In the wake of a global technology boom fueled by demand for artificial intelligence, a new and unsettling environmental concern is emerging. Data centers, the vast, humming warehouses of the digital world, are consuming staggering amounts of a resource more precious than data: water. These facilities use water directly for cooling, and even more indirectly for the electricity that powers them. According to a 2024 report from the Lawrence Berkeley National Laboratory, U.S. data centers consumed an estimated 17 billion gallons of water directly in 2023, with projections that this figure could more than double by 2028. But despite this monumental consumption, there is a distinct lack of transparency from the companies building them, obscuring the true environmental impact from public view and hindering efforts to manage these valuable resources.
The Unseen Resource
The water consumption of data centers can be categorized into two distinct types: direct and indirect. Direct water use refers to the cooling water that is pumped through pipes to absorb the immense heat generated by the computer equipment. Indirect water use, which can be significantly greater, is the water required to produce the electricity that powers the entire facility. This indirect consumption increases dramatically when the source is fossil fuels, which require vast amounts of water for cooling in their own power generation processes. The Lawrence Berkeley Lab estimated that in 2023, this indirect water use was a stunning 211 billion gallons in the U.S. alone.
This is not a theoretical problem, but a very real one, especially in regions with high water demand. Researchers based on the shores of Lake Michigan note that tech companies are eyeing the Great Lakes region as a prime location for new data centers. The region offers a relatively cool climate and an abundance of water, making it a natural fit for these “hot and thirsty” facilities. However, the Great Lakes are a critical binational resource that serves as the drinking water source for over 40 million people and supports a US$6 trillion regional economy. The demand from new data centers can compete with these existing uses and potentially deplete local groundwater aquifers, raising a serious conflict between technological expansion and human necessity.
Behind the Screens: How Cooling Systems Work
To understand the scale of water consumption, one must first understand how data centers use water to cool their servers and routers, which generate a tremendous amount of heat. Two of the most common methods are evaporative cooling and closed-loop cooling. In an evaporative cooling system, cold water is pumped through pipes and absorbs heat from the servers, turning into steam that is vented out of the facility. This method, while effective, requires a constant and massive supply of new water, as the water that evaporates is consumed and not returned to its source.
A closed-loop cooling system, on the other hand, is designed to be more water-efficient. The water absorbs the heat, but instead of being vented as steam, it is cooled down by air-cooled chillers and then recirculated back into the system. While this method uses far less water, it requires a lot more energy to run the chillers. The authors note that the specific numbers showing these differences are often not publicly available, but the aggregate numbers from major companies provide a glimpse into the scale. For example, Google reported consuming over 6 billion gallons of water in 2023 to cool its data centers, while Meta’s data centers accounted for 95% of its total water consumption, a figure of 776 million gallons.
The Transparency Problem
Despite their immense water footprint, technology companies rarely provide a clear and consistent picture of their water consumption. The authors’ analysis of corporate sustainability reports, government documents, and public records found a significant lack of transparency. While some companies release sustainability reports, they are often voluntary and vary wildly in what they choose to disclose, making it difficult to combine or compare data across the industry. For instance, Amazon releases annual reports but does not disclose its water use at all. Microsoft provides water demand data for its overall operations but doesn’t break down the numbers for its data centers, while Meta provides a company-wide aggregate figure without a more granular breakdown.
Only a few companies, like Google, provide individual figures for each data center, but even these disclosures are incomplete. The authors note that these reports “do not consistently include the indirect water consumption from their electricity use,” which a recent estimate showed was 12 times greater than direct use. Without a clear and consistent reporting method for both direct and indirect water use, the public and government officials are left with an incomplete and misleading picture of the industry’s environmental impact.
A Closer Look at the Tech Giants
A deeper dive into the available data for Google and Meta reveals just how central data centers are to these companies’ water use. In 2023, data centers comprised 95% of Google’s total global water consumption, a staggering 6.1 billion gallons. Meta’s numbers were similar in proportion, with its data centers consuming 776 million gallons of water globally, accounting for the same 95% proportion. The high-volume consumption is driven by facilities like Google’s data center in Council Bluffs, Iowa, which consumed a billion gallons of water in 2024 alone. This is an almost unimaginable amount, enough to supply all of Iowa’s residential water for five days.
This massive consumption also highlights the stark difference in technology use. For example, a Google data center in Pflugerville, Texas, which is air-cooled, consumed only 10,000 gallons of water in the same year—about what a single Texas home would use in two months. However, the authors emphasize that even this comparison is incomplete because Google’s disclosures do not pair water consumption data with the size of the centers, the specific technology used, or the critical indirect water consumption from their power sources. This lack of context means that even the most detailed reports offer only a “partial view, with the big picture obscured.”
The Obscured Big Picture
The data center industry is expanding at a rapid pace, a trend that will only accelerate with society’s growing reliance on AI. As new data centers are proposed and built, they will require massive amounts of water, especially in a world grappling with climate change and increasing water scarcity. But without a consistent and transparent way to track water consumption, society is flying blind. The public and government officials lack the complete information needed to make informed decisions about where to locate these facilities, how to regulate them, and how to plan for sustainability.
The absence of a standardized reporting framework means that each new data center is a step into the unknown for the communities that host them. These hot and thirsty buildings are competing for a finite resource with human communities, agriculture, and ecosystems. To ensure a sustainable future, the industry must move beyond voluntary disclosures and partial reporting to a system of mandatory and comprehensive transparency. Only then can we make sure that the digital future we are building does not come at the cost of one of our most precious natural resources.
