Are A.I. Data Centers Really Ruining Our Water Supply?
The short version: Hard Fork's chat with Andrew Masley on AI water use is worth a watch, but national averages can hide real local strain on stressed grids and aquifers.
Our Take
Credit to Kevin Roose and Casey Newton for bringing Andrew Masley of Effective Altruism DC on to talk through the water math behind AI data centers. It's the kind of nuanced, numbers-first conversation this debate needs, and we're glad it's reaching a big audience.
Here's our caveat: national or per-query averages can be technically true and still miss the point. A data center's water and power draw isn't spread evenly across the country — it lands on one county, one aquifer, one utility ratepayer base, one city council that signed a tax abatement. A modest-sounding number citywide can be a very big deal for a small water system in Georgia or a drought-stressed basin in Arizona. The question communities should be asking isn't "is this small compared to agriculture nationally," it's "what did our local officials agree to, and what happens to our bills and our water table when three more facilities get approved next year."
That's the gap we try to fill. Check our facility map to see what's been permitted near you, and visit take action for tools to push your local officials for real answers before the next deal gets signed.
This is GridWatch the USA’s original commentary. The video above is the work of Hard Fork, published on YouTube — full credit to the creator.