Farming in the Sky

http://www.paulgraham.com/submarine.html

Historically these greenhouses depended on being able to sell a higher quality product during winter months than produce imported from California, Florida and Mexico. The technology is better and now that more people are interested in organic there's an audience that will pay higher prices than the supermarket year around. But it is a small audience, these greenhouses will never feed large numbers of people.

The key to growing more high value food (protein and fat as opposed to low nutrient density fiber, vitamins and water foodstuff like this system provides) is to remove animals from the equation. This is the terrible inefficiency in modern agriculture and solving it would result in tremendous gains.

I'm very excited about things like tank grown nutrient yeasts. There was an article a few weeks ago about a type that has been engineered to produce milk for instance. This type of thing is the future I believe and has great potential. A few off season indoor vegetables are great too. As condiments.

When they get alternatives to grains, pulses, meats and milk then real advancement in food production will being made.

I mean imagine building a home for every plant you eat. It'd be kind of crazy right? You'd need to provide many of the things a typical home has: shelter, electricity, plumbing, food supply, structures etc.

But this is really what is proposed in these types of farms. Instead of available soil, you build a large structure, pots, racks and tracks. Instead of rainwater and rainwater distribution systems (clouds) you build pipes and tubes and plumbing and install pumps to move it all around. And you provide the plants nutrients in this way. And then you build a large shelter to keep everything away from the elements that mostly grow the plants outside. And then you install artificial lighting and electricity systems to pump energy into the solar panels that we know as leaves, to generate food.

For the longest time this has been hugely expensive. And to feed 7 billion people with about 2-3 plants or whatever it is, worth each day (in cereal, greens, fruits and legumes etc), you need to build billions of little plant homes.

And that was just never very economical. Let's not forget that a hectare yields an average of about 5.000 kg of cereal. Wheat goes for about $200 per metric ton, so a hectare nets you about $1k.

Now that's 107k square footage. That's incredibly cheap. Now imagine you want to replace that with a hectare of factory or say the plant houses I described earlier. That's a huge capex investment in the tens of millions compared to the hectare of soil which is priced around $10-15k and can be leased for a few hundred. To compete on price you'd need very high production in cheap economies of scale to offset your fixed costs on building this huge greenhouse structure with infrastructure for every individual plant.

But that's also where the land is much cheaper. It's virtually impossible to run vertical farms on soil, so you need a soilless medium which isn't a problem, but it means you need plumbing to every plant. No biggie. But the same for light, you can't do a vertical farm on scale (i.e. rows and rows) and expect sunshine to reach the bottom plants who sit in the shade. So you need artificial lighting and heating, and that's very expensive compared to sunlight.

High value crops, sure it's fine. (e.g. see every indoor weed farm with artificial lighting etc). But as the future of agriculture (where staple foods worldwide are cheap cereals, legumes, corn etc), no way.

Unless you have very large advances in material science (cheap, sturdy materials to pump out large greenhouses cheaply and sustainably) and energy (cheap solar pv/thermal, wind, geothermal). And that's what we're seeing now. So it's definitely getting more appealing by the day and it's something I've been very interested in pursuing because I just adore the idea of automated farming on some levels (I love to hike in nature, but large corn fields aren't really nature to me and don't excite me more than the idea of an automated corn factory!), but I still don't see how the economics make sense.

And it's more than finance, it's touted as an environmental solution because these farms can be located near the point of consumption. But that's just one part of the story, and incredibly myopic if it's the only part of the story that's considered. For example the problem in community supported agriculture I see is that people will drive a car individually, once a week, to a small location that doesn't get economies of scale, to pick up a tiny amount of food. Because they're under the impression it's more environmentally friendly than a large shipping container moving their beans across the ocean. When in reality their food miles is a bit like saying it's environmentally more friendly drive 2 miles to a store, than to cycle 5 miles to a store for your groceries because 'it's less food miles that way'. Sustainability is a very serious issue and I'm not entirely convinced such factory-style farming is more sustainable (see inherent investments to move from soil to building plant houses). In terms of water usage, it's actually amazing to use a soilless medium, and water is one of this century's biggest challenges. But in terms of materials and energy... I'm very skeptical.

In my area - Southern Ontario -- farmland is cheap and plentiful, a lot of it turning to brush or "hobby" farms (which usually means farms that don't really produce). In my area just outside of the metro area, everyone's property, with long growing seasons and excellent soil -- are in the measure of acres, and only grow decorative plants.

We recently as a family started considering the notion of mixing up life and moving somewhere else in Canada -- 100+ acre farms in PEI, NB, and Nova Scotia can be had for under $100,000. Many are now considered vacant land having been abandoned as farmsteads.

Whenever these sorts of articles talk about the "solution" to locally grown or a food crisis, I compare these two facts and something is not meshing.