How our century of supersized farm machines will end

The American farm industry is truly amazing to behold.

We took the landrace grain of the indigenous Americans and–with an arsenal of anhydrous ammonia and Roundup Ready genetics–went down to Mexico, smashed open their gates with NAFTA, and hammered their local corn producers to death in just a few years¹. They couldn’t compete with Supercorn USA, and are only now beginning to find a spine².

We blew their ag sector up so badly that the economic fallout supercharged the narcocrisis and triggered our own domestic political crisis thanks to illegal immigration and illicit border crossings.

Yes, I am claiming that American grain producers inadvertently set into motion a chain of events resulting in the 2016 and 2024 American presidential outcomes. I do genuinely believe this.

A key part of this “success” is the cycle of innovation and invention by major American agricultural equipment manufacturers like John Deere, Massey Ferguson, International Harvester, and others. They responded quickly to market needs. The producers kept asking for bigger machines, and their prayers were answered.

Unfortunately, I think America’s large farming operations are headed for a bad place by the end of the decade as their capitalization in equipment becomes dead weight.

Trying to extrapolate the expected timeline for this change is hard because the last major transition (horses to tractors) involved going from many small, cheap investments (individual animals) to a single large, risky investment. This time, we’re reversing that cycle - farmers will have the option to make smaller, safer investments into autonomous humanoid labor.

In fewer words: how many of those expensive farming implements could compete with an army of autonomous serfs working around the clock?

Let’s walk through a few illustrative points.

Combine vs serfs

A big grain operation runs a combine that costs a million dollars (new) and harvests approx. 100 acres a day³. From old farming records from the early USA, we know that a man can harvest about two acres a day by hand. Given that robots don’t get arthritis and don’t mind working in the sun, you could probably get 200% (4 acres / day) of human’s performance over the same 24 hours. This tells us that roughly 40 units of autonomous labor would be needed to match the million-dollar combine with a break-even price of \(\$1,000,000 / 40 = \$25,000\) . Given that new humanoid robots are coming out at or below that level, I see unfavorable signs for the future of big iron.

If you are thinking of investing in a brand new combine, do you really want to bet that big tech cannot produce a robot smart and durable enough to harvest corn?

There’s also a ton of fringe benefits from downsizing.

Big equipment is hard to maneuver and keeps you from farming right up to the treelines. It’s heavy, too, and can’t be deployed in wet conditions that would otherwise be fine for planting or harvesting. Since the 80s, nearly 30 million acres of farmland have been taken out of production⁴, mostly because the parcels were too small or awkwardly placed.

With smaller units of equipment, the probability that all of your units are out-of-commission simultaneously due to mechanical failures is exponentially smaller. You can’t split up a combine and only use half of its capacity at a time. With a single large piece of equipment, you are hosed if it fails.

Another thing - using big machinery forces you to plant in rows. Rows are not the optimal solution to the sphere-packing problem. Hexagonal or offset-row designs would allow for denser plantings for a given minimum distance between seedlings. We’re leaving a possible 10% of yield on the table just because straight rows are not space-efficient. This effect is especially pronounced in more northerly regions⁵ where sunlight is more limited.

Most importantly, you can’t use a combine for anything else, like planting trees, stocking fish ponds, or pressing oils. The average utilization for the harvester referenced above is roughly two percent⁶. This is an insanely weird use of capital, made possible only by the fact that this equipment is too bulky to move over large distances and you can’t share your neighbor’s harvester because they need it too. Compare this to the manufacturing industry, where average utilization is more like 70%.

Diversified farming from the labor surplus

We used to have apple trees on our farmland. I’d love for us to have more of them, but they can be a lot of work. If you offered me ATaaS (Apple Trees as a Service), would I bite?

You can bet your ripe, red bottom that I would.

Everyone loves the idea of diversified farms producing a range of tasty goods like bacon, applejack, and butter all from the same 40 acres. Enviro-lefties appreciate that the chemical requirements and ecological side effects are lessened. Some rural conservatives vaguely remember what it was like to milk a cow and dress a chicken when they were young farmers-in-training.

Diversified farms went away for a simple reason: they’re a lot of work for less money, and the opportunity cost got really high. Family sizes dropped and kids left the farm, leading to a decrease in cheap labor. Suppose you take the analysis from the previous section to heart and got yourself a team of a few dozen autonomous robots. They’ll be busy during harvest time, of course, but what about the rest of the year? Could they perhaps be sweeping out the chicken coop and pruning the fruit trees?

My expectation is that this abundance of versatile humanoid farm machinery will lead to a resurgence in diversified ag production, simply because there will be excess autonomous labor available outside of the harvest season.

Unscrambling our housing market

I also think autonomous humanoid robotics for farming will have substantial knock-on effects in the desirability of rural America from a housing perspective. Having a few acres with your own fruit trees, a walking path, a few fences or stone walls, a garden, a few small animals, and nice landscaping is in line with our current fantasies⁷. But, again, it is an issue of time and labor.

In the USA, we are so used to experiencing rural areas which are aesthetically ugly–not because we have no sense of taste, but because there is so much other work to do–that I want to speculate a bit: make farms pretty again, and you can add half a trillion dollars in real estate value to America’s interior.

There’s yet another benefit - all the awful red tape and NIMBYism that ensnares would-be homebuilders generally doesn’t exist in rural areas, and there are fewer hurdles (aside from availability of labor) to new construction.

Economic analysis

Bringing these ideas together, we can start to form a concrete analysis of the actual delta in value from debiggifying mechanized agriculture:

Conclusion

In all, I’m unreasonably optimistic about the future of American farmland despite the avalanche of technological change that’s about to sweep through labor, markets, and society overall. Hopefully we’ll get right to repair as well!

---