US Cities Move to Curb AI Data Center Expansion
The rapid build-out of generative AI is driving demand for hyperscale data centers whose servers require substantial electricity and cooling. A United Nations University report released in June 2026 estimated that global data centers used 448 terawatt-hours of electricity last year, with generation of that power consuming about 1.2 trillion gallons of water. The findings have sharpened concerns over whether local grids and water systems can absorb further expansion.
Austin’s City Council voted unanimously on Aug. 27 to begin drafting rules covering data centers’ water and electricity use, noise and lighting, while considering a ban on hyperscale facilities. Projects seeking more than 20 megawatts of power or 85 million gallons of water annually must be flagged to the council. Binding rules are due by the end of 2026, and large projects will require council approval in the interim.
All Coverage
1 original reportsThe Backstory
The history behind this eventU.S. AI Data Center and Chip Fab Expansion Deepens Water Risks in Drought-Hit Areas
Generative AI infrastructure consumes water not only to cool data centers but also in semiconductor manufacturing and power generation. As TSMC, Meta, Nvidia and others accelerate their U.S. expansion, water allocation has emerged as a key issue for regulatory reviews in drought-hit states, industry growth and environmental sustainability.
As of July 20, 2026, the latest analysis showed that 66% of new U.S. AI data centers are being built in drought-prone areas, adding pressure on water supplies for local households and agriculture. The reports did not disclose total investment, but warned that the simultaneous expansion of chip fabs, data centers and electricity demand could prompt several states to tighten water-use regulations.
Subscribe to Mark Radar Weekly
Every Friday, the week's strongest signals in your inbox. Unsubscribe anytime.
If you search news on Google, you can set Mark Radar as a preferred source—our coverage will show up more often in your results. Set as preferred source on Google →