t measuring accurately. The second step would be to establish clear, statewide standards for cooling efficiency, pushing operators toward best practices rather than the bare minimum. Finally, the third and most crucial step is to actively streamline the process for using non-potable water sources. Instead of it being a bureaucratic nightmare, regulators should make it the easiest, most attractive path for new developments. This combination of measurement, standards, and incentives creates a framework that guides the industry toward responsible growth.
The report warns against increasing electricity demand to save water. Can you provide a concrete example of this trade-off and walk us through the holistic decision-making framework an operator should use to balance these competing environmental metrics when designing a new facility?
A classic example is choosing between a traditional evaporative cooling system and a closed-loop air-cooled system. Evaporative cooling is like sweating; it uses the evaporation of water to dissipate heat very efficiently from an energy perspective. But it’s incredibly water-intensive. A closed-loop system, on the other hand, uses no water but relies on massive fans to force air over heat exchangers. This requires a significant amount of electricity to run those fans, driving up energy consumption. An operator needs a “Total Resource Impact” framework. This goes beyond just looking at the monthly utility bills for water and power. You have to model the carbon footprint of that extra electricity, especially if the local grid is fossil-fuel-heavy. You have to assess the long-term risk and cost of water in that specific region, factoring in climate change projections. It’s a multi-variable equation where you weigh kilowatt-hours, cubic meters of water, carbon emissions, and regional resource resilience to find the genuinely most sustainable solution for that specific location.
What is your forecast for the data center industry’s water usage over the next decade, especially with the explosion of AI? Will technological innovation and policy changes be enough to offset the resource demands of this rapid growth, or are we heading toward a crisis?
My forecast is that we are at a critical inflection point. The raw demand for water from the data center industry is set to skyrocket, driven directly by the heat-intensive demands of AI and high-density computing. The growth is simply too fast to ignore. Technological innovations are promising—immersion cooling, for example, is incredibly efficient—but technology alone will not be a silver bullet. Its adoption takes time and significant capital investment. I firmly believe that without rapid, coordinated action from policymakers and regulators, we are heading toward a crisis in many water-stressed regions. We’ll see more conflicts between industrial growth and community water needs. The technology to mitigate this exists, but it will not be deployed at the scale and speed required unless it is guided by intelligent, forward-looking policy. The future isn’t set in stone, but the current trajectory is a collision course.