By Charly SHELTON

Artificial intelligence (AI) has become the subject of countless headlines over the past few years. AI has been praised as the next great technological revolution and criticized for everything from job displacement to misinformation. More recently, another concern has entered the conversation: water.

Stories about AI’s enormous water consumption often cite some staggering numbers, warning that the data centers powering today’s AI systems consume millions of gallons of water every day. Those figures are accurate but like many statistics they tell only part of the story.

The important factor isn’t just how much water AI uses – it’s where that water is being used and how those numbers compare with the other demands that are placed on the water supply every day.

Modern AI systems rely on massive data centers filled with thousands of computer servers. Those servers generate a huge amount of heat, requiring liquid cooling systems to keep that equipment operating safely. Many facilities use evaporative cooling, a process that consumes water as part of dispersing that heat.

Miguel Yañez-Barnuevo, in a study for The Environmental and Energy Study Institute, noted, “A medium-sized data center can consume up to roughly 110 million gallons of water per year for cooling purposes, equivalent to the annual water usage of approximately 1,000 households. Larger data centers can each ‘drink’ up to five million gallons per day.” 

That number is understandably alarming at first glance. Five million gallons is difficult to visualize and when presented without context it can sound like an unsustainable burden. But context is exactly what these headlines often lack.

According to the U.S. Environmental Protection Agency, Americans use nearly nine billion gallons of water every day watering lawns and landscapes. Outdoor irrigation accounts for roughly one-third of residential water use nationwide and in arid regions such as Southern California, where the rainfall and snow runoff is less than back east, that percentage can be even higher.

Viewed nationally, lawn irrigation uses many times more water each day than all U.S. data centers combined. That doesn’t mean concerns about AI’s water use are to be ignored – it just means the issue is more nuanced than a single statistic suggests.

Water is a local resource. A community with an abundant water supply may not even notice the arrival of a new data center. A drought-prone community, however, could feel the impact immediately. That’s why researchers caution against treating water consumption purely as a national issue.

That distinction is becoming increasingly important as technology companies race to build new AI infrastructure. The United States is seeing billions of dollars invested in new data centers to support growing demand for artificial intelligence, cloud computing and online services. Those facilities bring jobs, tax revenue and economic investment but they also increase demand for electricity and water.

For communities considering new developments, the conversation becomes one of balance rather than absolutes. Southern Californians understand this better than most. Residents have lived through years of drought restrictions, mandatory watering schedules and constant reminders that every drop matters. Most homeowners know which days they’re allowed to water their lawns. Restaurants routinely ask whether customers want water instead of serving it automatically. Conservation has become part of everyday life.

Against that backdrop, it’s reasonable to ask whether AI companies should also be expected to use water responsibly. The answer is almost certainly yes.

“[The rate of freshwater consumption] is concerning, as freshwater scarcity has become one of the most pressing challenges,” one study from researchers at UC Riverside suggested. “To respond to the global water challenges, AI can, and also must, take social responsibility and lead by example by addressing its own water footprint.”

But it is equally reasonable to compare AI’s water consumption with the many other ways society already uses water. Agriculture, landscaping, manufacturing, power generation and recreation all compete for finite resources. Artificial intelligence is becoming another item on that list – not necessarily the largest, but certainly one that deserves careful attention.

Maybe regarding water usage the most important distinction isn’t about artificial intelligence versus lawns; it’s about numbers. 

Humans are notoriously poor at understanding very large numbers. A figure measured in millions of gallons sounds enormous until it is compared with billions. And each of those is hard to fully comprehend when the largest measurable amount of liquid usually encountered is a gallon of milk or the five-gallon water jugs in the water cooler. National averages can obscure local impacts while local headlines can make national trends seem larger than they are.

As AI continues to become part of daily life, conversations about its environmental footprint should continue. But those discussions are most useful when they include perspective as well as data. Artificial intelligence does consume water. In some communities, it consumes a lot of water. Whether that’s cause for alarm depends less on the headline statistic than on where that water comes from, who else needs it and how wisely people choose to use it.