Smart Technology, Dumb Choices: Rethinking AI’s Resource Hunger

We live in an age where artificial intelligence can predict droughts with stunning accuracy, optimize water distribution networks, and model climate scenarios decades into the future. Yet the very infrastructure that powers these miraculous capabilities is consuming water at such a staggering rate that it’s creating the exact scarcity problems AI was supposed to help us solve. The smartest technology humanity has ever created is being built on remarkably foolish environmental choices—and the consequences are already flowing from server farms to farmlands across America.

The Hidden Thirst of Intelligence

When we interact with ChatGPT, generate an image with DALL-E, or ask Alexa a question, we rarely think about the vast infrastructure humming behind our screens. But the numbers are staggering. Large data centers can consume up to 5 million gallons per day, equivalent to the water use of a town populated by 10,000 to 50,000 people.¹

To put this in perspective, Google’s data center in Council Bluffs, Iowa, consumed 1 billion gallons of water in 2024.² That’s enough to supply drinking water to roughly 30,000 people for an entire year, used instead to cool the servers that power our digital lives.

This consumption isn’t just growing—it’s accelerating. More than 160 new AI data centers have been built recently, representing a 70% increase from the prior three-year period.³ As companies like Microsoft, Google, Meta, and OpenAI race to deploy ever more powerful AI models, each iteration demands more computational power, more cooling, and ultimately, more water.

The Collision with Agricultural Needs

Data centers rank among the top-ten water-consuming commercial industries and the United Nations has predicted that by 2025, 50% of the world’s population will live in water-stressed areas.⁴ This timing couldn’t be worse, as the AI boom coincides with increasing drought conditions and water scarcity in many regions where data centers are being built.

The cruel irony is unmistakable: as we build AI systems that could potentially optimize agricultural practices, predict weather patterns, and help manage water resources more efficiently, the very infrastructure supporting these systems is competing directly with farmers for the water they need to grow our food.

The Corporate Race and Its Consequences

The competition between tech giants has transformed from a healthy rivalry into something resembling an arms race, with environmental consequences that scale exponentially with each new model. The tech giants both report surges in greenhouse gas emissions as they double-down on adding artificial intelligence to all of their products, as NPR reported about Google and Microsoft’s recent emissions increases.

Each company’s quest to achieve artificial general intelligence (AGI) first has created a feedback loop where environmental responsibility takes a backseat to competitive advantage. Microsoft partners with OpenAI, pouring billions into AI research while simultaneously pledging carbon neutrality. Google races to match GPT capabilities while its water consumption soars. Meta develops its own AI infrastructure while expanding its “Tent” data centers designed for rapid deployment.

A single AI-focused data center can use as much electricity as a small city and as much water as a large neighborhood, according to the Union of Concerned Scientists. When multiplied across hundreds of planned facilities, the scale becomes almost incomprehensible.

The Promise That Keeps Us Hoping

Yet we cannot simply dismiss AI’s transformative potential. The technology offers genuine solutions to existential challenges: climate modeling that could save millions of lives, drug discovery that might cure diseases, educational tools that could lift entire populations out of poverty, and optimization systems that could make existing industries far more efficient.

AI has already demonstrated remarkable capabilities in predicting crop yields, optimizing energy grids, and modeling complex environmental systems. The same computational power that requires massive water cooling could theoretically design more efficient desalination plants, predict drought patterns with unprecedented accuracy, or develop new materials for water conservation.

The question isn’t whether AI can benefit humanity—it’s whether we can harness its benefits without destroying the resources we need to survive.

A Path Forward: Innovation Within Limits

The solution lies not in abandoning AI development, but in fundamentally restructuring how we approach it. This requires several crucial shifts:

Efficiency Over Scale: Instead of building ever-larger models that require exponentially more resources, the industry must prioritize efficiency. Recent developments in model compression, edge computing, and specialized chips show promise for delivering AI capabilities with dramatically lower environmental footprints.

Location Intelligence: Data centers should be built where water is abundant, not where land is cheap or tax incentives are generous. The expansion of data centres across the US is prompting water utilities to reconsider how they will manage increasing water requirements.⁵

Cooling Innovation: Microsoft pledges a 95% reduction in evaporative water use by 2024, while others explore AI-optimized systems for predictive efficiency.⁶ The industry must accelerate adoption of alternative cooling methods, from liquid immersion to geothermal systems.

Regulatory Framework: Policymakers must establish water usage limits for data centers, particularly in water-stressed regions. The current approach of allowing unlimited consumption in exchange for economic development is unsustainable.

Transparent Reporting: Companies rarely disclose exactly how much water their data centers consume.⁷ Mandatory disclosure would enable communities to make informed decisions about data center development and hold companies accountable for their environmental impact.

The Choice Before Us

We stand at a crossroads where the tools that could help solve our greatest challenges might themselves become our greatest challenge. The AI revolution need not become an environmental catastrophe, but only if we act with the same urgency and innovation that drives technological development.

The farmers in Cheyenne aren’t anti-technology; they’re asking a fundamental question about priorities. In our rush toward artificial intelligence, we cannot afford to sacrifice the natural intelligence that has sustained human civilization for millennia—the wisdom to live within our means and preserve the resources that future generations will need.

The race for AI dominance will be won not by the company that builds the largest models, but by those who prove that transformative technology and environmental stewardship can coexist. The question is whether our leaders have the vision to choose that path before it’s too late.

The promise of AI remains profound, but only if we can fulfill it without drinking our planet dry.

The conversation about AI’s environmental impact is just beginning. As consumers, investors, and citizens, we all have a role in ensuring that our technological future doesn’t come at the expense of our environmental foundation.

References