Source: AgriFlats, by Wheeler Kearns Architects.
Architects are increasingly asked to put a working farm inside a building. Rooftop greenhouses, ground-floor grow rooms, glass-wrapped food districts. It is one of the most exciting briefs in the built environment right now, and one of the easiest to get wrong. The difference is almost never the quality of the design. It is the order of operations. When the building gets designed first and the farm economics get reverse-engineered to fit, the farm fails. When the growing operation is modeled first and the building is shaped around it, the project has a chance. This piece is about that sequence, and why the moment to bring in agronomy, capital, and climate expertise is before the rendering, not after.
Consider AgriFlats, a proposal by Wheeler Kearns Architects for Chicago's West Side and a finalist in World Business Chicago's 2050 visioning exercise under the Horizon Lines banner. The entry clusters production greenhouses with housing to "grow growers" and seed a CEA hub in Chicago. It is the anchor of a planned Food District: about 16 acres under glass on a roughly 21-acre site. The firm projected hundreds of permanent jobs and water use at a fraction of field agriculture. This is not a naive rendering, it is one of the more disciplined urban agriculture proposals we have reviewed recently. Wheeler Kearns clearly thought hard about energy, water, and community, which is exactly why it makes such a useful case.
Still, when this concept was shared publicly, the response converged fast, and it went straight to the seams. Architects, mechanical engineers, real estate financiers, and greenhouse operators converged on the same set of questions, and none of them were about whether it was beautiful. The question that kept surfacing was blunt: does the growing operation get modeled before the building, or after? That reaction is the whole lesson.
Here is why the reaction was so consistent. A farm inside a building is not a single design problem. It is the collision point of four disciplines that rarely sit in the same room. Agronomy sets the yield. Capital sets the pro forma. Engineering sets the climate. Policy sets what you are allowed to build. Draw the envelope beautifully and you can still break every one of them.
The financiers in that thread made the sharpest version of the capital argument. In a dense market like Chicago, the binding constraint is not design. It is the opportunity cost of land. Every square foot of low-margin, single-story greenhouse is a square foot not earning the net operating income of multi-story residential or commercial space. A horizontal agricultural envelope struggles to compete with vertical real estate value. That is not a reason never to build. It is a reason the numbers have to be run before the massing is fixed.
The engineers went after the physics. A greenhouse tall enough to look dramatic in a rendering, thirty feet to the ridge, is a greenhouse bleeding the heat you are paying to conserve. Shading from adjacent housing can starve the crop of the light the system was designed around. Form drifts out of alignment with function quietly, in ways that only show up later in the energy bill and the harvest data.
And the operators reminded everyone that a farm is not a building you finish. It is an operation you run, for years, sized to a specific crop, a specific distribution model, and a specific pro forma, kept alive by a skilled team over the full lifecycle. None of that is visible in the drawing.
The pattern underneath all of these critiques is the same: the building gets designed first and the economics get reverse-engineered to fit, when it has to run the other way. Several architects quoted the line back approvingly, because they have watched it happen.
It happens for an understandable reason. Architects are trained to optimize form. The farm arrives as a program requirement, gets penciled in, and then gets value-engineered late, once the envelope is already fixed. By that point you cannot retrofit agronomy. You cannot move the glass, change the bay depth, or re-solve the climate strategy without unwinding the design. As one mechanical engineer in the field put it, when an architect is involved they often do not yet know which questions to ask, so form and function end up at odds. Equipment vendors rarely close that gap. They tend to sell what they sell.
There is an irony here that the architects in the conversation caught immediately. "Form follows function" is Louis Sullivan's principle, coined in 1896, and it is a cornerstone of modern architecture. A farm inside a building is the most literal test of that principle you will ever encounter, because the function is not human comfort. It is photosynthesis, climate control, water chemistry, and unit economics, all of which are unforgiving.
This is why our own farm design methodology starts from the same rule: form follows function. We model the growing operation first, then shape the building around what the farm actually requires: the light it needs, the height it can afford to heat, the climate strategy that pencils, the logistics of moving product out the door. Our consulting philosophy is four words: Know before you grow.
That does not reduce the building to a container for a crop. The best integrations are the ones where the farm and the building create value for each other: a rooftop greenhouse that improves the building's thermal performance, an operation that anchors community programming or education, a food hub that gives a mixed-use development a genuine reason to exist. Those secondary benefits are real, and they can justify what crop economics alone cannot. But only when they are designed in from the start, not bolted on at the end.
Method matters more than opinion here, so here is ours.
We begin with feasibility, not form. Market first: who buys the product, at what price, in what volume. Then the business model, the crop plan, and the growing system. Then a bottom-up financial model built from real data on comparable operating farms, the kind of benchmarks we track through the Global CEA Census, covering capital expenditure, operating cost, revenue by channel, payback period, and break-even yield. We build these numbers before a dollar is committed to construction, because they are the design brief.
Our planning platform, Agritecture Designer, lets a team model a farm's unit economics in a single working session and compare systems, crops, and automation levels side by side, with capital and labor estimates drawn from vetted vendor data. It is the fastest way to find out whether an idea pencils before it becomes a drawing.
Only then do we move to design: layout, system specification, and a climate strategy that hand the architect a brief that is operationally sound and investor-ready. Every engagement puts agronomists, architects, financial analysts, and engineers on the same team. And because we do not sell equipment or take referral fees, the only outcome we are optimizing for is a farm that works.
The argument for engaging cross-domain expertise early is brutally simple. The cost of the wrong sequence gets poured into the foundation. Once the envelope is set, the mistakes become structural, and structural mistakes are expensive. Bring the farm logic in at concept, before the rendering is locked, and you are not constraining your design. You are protecting it, because you are designing around real requirements instead of around a farm that cannot run.
We have watched this work. In 2021 we brought together an international cohort of architects and gave them access to Agritecture Designer plus direct consulting from our team. The concepts that came out of it, from repurposing a 200,000 square foot concrete IKEA parking lot in New Jersey into phased food production, to reprogramming Milan's abandoned Ex-Macello slaughterhouse site, were strong precisely because the architects nailed down the farm's financials, resource use, and yields before finalizing the form. They designed food into the built environment in a way an investor could take seriously.
That intersection, architecture and agriculture done with discipline, is where controlled environment agriculture genuinely belongs. It is also where most of it still fails, for want of the right sequence.
The developers and architects who win the next decade of this work will be the ones who treat the farm as a system to be modeled, not a feature to be drawn. The rendering is the easy part. The farm that is still running, and still paying for itself, in year seven is the hard part, and it is decided long before the first line is drawn.
If you are integrating agriculture into a project, bring the cross-domain read in before the design is locked. It is the cheapest insurance you will ever buy on an expensive building.