This is a somewhat long read, but I’ve been spending a ton of time lately thinking about how AI might disrupt agriculture and how many of my ag-related investments could be impacted. I tried to drag together a lot of my thoughts, and I feel like the comparison between AI and the railroad is somewhere we can all find common historical ground, and this might help us better understand how things will play out with AI. I find all of this crazy important and interesting, so I wanted to share some insight…
Historians around the world argue that the railroad was the single most important tool that enabled the United States to become the world’s greatest economy. It wasn’t so much the transportation of people as the free movement of goods and commodities across such vast land that enabled explosive growth! In fact, it was on this week in 1869, the presidents of the Union Pacific and Central Pacific railroads met in Promontory, Utah, to drive a ceremonial “last spike” into a rail line that connected their two railroads. This made transcontinental railroad travel possible for the first time in U.S. history. No longer would western-bound travelers need to take the long and dangerous journey by wagon train, and the West would surely lose some of its wild charm with the new connection to the civilized East.
Modern analysis suggests that AI is currently following the same pattern as the railroad, electricity, and the internet. The adoption of artificial intelligence and the railroad both represent transformative technological shifts. Historically, the railroad vastly improved transportation efficiency, connecting markets and reshaping industries. AI is similarly expected to enhance productivity and create new economic opportunities. Both technologies require significant infrastructure development. The railroad involved physically building tracks and stations, while AI relies on advancements in data processing and cloud computing. The adoption process can lead to shifts in the job market, as certain tasks become automated and new roles emerge. The long-term impact on society and the economy will continue to evolve, presenting both benefits and challenges.
The development of the railroad in the 19th century was a catalyst for radical shifts in the American economy, simultaneously creating and destroying entire categories of jobs and businesses. The railroad didn’t just move things; it fundamentally enabled explosive economic growth by facilitating the free movement of commodities across vast distances. Below are a few interesting examples…
- Emergence of “Boom-Towns”: Previously uninhabited or sparsely populated areas became thriving towns almost overnight when a rail line came through.
- Terminus, Georgia, originally a small town of thirty, was chosen as a rail terminus and was renamed Atlanta, eventually becoming a major Southern hub.
- Kansas towns like Abilene and Dodge City became critical transshipment points for the cattle industry.
- New Industrial Sectors: The massive engineering challenges of building the transcontinental railroad led to numerous technical innovations that spurred new business sectors:
- Bridge Engineering: The need to cross waterways led to the development of truss and cantilever bridges.
- Steel and Iron Manufacturing: Early wooden tracks were insufficient for heavy locomotives, leading to a massive demand for iron and eventually steel rails.
- Agricultural Equipment: The ability to ship products to distant markets allowed companies like McCormick Harvesting Machine Company to build vast sales and repair networks that spanned the continent, something impossible before the speed of rail.
- Specialized Logistics and Services:
- Cattle Industry: Texas cowboys drove herds to railheads for shipment to stockyards in cities like Chicago or Kansas City. Eventually, local butchers and small-scale slaughterhouses were replaced by massive centralized stockyards. Next, the invention of the refrigerated railcar (or “refer car”) in the 1840s allowed fresh meat to be shipped thousands of miles, changing the American diet from mostly salted meats to fresh beef (another major shift or change).
- Agriculture & Fresh Produce: The railroad ended the era of “seasonal-only” eating by creating regional produce specializations. Farmers were no longer limited to selling in their local village. If you were in Georgia, you could now sell peaches in New York; if you were in Washington, you could ship apples across the country. This created a massive demand for icing stations along rail lines, where workers would reload cars with up to 11,000 pounds of ice to keep cargo fresh during transcontinental trips.
- Flour Milling: The railroad allowed Minneapolis to become a center for the grain trade, with entrepreneurs like Charles Pillsbury leveraging rail to build worldwide brands.
- Retail & National Branding: The railroad allowed local craftsmen to become national household names. Brands like Stetson were now able to turn local prototypes (the “Boss of the Plains” hat) into national necessities by shipping them via railroad to cattle towns and mining camps. Eventually, local hatmakers and toolmakers found themselves competing with national brands, and this greatly disrupted the local economies.
While the railroad eventually created a “positive-sum force” for the overall economy, but it also displaced older, less efficient industries. So, not only do we need to think about all the great things AI will bring to the economy, but we also have to think long and hard about what AI will displace. Below is just a small sample of the industries the railroads displaced.
