The Future of Food from First Principles

This article contains three parts:

1. An update on vertical farms, based on what I’ve learned since I first published on the topic. The gist of it is that I now think that, in the next few decades, vertical farming won’t be viable for the most important crops, but it will be for some, in certain circumstances.

2. Do we even need vertical farming? (paywalled from here)

3. How should we think of food from first principles? As energy transformation.

I do not discuss yet animal protein, as it’s an entirely different animal. I might in the future.

---

1. How Viable is Vertical Farming in the Short Term?

In The Promise of Vertical Farming, I explained why I thought vertical farming was the future of farming: It is much more efficient than outdoor farming in terms of water, fertilizer, pests, transportation, land… Its downsides are the higher costs of investment, energy, and labor, but all three will shrink as the industry is refined. As a result, I concluded that vertical farming would take over the world at some point.

In How Can Vertical Farms Become Viable?, I explained the path to widespread establishment of vertical farms:

• Start in places where the strengths of vertical farms outweigh the downsides, like in the UAE or Japan.

• Start with certain crops, like lettuce and tomatoes.

• Start with greenhouses and gradually transition to vertical farming, improving efficiency.

• Further improve efficiency with genetic engineering.

In the comments, reader Samuel Lovat shared his draft paper where he details math on vertical farming viability. The paper’s key insight compares the cost of outdoor-farmed food with the cost of energy of vertical farms. I think it’s interesting to look at that math (takeaway in bold if you are allergic to math).

He considers that, in vertical farms, you need 250 kWh of electricity to produce one kg of dry food from plants, which translates to a transformation of just 2% of the electricity energy into dry food energy. Since the cost of a kWh is about $0.04, the result is that producing 1 kg of dry food from plants in vertical farms costs $10 in energy alone.

This compares to a cost of about $0.5/kg for dry food from outdoor farming. Obviously, since just the energy cost of vertical farming is 20x more than all outdoor farming costs, it means vertical farming is not viable for the average crop.

So how can we reduce that cost?

Left: plants with a share of dry food as high as 95% include staples like wheat, rice, and corn. Right: plants with a share of dry food as low as 5%, like herbs, tomatoes, and lettuce.

This dry matter thing is interesting. It takes the water out of the food. Since staples like wheat or rice are nearly all dry food, a kg of rice is basically a kg of dry food. But 95% of the weight of a tomato or a lettuce is water. So growing 1 kg of lettuce in a vertical farm (VF) just needs $0.5 in energy—we’re in the right ballpark. 

That’s why lettuce, herbs, and tomatoes are already grown in VFs: They’re cheap to grow because they’re mostly water, and it costs little energy for plants to use water. Since VFs waste less water than outdoor farming, plants with a high water content can become cheaper in VFs.

If electricity prices drop with increased use of solar electricity, we could assume the $0.04/kWh might shrink by a factor of 4x. 

We can also assume that the share of electricity that becomes food will increase. This article suggests this path:

This is an 8x improvement, from 0.6% to 5.4%. 

The paper I cited before posits instead that we can go from 2% energy-to-food efficiency today to 10% (through improvements in things like LED lights), for another factor of 5. 

Together, these improvements (electricity cost and photosynthesis optimization) would yield 20-30x improvement in productivity. That would make the cost of the energy to grow staples in VFs similar or slightly lower than the overall cost of outdoor farming.

It’s important to note here that, from an energy standpoint, these types of optimizations would make growing non-staple vegetables in VFs viable much faster. Things like cucumbers, zucchini, bell peppers, carrots, cauliflower, eggplants, peas…

For us to consider growing staples like wheat, corn, and rice in VFs, electricity costs might just be too high for a long time. We would need significantly cheaper electricity and aggressive genetic engineering.

What type of genetic engineering are we talking about?

The carrot to the left has a “low harvest index”: The share of the plant that is edible is low. The carrot to the right has a high harvest index close to its limit of ~1. Source.

Engineering designed to:

• Reduce non-edible parts of the plant

• Change the circadian rhythm of plants so they can grow for more hours

• Stop wasting oxygen through photorespiration during photosynthesis

Overall, these papers and articles have made me realize how hard it will be to reduce the use of agricultural land at scale: For that, we need to replace outdoor farming of staples to VF farming, and just based on electricity, there’s no way that will happen in the next few years.

Aside from that, this research supports my previous conclusions:

• VFs are niche today.

• There’s a path for more and more vegetables to be grown in VFs, especially as electricity costs plummet and agricultural technology efficiency increases.

• But to tackle staples, we will likely need genetic engineering.

Since this process will take many years, there’s another question here: Will we even need VFs?

2. Will We Even Need Vertical Farms?