Two very different stories get told about data centers and water. The activist version: AI is draining aquifers and stealing water from communities. The industry version: it's a rounding error, and we'll be water-positive by 2030. Both framings are using real data. Both are misleading in specific and documentable ways. This episode walks through the honest, calibrated picture.

AI-generated (NotebookLM) audio overview. Source: HexLocal in-house research — "Data Centers and Water Consumption: A Calibrated Research Brief" (Dr. Priya Nair, June 12, 2026). Primary sources include Lawrence Berkeley National Laboratory, USGS, the IEA, peer-reviewed environmental journals, and named-source investigative reporting.

- The three frames you have to hold at once: nationally small, locally sharp, and larger than disclosed once you count the power plants generating the electricity

- How cooling actually works: why evaporative systems consume water, why switching to air cooling doesn't fix the problem — it moves it off-site to the grid

- WUE vs. PUE — and the disclosure gap that lets a company report a low water score while quietly pushing most of its water footprint to a different zip code

- Water rights in the American West: why operators secure senior rights, what that means for communities already using them, and how Arizona's 1980 Groundwater Management Act is now biting the data centers, not the farmers

- The "bottle of water per ChatGPT query" claim: why the widely-cited 2023 figure is outdated by a factor of 25 to 150x, and why the vendor numbers that replaced it aren't the full story either

- Where activists distort, where industry distorts, and what remains genuinely unresolved in the science

- The trajectory: IEA projects global data center water consumption roughly doubling by 2030 — but the leading edge of AI cooling is trading on-site evaporation for off-site electricity, not eliminating the footprint

Sources: Lawrence Berkeley National Laboratory 2024 U.S. Data Center Energy Usage Report; USGS water-use data; peer-reviewed research by Siddik, Shehabi and Marston (Environmental Research Letters 2021), Mytton (npj Clean Water 2021), and Li, Yang, Islam and Ren (Making AI Less Thirsty, 2023); IEA Energy and AI report (2025); The Green Grid WUE white paper; sustainability disclosures from Google, Microsoft, and Amazon Web Services; reporting from The Oregonian, Fortune, Bloomberg, Brookings, the Lincoln Institute, EESI, and Data Center Dynamics.