Crops
Cereals won't work in a vertical farm. Why falling costs won't change that
“It’s no good now, but once costs come down, will we be able to grow cereals in a factory?” — plenty of people, handed cereal factory-growing in a food security context, are carrying this question before they even reach the search box. Most people figure that if you just wait for technology to advance, an answer will eventually arrive. But look at it through today’s numbers — current prices, current technology — and the call on whether to enter is already in. It’s a question you don’t have to defer.
The three axes of cereals, where everything runs the opposite of leafy greens
Let me unpack what it means to grow cereals in a vertical farm. Factory growing of leafy greens and herbs is now standard. So it’s tempting to think, “Well then, wheat and rice too — surely those will work in a factory eventually, once the technology advances.” When food security comes up, this question often comes up on the reader’s side as well. The catch is that word “eventually”: it lets you keep deferring the call forever.
The trigger is usually that one big wave — food security. If leafy greens work, surely we can grow our own staple foods in a factory too. You also hear it framed as resilience against disasters and emergencies. But there’s a snag. Leafy greens fit the factory because they have “a high unit price, fast turnover, and short stature.” Cereals run the opposite of all three. Can you really talk about them as an extension of the same line?
Let me lay out the three axes against today’s numbers. On unit price, wheat and rice are an order of magnitude lower than leafy greens. They’re the textbook commodity, and on the ground they sit in a different league entirely from leafy greens. On growing period, leafy greens take weeks while cereals take months: wheat takes about 6 months from seeding to harvest, soybeans 3–4 months. Plant height reaches over 1m for wheat and 60–70cm for soybeans. The edible part is only the grain, and to support that grain you have to keep lighting a tall plant. You’d expect to gain area efficiency by going vertical, but the height makes it hard to add more tiers. Unit price, growing period, and area efficiency all point the wrong way. So it isn’t a matter of “eventually, the technology.” Lay these three axes against today’s numbers and whether it fits is already plain to see.
Even lettuce, the flagship leafy green, gives about 11x the yield of open-field farming on the same area under hydroponics — but at the cost of about 82x the energy, by one estimate (Ref. 1). It’s a structure where the more yield you stack up, the more electricity comes with it, and lighting accounts for much of that electricity (Ref. 2). And that’s leafy greens. For cereals, whose unit price is a tier or two lower, that same electricity structure lands directly on the books as dead weight.
Fixing just one axis doesn’t move the whole
It’s tall, so you can’t stack tiers, and that costs you area efficiency. The growing period is long, so you can’t make money on turnover. And on top of that the unit price is low. These three axes aren’t independent. They all compound, multiplicatively. Improve just one with technology and the others drag, so the total doesn’t move. Put another way, you could also say that if the structure changes on even one axis, things shift all at once.
This “not independent” is the crux. Even though they compound, push it far enough and the effect concentrates on the cheapest axis: unit price. For instance, you’re tempted to think, “if we could just make the lighting more efficient.” But in a vertical farm electricity eats most of the operating cost — lighting most of all — and the headroom for improving LED efficiency is already small. There’s even an estimate that growing wheat indoors costs an order of magnitude more than outdoors. In other words, the margin behind “if efficiency just improved a bit more” is thin to begin with. So a scenario where fixing one axis moves the whole isn’t much to count on. And “change one axis and it all shifts” is true in principle, but what can change is the premise rather than the technology. It comes closer to swapping out the premise — recasting the cereal itself as a different, short-statured crop.
There are numbers backing this. At a Swedish vertical farm, the electricity alone to grow wheat ran over USD 40,000 per ton. That’s roughly 100x the international price of wheat, by one estimate (Ref. 3). Another estimate has more than half the operating cost of an indoor wheat facility going to lighting electricity, with a cost-to-revenue ratio of about 46 to 1 — a level that’s off by orders of magnitude from penciling out under the current structure (Ref. 4). And the LED efficiency you’re banking on has entered the flattening side of the S-curve, where it’s hard to expect it to keep getting exponentially cheaper from here, as has also been pointed out (Ref. 3). The thin margin behind “if efficiency just improved a bit more” is this overlap.
