The Cloud Is Not Weightless

2026-02-27 · Source: AI on Medium · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure, Environmental Science & Earth Systems · Depth: Intermediate, long

Summary

Microsoft's data centers in West Des Moines, Iowa, consumed 11.5 million gallons of water from local rivers to cool GPU servers training GPT-4, prompting the local water utility to demand significant reductions in peak water usage for future projects. This incident highlights the substantial and often overlooked water footprint of AI and cloud computing, which are typically perceived as "ethereal." Data centers primarily use evaporative cooling, consuming vast amounts of water that do not return to the local supply. A 2024 report estimated U.S. data centers consumed 17 billion gallons directly in 2023, with projections to double or quadruple by 2028, plus an additional 211 billion gallons indirectly for electricity generation. Major tech companies like Microsoft and Google reported significant increases in global water consumption, with experts attributing much of this growth to AI workloads. The industry's "water positive" pledges are criticized for their fungibility, as water replenished globally does not offset local depletion.

Key takeaway

For CTOs and VPs of Engineering evaluating AI infrastructure, recognize that the "cloud" has a significant, often localized, water footprint. Your teams should demand transparency from cloud providers on AI-specific water consumption and prioritize solutions that genuinely reduce local water stress, rather than relying solely on broad "water positive" pledges. Ignoring these environmental costs risks exposing your organization to regulatory challenges and community backlash, especially in water-stressed regions.

Key insights

AI's substantial and often hidden water footprint is a structural feature of the AI economy, accelerating climate stress.

Principles

• Every computation generates heat requiring cooling.
• Evaporative cooling consumes water, which does not return.
• Optimizing for carbon and water can be contradictory.

In practice

• Quantify both direct and indirect water consumption.
• Scrutinize "water positive" claims for local impact.
• Assess data center siting for water scarcity risks.

Topics

• AI Water Footprint
• Data Center Cooling
• Environmental Justice
• Climate Change Feedback
• Corporate Sustainability

Best for: CTO, VP of Engineering/Data, Executive, AI Ethicist, Policy Maker, Director of AI/ML

Related on AIssential

• Can you explain to me the hatred of AI and its water usage? — AI's water usage controversy centers on data center cooling demands in water-stressed regions, impacting local communities.
• AI laws overlook environmental damage – here’s what needs to change — AI's substantial environmental footprint is largely unaddressed by current global regulatory efforts.
• When it comes to total water use, AI data centers are a drop in the bucket — AI data center water use is globally minor but locally significant, prompting corporate stewardship.
• Your AI habit is wasting precious resources. Here’s how to use it responsibly — Unnecessary AI use, particularly for simple tasks, incurs significant environmental costs, necessitating a shift towards responsible and efficient consumption.
• UNU-INWEH: By 2030, global data-centre electricity use could nearly double to 945 TWh, with water impacts comparable to the basic annual domestic needs of 1.3 billion people in Sub-Saharan Africa. — AI's environmental footprint extends beyond carbon, encompassing vast electricity, water, land, and e-waste, driven largely by inference.
• Big Tech Will Not Save Us From the Climate Crisis — Big Tech's AI-driven data center expansion threatens climate goals and local environments through immense energy and water consumption.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI on Medium.