A commenter on a Lake Mead water level video states that the content creator should look into [AI] Datacenters and the impact they're having on water levels and the communities they're built in.
With the rise in the media attacking AI Datacenters over their water use, and the fact that I've heard this story many times before, with different villains in the role, I felt like taking my own deep dive into the actual history of the Colorado River water shortages.
What I found was every major drought in the western United States seems to produce a newly fashionable villain in the media. The target changes from decade to decade, but the underlying problems remain remarkably consistent: governments approving more development than local water systems can reliably support, promising more water than rivers can provide, subsidizing water-intensive industries, and refusing to seriously confront outdated water rights and long-term over-allocation.
To be clear, many of the groups blamed over the years genuinely used large amounts of water, and some deserved criticism. The scapegoating came from presenting each one as though it were the primary cause of the crisis, rather than one user operating inside a government-controlled water system that had already been stretched far beyond its sustainable limits.
From the late 1800s into the early twentieth century, western development was guided by the belief that water reaching wetlands, desert basins, or the ocean without first being diverted for human use was being “wasted.”
I'm actually serious. Water flowing naturally through the environment was often viewed as a resource that had failed to be put to productive use.
That philosophy helped drive the massive era of dam construction, river diversion, aqueducts, and federally subsidized irrigation across the western United States. It also encouraged governments to promise communities and farms more water than many rivers could reliably provide.
The Colorado River Compact is one of the best-known examples. It was negotiated using a relatively short period of river-flow records that happened to reflect unusually wet conditions. That made the river appear more generous than its longer history would support, and helped create a legal system that allocated more water usage than the river could consistently deliver.
In the 1930s, our first major villain entered the crosshairs. The Dust Bowl farmer.
Farmers had removed enormous areas of native grassland, exposing loose soil to severe erosion when drought and high winds arrived. But that expansion did not happen in a vacuum.
High wheat prices and European demand during the First World War encouraged farmers to plough up huge areas of the Great Plains. Mechanized farming equipment made it possible to cultivate land on a scale that had never previously been practical. When wheat prices later collapsed during the Great Depression, many farmers ploughed even more land in an attempt to produce enough volume to remain financially solvent.
The decisions made by farmers may have contributed to the disaster, but those decisions had been encouraged by years of settlement policy, agricultural economics, easy credit, and the widespread belief that the water supply in the semiarid plains could support permanent intensive farming.
From the 1940s through the 1960s, the argument increasingly became farmers versus expanding cities.
Farmers did not want to surrender the water rights they had already been promised, while growing urban populations questioned why they should be expected to conserve water while agriculture continued irrigating land in naturally arid regions.
The deeper problem was that governments continued approving and encouraging both agricultural and urban expansion, while treating new dams, aqueducts, canals, and groundwater pumping as substitutes for confronting the actual limits of the water supply. They, quite literally, sold water that didn't actually exist.
By the late 1970s, particularly during California’s severe 1976–1977 drought, public attention shifted heavily toward the individual homeowner.
Green lawns, swimming pools, decorative fountains, sprinklers, long showers, and people washing their cars became the most visible examples of water waste.
Household conservation was not pointless. Outdoor residential use can represent a large portion of municipal demand, and reducing it can provide meaningful relief during a drought.
The problem was that the public discussion became increasingly focused on individual behavior while much larger questions involving water rights, development approvals, agricultural allocation, and long-term planning received far less attention.
Whatever the government intention behind those campaigns, the effect was to make ordinary citizens feel as though their lawns and showers were the sole reason the region had run short of water.
During the drought that began in 1987 and continued into the early 1990s, public attention turned increasingly back toward agriculture.
News coverage highlighted flood irrigation, groundwater pumping, water-intensive crops, and farming in naturally arid valleys. These were legitimate concerns. Agriculture was, and remains, the largest consumptive water user across much of the western United States.
However, those issues were often discussed as though farmers had independently designed the entire system. Their choices had been shaped for decades by senior water rights, federal irrigation projects, subsidized deliveries, agricultural lending, crop markets, and government policy.
That doesn't remove responsibility from farmers, particularly large corporate operations, but it does mean the problem continued to be the result of government mishandling rather than simply the result of "farmers wasting water".
In the 1990s, another villain entered the conversation: environmentalists. One of the most famous targets was the effort to protect the endangered Delta smelt in California’s Sacramento–San Joaquin Delta.
