The Peter Zeihan Podcast Series - Automation of US Agriculture || Peter Zeihan

Episode Date: September 4, 2025

It's not a stretch to say that US agriculture is one of the most automated industries out there...Join the Patreon here: https://www.patreon.com/PeterZeihanFull Newsletter: https://mailchi.mp/zeihan/automation-of-us-agriculture

Hey all, Peter Zion here coming you from Colorado. Today, we are taking a question from the Patreon page, and it's specifically, it's about automation, specifically the automation of agriculture. And the question is, with the way the technology is evolving, can it solve some of our personnel and worker problems? Because we're facing a worker shortage in pretty much every sector of the economy right now, with agriculture, arguably one of the two or three sub-sectors that is most affected.

The answer is kind of eh. First, back story. I would argue that today agriculture is already the most automated industry in the United States. If you think back to as recently, as the year 1900, well over 90% of the American population was involved in agriculture in some way. But then we started to industrialize and urbanize, and people moved from the farm into the cities. And today, it's only about 1% of the American population. that lives on the farm and the ranch. That means that we've had to expand our agricultural output

against a backdrop of ever-shinking labor forces. And so when you think about what happens in a modern row crop farm these days, whether that's corn or soy or whatever, it's already mostly automated. I mean, the very concept of mechanical planters and especially combines for putting the seed down and then harvesting the end result, that's largely automated. And what we've seen in the last decade, roughly, is even those things don't need to be driven anymore.

If you step into a modern combine, you're going to feel like you're on the deck of a Star Destroyer in Star Wars. You've got readouts everywhere. It's GPS run, and the farmer's there for when things go wrong. I don't mean to suggest that they're just taking a nap behind the wheel, but a lot of the work is already programmed in. Once you harvest it, you get it into your truck. you then deliver it to an elevator where it's loaded automatically, hasn't been done by shovel for a very long time, and eventually ends up on a truck going somewhere or in a rail car, where it goes into, say, a flour mill, where the wheat is turned into flour automatically. They don't even bother turning the lights on.

You then go into something like a bread factory, a bakery. Most of that is automated too. So there are tight spots in agriculture. But there aren't very many of them. Those happen primarily when you're talking about higher value added products, specifically meats and fruits and vegetables. So with meats, you have individual critters that are running around. They need to be managed, whether by horseback or on the back of an ATV. That is something we haven't figured out to automate.

And if you remember back to your early civilizational lessons, the reason why man's best friend is dog is, is dogs are pretty good at managing herds, and we have yet to come up with us something that's better than a canine. When you're talking about processing, you're also dealing with something where eyes and fingers are what we need. When you take a carcass and you process, you're doing one at a time,

and each carcass is a little bit different. So we have found out that there are certain types of automation that we can include to look for occlusions and gross and things like that that you don't want in your, ground beef and we have figured out how to move the carcasses through the facility, but you still need people hacking things apart. We made a bit of an advance during COVID when for health reasons we couldn't have people on the line, but it really probably only shaved a 5 to 15% off of the manpower's necessary to make it all function. It's difficult for me to see the technology

in its current form, having a significant breakthrough in the next 20 years. Love to be surprised, but I don't see it. Then you're talking about fruits and vegetables. Here, it's an issue of a combination of identification and delicacy. Things like apples bruise pretty easily. Things like arugula have to be cut or asparagus, have to be cut at the right rate. And the visual recognition is getting better. And if we're going to have improvements in visual recognition,

then, you know, maybe we're talking about something. But there's three steps to it. You have to identify what needs to be cut and where. Number two, you have to have a physical arm that can go out in graphs and sever it. And then three, you have to be able to relocate that recently harvested whatever it is into the rest of the processing system. Here's the problem. When you do this on a line for manufacturing, you have a moving conveyor oftentimes that moves the product through in its intermediate form.

And then your automated system is doing the same thing to say every item and you have humans and machines to a degree working in parallel with the humans either maintaining the machines or going back and forth on the line doing the things that require our brain and fingers. The machines, even with AI, are pretty dumb. They might be able to do an individual step better than human, but they can really only do that individual step. There are tens of thousands of individual steps. You can have machines doing a lot of that, but the product has to be brought to them. These facilities are not mobile. These machines are not mobile. When you're doing with agriculture, I'm sorry, but the apple tree is not going to rotate so that the fruit is facing the direction it needs to.

You have to bring the machine to what you're harvesting. And that requires a degree of locomotion and energy intensity and programming that we just don't have yet. So incremental gains, sure. I would say probably apple harvesting is coming along the fastest in this regard. It's a relatively high capital industry. And as a rule, the apples are a different color from the tree when they're ready to be harvested. That provides a lot more visual cue. And when you have them weighing down, they separate from the leaf bundles a little bit. But if you want to do this for spinach, sorry, you need Guatemalans for that. And one of the things we're going to be seen over the remainder of this year is a separation in terms of pricing between the parts of agriculture where maybe a degree of automation can help and those where it really can't. And unfortunately, the sectors within the United States that are most reliant on the irregular immigration that has become a political hot potato in this country are the ones that are actually least capable of being automated.

So if you're in an area that imports its fruits and vegetables from elsewhere in the country, you're going to notice that at the grocery store. That's unavoidable at this point.