Since 2022, at least ten venture-backed indoor farms have gone bankrupt, been liquidated, or wound down across the US and Europe. One of them raised more than $940 million — more than almost any indoor-farming venture in history — and built a showcase facility that was supposed to prove vertical farming could work at commercial scale. It proved the opposite.
The easy conclusion is that indoor farming does not work. That conclusion is wrong, and the survivors are proving it wrong right now, operating at real margins.
The harder, more useful conclusion is that this was not a failure of the concept. It was the failure of a playbook: build the technology, raise enormous sums, then go looking for a market — and run the farm on hindsight the entire way.
Naming the Real Problem
Every operator in a controlled environment is making million-dollar decisions on lagging information.
Consider what that means in a sealed room. Outdoors, a problem announces itself slowly, and there are weeks of weather and biology in which to react. Indoors, the environment engineered for total control turns against the grower quickly when something drifts — and the drift is usually discovered once it is already visible. By then, the decision that would have cost a spot treatment costs a room.
AGEYE calls this growing by hindsight: managing a facility built for total control while flying blind on the one variable that actually determines the outcome — what is about to go wrong.
It is quietly the most expensive input in the entire operation. Not the lights. Not the labor. Not the rent. The hindsight.
The Cost, in Numbers
Growing by hindsight is not a metaphor. It carries a price, and the research puts hard figures on it:
- A disease caught after it is visible costs roughly 3 to 5 times more to treat than the same disease caught 7 to 14 days earlier, because a targeted spot treatment has become a facility-wide intervention.
- Timing is nearly everything a sealed room takes away. The same infection that costs about 5% of yield when caught at the right moment can cost upward of 90% when caught late.
- The clock is unforgiving. When climate control slips in a sealed environment, condensation and pathogen conditions build within 24 to 48 hours, and total crop loss can arrive within 72 hours to 7 days.
Now stack that against how most facilities still run: dashboards that report what already happened, schedulers that assume nothing goes wrong, and a grower’s experienced eye that — however good — can only catch a problem once it is visible. That is not a criticism of growers. It is a criticism of the tools the industry has handed them.
Why Another Dashboard Will Not Fix It
Here is the uncomfortable part, and it applies to AGEYE’s own sector.
Controlled-environment software was built on a simple promise: growing indoors generates more data than any human can manage, so the vendor sells a subscription to manage it. Climate dashboards. Irrigation schedulers. Crop-monitoring tools. Each solves one slice of the problem, charges a recurring fee, and leaves the operator to stitch the systems together.
That arrangement is now under real pressure. The marginal cost of building a mid-tier precision-ag software feature — field mapping, scheduling, monitoring — is approaching zero as AI tooling collapses months of development into weeks. Dashboards are becoming a commodity. Anyone can build one.
Which means the dashboard was never the defensible thing. The intelligence is — the layer that does not merely show what happened, but learns from the enormous volume of data a sealed environment generates and reports what is coming while there is still time to act on it.
The Shift
So the shift every serious operator and investor should be watching is not “more software.” It is a change in what the software is for.
- From explaining the past to seeing the next problem first.
- From every facility as an isolated black box to one intelligence layer across all of them.
- From growing by hindsight to growing by foresight.
That is the difference between a farm that reacts and a farm that anticipates — and increasingly, it is the difference between a farm that survives and one that becomes another cautionary tale.
The survivors of the last three years did not win because they had better lights or more capital. Several of them had far less. They won because they ran disciplined operations that learned from their own data instead of being surprised by it.
The next phase of this industry belongs to the operators who stop paying the hindsight tax — and to the technology built to end it.
For anyone building, operating, or investing in controlled-environment agriculture, the question worth sitting with is not “what is my yield?” It is “how late am I finding out when something is about to cost me one?”
Because in a sealed room, by the time the problem is visible, it is already being paid for.
AGEYE builds the intelligence layer for controlled environment agriculture — the system that helps growers see the next problem before it costs them a harvest. The full argument for the category, and the numbers behind it, is set out in Growing by foresight.