Most controlled-environment agriculture inherits its layout from the greenhouse: fixed benches, shared irrigation, one big atmosphere. It photographs beautifully. It also means every plant in the room shares fate with every other plant in the room. One pythium outbreak in the nutrient loop, one thrips introduction, one failed dehumidifier over a long weekend — and you write off the room.
When your crop is a commodity lettuce, that is an insurance conversation. When your crop is the feedstock for an assayed pharmaceutical-grade product with published certificates of analysis, it is an existential one. So we changed the unit of cultivation.
The unit of failure should be the pallet, never the room.
The pallet as the unit of everything
In our system, plants grow on standard-footprint pallets — self-contained racks with their own nutrient reservoir, their own sensor package, and their own quarantine status. Pallets move through the building on ordinary material-handling equipment, from germination to grow-out to harvest, the way work-in-progress moves through any factory.
Isolation is the point. A contamination event is detected at the pallet, quarantined at the pallet, and destroyed at the pallet. The blast radius of the worst day is one pallet: a known, bounded, insurable quantity. The unit of failure should be the pallet, never the room.
What it does to the numbers
Bounded failure changes the economics more than any yield improvement we have found. Loss events stop being tail risks and become a small, predictable scrap rate — under two percent of pallets in our pilot data — which is what lets us sell an assayed 500 u/g product at a wholesale price the industry assumed was impossible.
It also changes the science. Because each pallet is its own experiment — own reservoir, own logs — every production run doubles as a controlled trial. Cultivar screens and stress protocols that used to need a dedicated research bay now ride along on ordinary production, one pallet at a time.