AI drinking up our water? | How datacenters may cause water scarcity #fypp #ai
The short version: A short explainer video on AI's water footprint is a good entry point, but the real accountability lives in county permits and utility filings near you.
Our Take
Credit to creator Nilesh Savale for putting a number on something most people never think about: the fresh water it takes to cool the servers behind every chatbot query and AI image. The comparison to fast fashion is a smart hook — it reframes water use as a everyday-choices issue, which is fair, but it can also let the real decision-makers off the hook. Nobody voted for a hyperscale campus to draw millions of gallons a day from a municipal system already stressed by drought; that decision usually got made in a closed-door economic development meeting, with tax abatements attached.
The video is right that location matters — water-stressed regions in the US are exactly where a lot of this buildout is landing, often because land and power are cheap and permitting is fast, not because the water math makes sense. What's missing from most of these explainers is the paper trail: which utility approved the water allocation, what the facility disclosed (or didn't) to local zoning boards, and whether ratepayers are quietly subsidizing the cooling costs through their own bills. That's the layer GridWatch tracks.
If this video got you curious about what's actually being built near you, check our facility map to see the footprint in your state, and if you want to push back on a project before the permits are final, our take-action page has the tools to find the meeting and show up.
This is GridWatch the USA’s original commentary. The video above is the work of Nilesh Savale, published on YouTube — full credit to the creator.