That said, the same estimate also says something else. If renewable energy makes electricity cheaper, and automation and indoor-oriented breeding advance, that 46 to 1 could narrow to about 6 to 1 in the future (Ref. 4). Even 6 to 1 still doesn’t pencil out commercially, but it’s not “a wall that absolutely won’t move” — rather, “the path to shaving off one order of magnitude is on the technology side.” If it opens, it opens from here, not from an extension of mass production.
”It’ll come down once it scales” — the answer is already in
After the talk of fixing one axis with technology comes the talk of scale. “It’s still early days, so as more operators come in and mass production advances, costs will fall.” It’s a common counterargument. You’ve heard it, haven’t you? But in countries where a fair number of vertical farms have actually started up, did that “costs fall with mass production” happen? What you hear more is the other story: deficits dragging on, and subsidies not changing things.
Japan is a country where a fair number of vertical farms have started up. A considerable cumulative sum has been poured into subsidies too. And still, profits haven’t improved all at once. A 2017 industry-magazine report had roughly 75% of vertical farms in deficit at the time (Ref. 6). But that figure is old and can’t be carried over wholesale to the present. The latest public field survey (FY2025) has non-deficit operations (profitable plus break-even) above 60% overall, with deficits down to roughly the low 30% range. And the year-to-year swing is large, so you can’t quite call it a clean improving trend either (Ref. 9). More important still is the difference by facility type: deficits reaching around half are concentrated mainly in PFAL, while Greenhouse and hybrid types run around 70% non-deficit (Ref. 9). So you can’t lump it together as “X% are in deficit, so it’s no good.” What bears on the cereal question isn’t the absolute value of that deficit rate itself but the fact that even in a country with a decent sample size, mass production alone didn’t improve the bottom line by orders of magnitude.
The claim that mass production brings costs down also turns out, by one estimate, to rest on weak economies-of-scale leverage. The scale elasticity of construction cost is −0.17. That is, scaling up 100x cuts unit construction cost by only about 55%, by one estimate (Ref. 7). Half remains; it isn’t a matter of order-of-magnitude leverage. And what this elasticity acts on is construction cost — operating cost (electricity and labor) is outside the scope of this figure. So using “once it scales” as a cost-reducing lever for the whole, operating cost included, misreads the source.
And when it comes to cereals specifically, scale isn’t even the point to begin with. The same estimate states flatly that cereals like wheat are “out of the question.” But the reason is neither construction cost nor scale — it’s the single point that, at current prices, present-day cereal production itself generates no surplus (Ref. 7). It’s not “no good because construction is expensive,” nor “no good because scale is insufficient.” The unit price is low, so even when you grow and sell, the balance doesn’t come out positive. The cereal math, pushed far enough, comes down to this one point: unit price.
Still, to gauge how true “wait and it’ll come down” really is, let me draw one comparison. Even lettuce, which now looks like the flagship paying crop, was — by the same estimate — at a level that generated no profit, just like cereals, up until around 2017. Then LED improvements, indoor-oriented breeding, and accumulated cultivation know-how sent yield soaring, and it moved over to the paying side (Ref. 7). So there is precedent for a reversal. But it happened not “because mass production grew the sample size” but “because breeding, electricity, and technique swapped out the premise.” Whether the same happens for cereals is unknown, but if it does, that too will come from the technology side, not an extension of mass production.
The high-value-add exception puts cereals on the outside
Within cereals, isn’t there at least a partial exception? Have you ever wondered? It’s the often-heard line that “it works if it’s high-value-add.” Even if cereals as a whole have all three axes pointing the wrong way and won’t move, there’s branded rice even among rice, and high-priced varieties even among beans. With items whose unit price sits a tier higher and that fetch a price even in small quantities, you’re tempted to think the picture might look different even for a cereal.
But as you look through the paying exceptions, what stands out are things close to functional nutrients or pharmaceutical ingredients, whose unit price is in a different league. A world of tens or hundreds of times the unit price of an ordinary crop. This is, if anything, the polar opposite of cereals. Cereals are a commodity — the flagship of the lowest-unit-price side. So the condition “it works if it’s high-value-add,” rather than opening an exception, actually confirms that cereals sit at the very outermost edge of the region that pays. And one more thing. Even if you did aim a factory at such a high-unit-price, small-quantity item, that’s already a different matter from the original question of “mass-producing staple foods to feed ourselves.” Even if it can hold up as an exception, that exception doesn’t bear on the original theme of food security or staple self-sufficiency. It’s not that I want to deny every exception — it’s that the place it works is different.
