Agriculture needs another revolution
Summary
Vertical farming has the potential to unlock multiplicative yield gains per area of land and catalyze development of new technologies (precision farming, rapid genetic engineering, robotics based automation, etc.). Making this transition will exponentially benefit humanity (and other denizens of earth) in multiple ways.
However, the current focus of indoor / vertical farming is fresh produce. Here, I make the argument that in order to reduce reduce agricultural land use and obtain maximal benefit of vertical farming, we need to primarily focus on cereals, pulses and oil crops, and not just fresh produce.
Background
There are several important reasons to reduce agricultural use of habitable land use:
Agriculture use is still driving deforestation across the world.[1][2]
Preventing further wildlife biodiversity decline, reverting back agricultural land to natural habitats and making earth equitable for the other denizens of earth[3]
Reforestation still remains an effective, viable option of scaling up carbon removal and preventing large offset climate change.[4][5]
Currently, about 50% of the habitable land on Earth is used for agriculture.[6]About 33% of the total agricultural land is used for growing crops (crops for food, feed, biofuels), with the rest 67% used as grazing land for livestock. For this post, I will only focus on the land use for growing crops.[7]
Proposal: Scale Yields Multiplicatively by Growing Food Vertically
In essence, agriculture is massively spread out over large regions of land. Because of this, inefficiencies creep into the system making it harder to:
Maintain consistent yields throughout. Although this averages out globally, the variabilities (critically, losses) are often borne by farmers. This makes agriculture a very uncertain sector, which will be made even more uncertain due to the effects of climate change.
Genetically engineer crops for higher yield and viable seeds. Soil, water and climate conditions vary drastically across the large area of land use making it harder to genetically engineer crops to provide consistently high yield everywhere.
Roll out precision agriculture technologies. Embedding sensors to monitor key parameters such as water and nutrient sufficiency while detecting (and possibly eliminating) pests scales with land area.
Deploy automation and robotics. Robotics requires a constrained environment, at least currently. Agriculture spread over a large area, generally with animal activity, makes it hard to build robots to help with monitoring and harvesting.
Efficiently transport agricultural yield. Losses can occur over multiple stages (processing, storage, transport, etc.), with typical losses being around 10-20% but sometimes going as high as 40%.[8]
A common solution to all this is to figure out how to grow plants and trees, i.e. any flora of agricultural interest, in indoor environments of large multi-stored buildings. This automatically provides the following benefits:
Crop yield scales by the number of floors in the building, automatically producing multiplicative yields and can drastically reduce land use. In contrast, the global green revolution unlocked a 2.5x yield increase over the last five decades[9]
Precision agriculture and robotics becomes easier by designing buildings with monitoring and automation in mind. Climate controlled buildings increase reliability under the effects of climate change.
Spatially concentrated high yields allow self-sufficiency over counties, states and countries, making transportation and supply-chain management easier and reducing post-yield losses.
This is a necessity for humans to become a space-faring civilization ;)
However, I make a couple of key assumptions:
Using multi-storied buildings is cost-effective in the long term, even through they have much higher upfront costs. My rationale is that the advantages of monitoring, climate-control, yield optimization and automation would eventually allow for lower costs per yield.
Power costs will be substantially reduced in the future. Moving agriculture indoors would require generating light, maintaining climate, etc, which would require substantial power. I am assuming that newer technologies with high power capabilities like nuclear reactors (possibly fusion reactors) would reduce power costs substantially in the future.
Avenues for Effective Change
Most of the agricultural land is used for growing cereals/grains, oil crops, and pulses (~85-90% of the land; Fig. 1).[6]To make effective change, we need to focus on decreasing the land use required to grow these crops.