Politicians and agricultural interests argued that water was being “dumped into the ocean” to save a tiny fish while farms and communities were being denied water.
What those arguments usually omitted was that the Delta smelt was an indicator of a badly damaged estuary. Reduced freshwater outflow and massive water exports were important contributors to its decline, alongside invasive species, habitat destruction, pollution, altered food webs, and rising water temperatures.
The fish did not create the water shortage. It became the politically convenient symbol of environmental limits being placed on a water system that had already been heavily altered and overcommitted.
Then came the 2000s.
As Lake Mead declined, Las Vegas became an irresistible visual target. The headlines focused on casino fountains, swimming pools, golf courses, artificial lakes, suburban lawns, and people “growing grass in the desert.”
Some of that criticism was warranted. Building lush landscapes in an extremely dry region obviously increases water demand. That said, by that time Southern Nevada had already begun implementing some of the most aggressive urban water-conservation measures in the United States.
Golf courses were placed under strict water budgets. Large areas of ornamental turf were removed. New water features were restricted. Indoor water was treated and returned to Lake Mead for reuse. Homeowners and businesses increasingly replaced lawns with desert landscaping and drought-tolerant plants.
The giant fountains that became symbols of waste were also recirculating systems. They still lost water through evaporation, wind, splash, and maintenance, but they were not constantly replacing their entire volume with fresh water as was suggested.
Meanwhile, agriculture continued consuming the overwhelming majority of Colorado River water. It was another case of the most visually obvious and culturally irritating users received the attention, while the much larger government structure of agricultural allocation, population growth, water rights, and desert development remained harder to fit into a headline.
By the late 2000s and early 2010s, attention shifted toward fracking. A single well could use millions of gallons of water, making the industry an easy target during droughts. While fracking could absolutely create serious local problems, especially in small or already water-stressed communities, it still represented only a small portion of total national water use. Once again, a real local concern was presented as though it explained a much larger regional crisis.
During California’s 2012–2016 drought, the list grew even longer. Almond farmers were attacked because of the amount of water required to grow each almond. Bottled-water companies were condemned for extracting and selling water during a drought. Celebrities were shamed for maintaining green lawns and swimming pools, while ordinary homeowners were told to shorten their showers and allow their yards to die.
None of those concerns was entirely imaginary but the problem was still the same one that had existed for decades: governments had allocated more water than rivers and aquifers could sustainably provide, encouraged enormous agricultural and urban expansion, and allowed groundwater to be pumped faster than nature could replace it.
Then cryptocurrency mining arrived. Crypto was primarily criticized for its enormous electricity use, but the water required to cool mining facilities and generate that electricity became part of the argument as well. Like fracking, its impact could be significant in a particular community, however it remained relatively small compared with agriculture across the wider region.
Semiconductor factories were next. Chip manufacturing requires large volumes of extremely pure water, and plans to build enormous fabrication plants in places such as Arizona understandably raised concerns. Many of those facilities recycle a significant portion of their process water, but their arrival still added another major industrial user to a water system that was already under stress.
And now the fashionable villain is AI datacenters. The technology has changed, but the argument has not.
A visible new industry arrives, journalists report an alarming estimate of how much water it might use, politicians and the public point toward it as the explanation for shortages, and the century of government-approved over-allocation behind the crisis quietly disappears from the discussion.
That does not mean datacenters should receive unlimited water, or that communities should approve them without examining their local impact. Older evaporative cooling systems can consume substantial amounts of water, and even an efficient facility can strain a small municipal system if it is built in the wrong location.
However, many newer datacenters are being designed with closed-loop liquid cooling, dry cooling, treated wastewater, or other systems that sharply reduce freshwater consumption. They are not all using the same technology, and quoting the worst examples as though they represent every modern datacenter is misleading.
More importantly, AI did not create the western water crisis. Datacenters did not negotiate the Colorado River Compact during an unusually wet period. They did not encourage a century of agricultural and urban expansion in arid regions. They did not establish senior water rights that often reward historical entitlement rather than efficiency. They did not allow aquifers to be pumped for decades without any serious accounting of how quickly they could recharge.
AI datacenters are simply the latest highly visible user to enter a water system that was already badly overcommitted. They deserve scrutiny, however they do not deserve to be used as a convenient explanation for a crisis that began more than a century before they existed.