There’s a concrete example. Tomatoes engineered to accumulate the sweet protein miraculin had lines bred specifically to aim at mass production in a vertical farm, with production reported in a closed system (Ref. 8). But this is close to pharmaceuticals and functional nutrients — ultra-high unit price, the polar opposite of a staple commodity. In break-even estimates too, short-cycle lettuce crosses the paying line at a few tens of square meters, while on the same facility assumptions the break-even point jumps sharply once the crop changes (Ref. 7). That cereals sit at the very outermost edge of the paying region can be read off these individual estimates as well.
One note: in environments where stable supply takes priority over cost, the premise of the discussion changes. Where open-field farming can’t hold — outer space, polar regions — the vertical farm’s strength of “controlled production” comes into its own. The very research lineage of indoor wheat originally came from life-support systems for the Moon and Mars (Ref. 4). Where a yardstick other than commercial profitability stands up, cereal factory growing does take on meaning.
Concluding the entry call on today’s numbers
You’re at a board meeting and get handed, “Look into cereal factory growing from a food security angle.” Picture that scene. Instead of writing “awaiting technological progress” and deferring it, what can you check right now with the numbers in hand? And if you do pass on cereals, where should the resources you’d have spent on that study rightly go?
What you can check in hand is unexpectedly simple. Take a crop already turning a profit in-house — say the unit-price table for leafy greens — and add one row for the cereal under review. Unit price, growing period, revenue per area. Line up these three axes side by side in the same table. Then it becomes visible, right there, whether the gap is the kind that narrows as technology gets cheaper, or an order-of-magnitude gap — unit price included — that won’t structurally disappear. If all three axes point the wrong way at once and the unit price is lower by orders of magnitude, that’s a current-price, current-technology problem, so you may conclude on today’s numbers to pass. No need to wait for “eventually.” When conditions like “once technology advances” or “once subsidies come” come up internally, question on the spot whether that “technology” refers to an extension of mass production or to swapping out the premise — breeding, electricity, automation. If the former, ask whether it isn’t an excuse for not looking at today’s numbers. This bit of effort serves as the check.
The place to channel resources is the region that pays. In terms of the three axes, leafy greens — high unit price, fast turnover, short stature — line up best. So put leafy greens at the center. Fruiting vegetables tend to get relegated to second tier in PFAL because they take up space relative to their unit price, but tomatoes and strawberries are commercially viable in Greenhouse and hybrid greenhouses, so the picture differs by facility type. The high-value-add niche I set aside in the previous chapter as “doesn’t bear on staple self-sufficiency” is, as a commercial call, actually a target worth chasing. It won’t replace staple food, but it’s on the side that gets chosen because it pays. These two don’t contradict. Leaning your study time toward that kind of crop portfolio is the right move. One thing to separate out: don’t use food security, emergency preparedness, or subsidies as a substitute reason for something penciling out. If you keep cereals for social significance, keep them in a separate frame from the commercial call, as a separate-layer decision. Not mixing them is what matters.
Finally, let me add some backing to the point that this is not a matter of “eventually” but of present-day structure. The vertical farm’s promises of land sparing and urban self-sufficiency have been held up repeatedly — the tower greenhouses of the 1960s, the indoor-farm projects of the 1980s — and gone unfulfilled each time. So today’s cost wall is, rather than a temporary technological lag, something structural that current prices and technology create (Ref. 3). And the vertical farms that are commercially viable are still concentrated in leafy greens, herbs, and berries, which make up only a few percent of the world’s calorie supply (Ref. 5). The picture of cereals sitting outside the set of “crops chosen because they pay” is there at current prices and current technology. If it’s to be overturned, that comes not from an extension of mass production but from the technology side — breeding, electricity, automation. That’s the way to frame it.
If you’re a vertical farm already running established crops like leafy greens, there’s plenty of room left to raise profitability depending on your on-the-ground know-how.
172 hints for raising a vertical farm’s profitability
[Free] 13 templates for use in a vertical farm’s farm operations management