Lutnick claims that data centers, known for water consumption, ‘don’t use water’

When Commerce Secretary Howard Lutnick told CNBC that data centers "don't use water," he made a statement that flies in the face of well-documented industry reality. The cooling systems at large data centers—particularly those using evaporative cooling towers—consume enormous volumes of water. Google alone reported using over 5 billion gallons of water across its data centers in 2022, and Microsoft's usage has similarly drawn scrutiny from environmental regulators and local communities alike. Yet there's a nuance worth examining. Lutnick's comments may have been referencing newer liquid-cooling technologies that are increasingly being deployed in next-generation AI data centers. These systems use closed-loop cooling that recirculates water rather than evaporating it, which does dramatically reduce net water consumption. Some industry advocates argue that as these technologies become standard, the water footprint of data centers will shrink considerably. The problem is that his framing was absolute—"don't use water"—when the reality is far more complex. Current facilities do use water, and many communities from Oregon to Chile have pushed back against data center development precisely because of water scarcity concerns. The juxtaposition of massive AI expansion with drought-stressed regions is a genuine policy tension that deserves honest discussion, not dismissive soundbites. What's at stake here is credibility on both sides. When government officials make sweeping claims that ignore current industry practices, it undermines public trust in regulatory oversight. Conversely, when environmental advocates overstate the water impact of every facility regardless of design, they lose credibility too. The responsible path forward involves transparent reporting of water usage metrics, honest acknowledgment of regional disparities, and a commitment to accelerating water-efficient cooling technologies—not rhetorical shortcuts that obscure the real trade-offs. As AI infrastructure grows, the public deserves accurate information about its environmental footprint, because communities will be living with these consequences for decades.