The Hidden Costs of AI: Data Centers Strain U.S. Water and Power Resources

A sprawling rectangular building on the outskirts of Cheyenne, Wyoming, is one of many data centers constructed or approved across the United States to support the growing demands of artificial intelligence (AI). Painted in a nondescript color, the facility houses enormous heat exchangers to keep its powerful computing hardware cool. These data centers are critical to the digital economy, enabling everything from streaming video to complex AI models. Tech giants like Amazon, Microsoft, Meta, and Google’s parent company, Alphabet, are leading the charge, collectively planning to spend at least $320 billion this year on infrastructure and equipment, primarily for AI. According to Business Insider, there are now 1,240 data centers in America—nearly quadruple the number in 2010. Many of these are "hyperscalers," large-scale facilities designed to handle the substantial computational needs of AI. The potential benefits of AI are immense. Leaders in the tech industry envision a future where AI can replace human roles in healthcare, predict disasters with high accuracy, develop groundbreaking scientific theories, and even offer companionship. Cortney Thompson, the CIO and co-founder of Lunavi, which owns a data center in Cheyenne, emphasizes the importance of having the resources to innovate and create. However, the rapid expansion of these facilities comes with significant environmental and social costs. One major concern is water usage. About 40% of US data centers are located in areas with high or extremely high water scarcity, as identified by the World Resources Institute. For instance, Maricopa County in Arizona, an area marked for extreme water stress, is home to 48 data centers. Despite knowing their impact, data center operators are often only required to track water use if mandated by local regulations, leading to a lack of transparency. In 2018, US data centers consumed 34 billion gallons of water annually, a figure that has likely increased with the doubling of capacity in just seven years. Another critical issue is electricity consumption. US data centers are projected to consume more electricity than Poland did in 2023, with demand expected to triple over the next three years. This surge will necessitate keeping older, less efficient power plants operational, exacerbating pollution and public health concerns. Business Insider estimates that the pollution from these power plants could result in annual health costs ranging from $5.7 billion to $9.2 billion. To manage these challenges, tech companies are investing in green energy infrastructure. For example, near the Cheyenne data center, a massive solar farm is being developed alongside existing wind turbines. However, these efforts face skepticism. Critics argue that the environmental and social costs could outweigh the benefits if not properly managed. Additionally, local governments are providing substantial tax breaks to attract these data centers, hoping to revitalize their economies with high-tech jobs. Business Insider found that in some regions, these incentives can amount to $1 million or more per long-term job. While tech leaders maintain that they are responding to societal demand and that the investments will prove valuable, the immediate impacts include inflated local electricity prices and strained natural resources. The AI race is reshaping the landscape of the digital economy and regulatory frameworks. Virginia, home to 329 data centers, handles up to a third of global internet traffic and consumes nearly a quarter of the state's electricity. Despite local oversight, the actual resource consumption of these facilities has remained largely secret until recently. Business Insider’s detailed analysis, based on air permits and other data, sheds light on the true ecological footprint of these data centers, highlighting the urgent need for sustainable practices. While technological advancements may eventually mitigate environmental impacts, the current trajectory raises questions about the balance between progress and resource depletion. If the AI revolution delivers its promised benefits, the ledger might eventually balance. If not, the costs in terms of depleted water reserves and increased pollution may leave a lasting negative legacy. Industry insiders and experts acknowledge the challenges but remain optimistic. They assert that investments in renewable energy and more efficient cooling technologies could significantly reduce the environmental impact. However, the transparency and accountability of these efforts remain crucial. Tech companies are seen as key players in a speculative future, and their actions will shape not just technology but entire communities and ecosystems. As the race for AI dominance continues, the stakes for sustainability and responsible development are higher than ever.