ABOUT
The name is a joke. The receipts aren't.
ThirstTrap.ai is an independent, ad-supported research publication about the water, power, permits, and local infrastructure behind the AI and data-center boom. The voice is loud on purpose. The research is calm, nonpartisan, and source-linked on purpose.
Why this exists.
The public argument about AI and water is loud, viral, and mostly wrong in both directions. One side says every prompt drinks a bottle of water. The other says data centers are a rounding error next to agriculture, so nothing matters. Both claims quote real studies and both strip the boundary that made the study honest.
ThirstTrap.ai exists to put the denominator back. Every number on the site says who reported it, when, for which facilities, under which boundary, and whether it was measured or modeled. The goal is a clear, shareable answer that a resident at a public hearing, a reporter on deadline, a planner reading a docket, or an engineer defending a design can all cite without embarrassment.
How it is paid for.
Contextual display advertising in clearly labeled, fixed-height slots is the primary revenue model. Clearly disclosed direct and newsletter sponsorships may follow. Merch is the third leg. There is no paywall, no subscription, and no paid correction. Evidence, methods, tools, and corrections stay free to read and share. The advertise page lists every placement and rule.
Independence.
Advertisers and sponsors do not preview, approve, select, or suppress evidence, conclusions, rankings, corrections, or coverage decisions. Ads sit physically apart from receipts, source links, calculator results, and corrections. Advertiser categories are reviewed for conflicts with the topic, and material relationships are disclosed. If that ever changes, it will be announced on this page, dated.
The site is nonpartisan. It takes no position on whether any specific project should be approved. It takes a strong position on whether the numbers in the argument are being used honestly.
What exists today.
- 15 sourced answers, each on its own permalink page with boundary, evidence class, and verification date
- A homepage with headline receipts, a signal feed of recent filings and disclosures, and a company disclosure comparison that refuses to rank
- The Impact Lens: a calculator that shows what three published, non-comparable per-prompt figures would imply, with every caveat attached
- A public methods page with evidence classes, metric definitions, boundary rules, and a corrections log
- Machine-readable versions of all of it: sitemap, RSS feed, llms.txt, and structured data
What does not exist yet.
The handbook rule is simple: never imply coverage or capability that does not exist. So, plainly, the following are not built:
- Nationwide automated ingestion or real-time monitoring of permits and dockets
- A daily alert pipeline
- Facility-level maps or records beyond the sourced examples on the site
- Team workspaces, exports, or an API
- Subscriptions, paid tiers, or paid alerts (not planned)
Contact.
Tips, corrections, source suggestions, press, and advertising all go to thirsttrapai@pm.me. There is no form and no mailing list; email is the whole inbox. Research cutoff for launch content: .