A new analysis has drawn attention to an uncomfortable tension at the heart of the artificial intelligence boom, finding that the majority of new AI data centers planned across the United States are slated to be built on land that has recently experienced drought. According to reporting based on an examination of planned facilities, roughly two-thirds of 809 data centers in the pipeline are set for areas classified as drought-stricken over the past year, raising pointed questions about the sustainability of placing water-intensive infrastructure in regions already contending with water scarcity. The finding highlights a growing collision between the rapid expansion of AI computing and the environmental constraints of the places where that computing is increasingly being built.
The concern stems from the substantial water requirements of modern data centers, which typically depend on large volumes of water for cooling the powerful and heat-generating computer systems they house. As the computational demands of artificial intelligence have surged, so too have the cooling needs of the facilities that support it, making water an essential input for their operation. When such facilities are concentrated in regions that have suffered drought, the result is a direct competition for a scarce and vital resource, pitting the needs of data centers against those of agriculture, households, and ecosystems that also depend on limited water supplies.
The scale of the issue is captured in the specific figures underlying the analysis. Of the 809 planned data centers examined, some 517 are set to be built in areas that the National Oceanic and Atmospheric Administration’s drought monitoring system classified as drought-stricken in the past year. This concentration of new facilities in dry regions is not necessarily the product of deliberate disregard for water concerns, but rather reflects the various factors, including land availability, energy access, tax incentives, and proximity to infrastructure, that influence where data centers are sited. Nonetheless, the cumulative effect is a significant expansion of water-hungry infrastructure into precisely the areas least able to spare the water.
Public sentiment on the matter appears notably wary, reflecting growing awareness of the trade-offs involved. Polling cited in connection with the analysis indicates that roughly seventy percent of Americans would prefer not to live next to a data center, a striking figure that suggests these facilities, despite the economic activity and tax revenue they can bring, are often regarded as undesirable neighbors. This public unease, combined with concerns about water and other environmental effects, has begun to translate into political and regulatory attention, with some states considering new restrictions on data center development.
The situation reflects a broader and increasingly visible challenge associated with the AI boom, namely the considerable physical resources that the technology requires. While artificial intelligence is often discussed in terms of its software and capabilities, the systems that power it depend on extensive physical infrastructure with real and substantial demands for energy, water, and land. As the industry continues to expand, these material requirements are coming into sharper focus, and the question of how to reconcile the growth of AI computing with environmental sustainability is emerging as one of the more significant issues facing the sector and the communities that host its facilities.
Addressing these concerns will likely require a combination of approaches, from more careful siting decisions that take water availability into account, to investment in cooling technologies that reduce or eliminate the need for large volumes of water, to greater transparency about the resource consumption of individual facilities. Some operators have begun to explore alternative cooling methods and to commit to water stewardship, though the effectiveness and adoption of such measures vary considerably. The challenge is compounded by the speed of the current buildout, which can outpace the development of the policies and practices needed to manage its effects responsibly.
Ultimately, the finding that most new AI data centers are being built on drought-prone land serves as a reminder that the AI era carries real environmental costs that extend well beyond the digital realm. As society weighs the benefits of artificial intelligence against its impacts, the question of where and how to build the infrastructure that supports it will demand thoughtful attention from companies, regulators, and communities alike. Balancing the undeniable economic and technological opportunities presented by AI against the imperative to manage finite natural resources responsibly is a challenge that will only grow more pressing as the technology, and the infrastructure it requires, continues to expand.
Your First 10 AI Skills
10 practical AI skills, copy-paste prompts and a 7-day plan to start using AI with confidence.
Download the guide →