A cognition-core farm never arrives without warning. It just arrives without warning to you — the surveyors show up only once the land’s already spoken for.

The pattern is almost embarrassingly simple once you know what you’re looking at. Core farms are thirsty, power-hungry things. They need a fusion relay station close enough that the transmission conduits don’t bleed half their output as heat, and they need cooling water nearby that nobody’s currently litigating over. So the corps go looking for exactly that: cheap parcels sitting in the shadow of an existing reactor, near a river or aquifer, zoned loosely enough that a Colony Administration won’t ask too many questions until the concrete’s already cured.

Mega-corps have built proprietary siting models for this — encrypted, licensed, sold back to each other as a “predictive advantage.” What’s strange is that none of it is actually secret. It’s zoning maps, grid telemetry, and public reactor output logs. Anyone can read it. So I did.

I spent six weeks this cycle pulling public transmission-conduit records from three Core Systems settlements, cross-referencing them against land parcels that changed hands through shell purchasers in the last standard year, and plotting both against existing reactor footprints. You don’t need a cognition-core of your own to run this — a decent terminal and patience will do it. I did the first pass on a terminal older than some of my readers, because the point was never speed. The point was that it works without a license.

What the map shows, settlement after settlement, is a quiet land grab roughly eighteen months before groundbreaking. Water rights get bundled into unrelated agricultural sales. A substation gets “upgraded” for reasons nobody in the district requested. Property assessments creep up before there’s anything visible to assess. By the time the fabrication crews land, the district’s already negotiating from a position of catching up.

The methodology is below the fold, in full: the parcel-overlay method, the conduit-load threshold I used (anything pulling above 40% of a relay’s rated capacity is a candidate site), and the raw telemetry sources — all public, all free. Replicate it. Break it. Tell me where it’s wrong.

I’m also running an open seminar next cycle — no fee, no login wall, just a transmission anyone can tune into — walking through how to read your own district’s grid this way. Call it Core Ecology if you need a name for it. I’d rather you just call it looking.