Currently, the focus of indoor and vertical farming is to grow fresh produce. From what I gather, the reason for focusing on fresh produce is because of ease of growth in soil-less systems (like hydroponics, aeroponics, etc.), ability to scale them vertically to increase yield, and a high rate of spoilage loss, necessitating local centers of fresh produce.[10]
However, produce currently make about 3% of the total land use, so they are unlikely to make an effective dent. In addition, methods developed for these plants will likely not be generalizable for cereals, oil, and pulses. To effectively reduce land use for agriculture, the key focus should be on developing technologies that enable mass indoor farming of cereals, pulses and oil crops.
Fig 1: Agricultural land use by major crops (data source: our world in data)
Conclusion
This essay began as an initial question/thought of why vertical farming has not taken off and what are the key bottlenecks that is preventing this transition. A dive into this made me realize that while there is a huge potential for indoor farming to be a foundational and transformative, the current focus on it is not effective. To effectively deploy indoor farming and reap its multiplicative benefits, the focus must be on cereal, pulses, and oil crops that use ~85% of the land used for agriculture for human consumption.
I would love to hear what LW community thinks about this :)
Hannah Ritchie (2021) - “Drivers of Deforestation” Published online at OurWorldinData.org. Retrieved from: ‘https://archive.ourworldindata.org/20260518-093348/drivers-of-deforestation.html’ [Online Resource] (archived on May 18, 2026). ↩︎
Accounting for deforestation and land use, it seems like carbon released for agriculture is equal to (or greater than) the carbon released due to fossil fuels. Paper: Increased transparency in accounting conventions could benefit climate policy—https://doi.org/10.1088/1748-9326/adb7f2; Video. ↩︎
Hannah Ritchie (2021) - “To protect the world’s wildlife, we must improve crop yields — especially across Africa” Published online at OurWorldinData.org. Retrieved from: ‘https://archive.ourworldindata.org/20260518-093348/yields-habitat-loss.html’ [Online Resource] (archived on May 18, 2026). ↩︎
Carbon sink capabilities of forests: Pan, Y., Birdsey, R.A., Phillips, O.L. et al. The enduring world forest carbon sink. Nature 631, 563–569 (2024). https://doi.org/10.1038/s41586-024-07602-x ↩︎ ↩︎
Current advances in carbon removal technologies—refer to executive summary, point 2: The state of carbon dioxide removal: a global, independent scientific assessment of carbon dioxide removal. University of Oxford. https://doi.org/10.17605/OSF.IO/F85QJ ↩︎
Hannah Ritchie and Max Roser (2019) - “Half of the world’s habitable land is used for agriculture” Published online at OurWorldinData.org. Retrieved from: ‘https://ourworldindata.org/global-land-for-agriculture’ ↩︎ ↩︎
Reducing livestock related land use would require change in food preferences (covered in detail here) and/or development of plant-based meats (focus of good food institute). ↩︎
- ^
Number based on actual loss section in Post-harvest losses and Hannah Ritchie (2020) - “Food waste is responsible for 6% of global greenhouse gas emissions” Published online at OurWorldinData.org. Retrieved from: ′https://archive.ourworldindata.org/20251125-173858/food-waste-emissions.html’
Changes in cereal production yield in the last decade: https://ourworldindata.org/grapher/index-of-cereal-production-yield-and-land-use ↩︎
Isn’t this the most important disadvantage? I would expect that the energy conversion efficiency of vertical farms is inevitably far less than 1 because plants are that imperfect. As a result, such energy should originate not from oil [1] or gas, but from a renewable source or nuclear[2] power. Alas, renewable sources could turn out to be a similar waste of power because they cause a chunk of land to become unusable...
Unless our goal is to convert oil into food.
Or from artificial thermonuclear synthesis, but the ITER has yet to start working.
Energy consumption is definitely an important disadvantage—I indeed assume that energy is not a concern for the post, either through nuclear power or other non-land renewable sources.
Even with this key disadvantage, this direction for agriculture might be advantageous because:
Yield will scale linearly or even super-linearly with number of stacks in a multi-storied building because the constrained environment will allow for better optimizations (robotics, precision agriculture, genetic engineering) than current distributed horizontal farming.
Energy consumption for climate control, artificial lighting will scale sub-linearly for the same reasons, in addition to economics of scale.
Energy costs in general have been decreasing. Plus, widespread adoption of nuclear power due for AI capabilities might have a fortunate side-effect of reducing energy costs.
These are intuitive beliefs though and I currently don’t have numbers to back it...
A few numbers:
Most productive crops need on the order of 400-600 μmol/m^2-s of photons while growing, which corresponds to about 500-800 W/m^2 of electrical input for its “daylight” period of 10-14 hours per day (best duration depends upon type of plant). That’s an average of about 300 W/m^2 over each growing day, and we can roughly halve again to 150 W/m^2 for proportion of growing time per year.
Currently about 1.6 billion hectares are devoted to crops, which would then require an average of about 2.4 petawatts of electrical energy to illuminate suitably for growing. Currently average world production of electricity is about 0.004 petawatts, so we would need to scale it up by a factor of about 600 to meet current agricultural production.
This is extremely expensive, even if we reduce it by a factor of 10 for fantastic advances in crop genetic engineering and useful density and other technological breakthroughs.
There’s also the problem that this level of electrical power would start to directly raise global temperatures even without any greenhouse gas amplification.
We would almost certainly be better off not using vertical farming, but putting all those required agriculture technology advances directly into crops using existing solar illumination. We can still plan to use vertical farming somewhere that isn’t Earth.
Thanks for the numbers—I am starting to see that this might not be a good idea in terms of energy consumption.
I am curious about the 500-800 W/m^2 estimate—it seems too large—could you describe the calculations in a bit more detail? I am saying it is too large because on average earth only seem to receive about 250-300 W/m^2 based on solar irradiance data[1]. Additionally, this is over a wide range of wavelengths—plants don’t really use all the wavelengths so the effective power required in terms of energy of photons might be lower[2].
Just to double check, I tried to estimate the same thing in a different way. Sun irradiance provides ~800-2800 kWh/m^2 every year[3]. Assuming an average of ~1500 kWh/m^2, the total energy provided by the sun on the land devoted to crops (1.6 billion hectares) is about 24 EWh. Given that the total electricity generated is around 30 PWh[4], the energy provided by sun is 3 orders more.
What I am currently uncertain about is the the proportion of the total sunlight energy is meaningful required for a growing plant—that is:
how much effective area is used by the plant during its lifecycle (growth rate, leaf distribution, farming strategy, etc)
how much of the sunlight is actually absorbed by the plant (does not account for photosynthetic losses)
This is important considering that the land use for crop is not equal to direct photosynthesis/crop growth. If we assume only 1⁄2 of the energy is in usable wavelengths and 1⁄10 is used for effective crop growth, then we need 50x the current electricity production, very similar to your estimate.
However, what is not clear to me is the error margins in this estimate. Is it factors or orders of reduction of incident solar energy? That is, is there is an upper limit to the number of photons a photosynthetic cell can effectively use. If that amount is a couple of orders less than sun irradiance, then the electricity production requirement might be an overestimate.
Regardless, even if I am very aggressive in my estimate by assuming crops use only 1% of the total sun irradiance per year, it still is 5x times the current electricity production and does not seem feasible.
What I am currently wondering (hope to look up and a comment in a day or two) is:
Is there an expected doubling rate for electricity production and how long is it?
What are the cascading benefits of fractional vertical farming on earth? That is, if 1⁄10 of the crop area was replaced by vertical farming, is the energy cost worth the carbon capture and habitat restoration benefit?
It is clear to me that energy is a far bigger constraint than I thought—I definitely underestimated the amount of energy provided by the sun, especially considering the amount of electricity generated by humans.
From a quick estimate of this and this figure.
This figure. I did a quick estimate using STM G173 data—the power in 300-700 nm is about ~50%.
Based on this figure.
https://ourworldindata.org/grapher/electricity-prod-source-stacked