- Wagon Trains and Boat Travel: Journeys that previously took months by wagon train or weeks by boat were reduced to mere days by rail. This essentially rendered the business of long-distance wagon freight and passenger travel obsolete.
- Local Self-Sufficiency: The ease of shipping goods meant that small-scale local manufacturers and farmers who couldn’t compete with the efficiency of large-scale operations often went out of business.
- Craft Displacement: As automated machinery was shipped more easily, traditional hand-labor in various fields (like woodchopping for fuel, as whalers and woodchoppers were later affected by fossil fuels and electricity) began to slide into obsolescence.
Just as people feared the tractor would “break” the labor market (it dropped farm employment from 33% to 2% while tripling output), analysts argue that AI will not cause permanent unemployment. Instead, it will pull human work “up the stack” to solve more complex problems.
- New Frontiers: Just as the internet created “cloud migration” or “eGrocery” jobs that were previously unimaginable, AI is expected to unlock a universe of robotics jobs and new service industries that have yet to be imagined.
But what happens to the people who aren’t smart enough or skilled enough to be pulled “up the stack” to solve the more complex problems? This is the core tension of any industrial revolution. When we say workers are “pulled up the stack,” it sounds optimistic, but historically, that transition is rarely smooth or universal. Looking at the railroad and previous shifts, there are three distinct ways this plays out for those who don’t—or can’t—transition into more complex cognitive roles:
1. The Expansion of the “Service Economy” – The railroad didn’t just create high-end engineering jobs; it created a massive surge in service and physical labor that didn’t exist before. I’m thinking AI creates more physical blue-collar jobs serving data centers, robots, etc…
Railroad Era: While wagon drivers lost their jobs, the railroad created an explosion in demand for hospitality (hotels at rail stops), localized delivery (short-haul drayage), and maintenance.
AI Era: We may see a massive “return to the physical.” If AI can handle the “smart” tasks (legal research, coding, accounting), human value may pivot back to things AI cannot do well: trades, tactile care (healthcare/nursing), high-end hospitality, and specialized physical craftsmanship.
2. “Deskilling” vs. “Upskilling” – There is a concept called Efficiency-Driven Access. Sometimes, technology doesn’t require you to be smarter; it makes the job easier, so more people can do it.
Railroad Era: You no longer needed the survival skills of a frontiersman to move goods across the country. The “system” (the tracks and schedules) handled the complexity, allowing thousands of people to work in logistics with less specialized training.
AI Era: AI might act as a “leveler.” A person who isn’t a great writer or a math whiz can use AI to perform at a professional level. Instead of “moving up the stack,” the stack moves down to meet them, allowing people with average skills to produce more high-value results. Think about it like this… trains allowed almost anyone to become a merchant or seller of goods nationwide (if they wanted to). Whereas AI is allowing anyone to become a creator or analyst. AI is essentially “leveling the playing field” (making complex cognitive tasks easier for everyone).
3. The Structural Lag (The Hard Truth) – Historically, there is always a “lost generation” during these shifts.
Railroad Era: If you were a 50-year-old stagecoach driver in 1870, you likely didn’t become a telegraph operator. You likely saw your wages stagnate or were forced into lower-paying manual labor.
AI Era: The risk is a bifurcated economy. If cognitive labor becomes cheap, the “premium” on human intelligence could disappear, leaving a gap between those who own the AI (capital) and those who perform the remaining manual tasks (labor). Think about it like this… trains brought about the rise of labor unions to protect remaining workers (think about the Teamsters, who were founded as more local delivery drivers were needed as the trains started delivering more goods to terminals across the country). On the flip side, AI is already bringing about more and more discussion around Universal Basic Income, shorter work weeks, etc… In other words, “lump-of-labor” fallacy suggests there is always more work to be done, but it doesn’t guarantee that the new work will pay as well as the old work for every individual.
If the railroad taught us anything, it’s that the “losers” are usually those who provide a slow, expensive version of what the new technology makes fast and cheap (like the wagon drivers). The “winners” are those who use the new speed to solve much bigger, previously impossible problems. This comparison strikes at the heart of how “General Purpose Technologies” (GPTs) like the railroad and AI don’t just improve efficiency—they change the unit economics of human endeavor. In other words… In the railroad era, the “losers” were people like wagon teamsters and canal boat operators. Their failure wasn’t a lack of effort; it was that they were competing on the same dimension as the railroad: moving a thing from Point A to Point B. When the railroad reduced a six-month wagon journey to six days, the wagon driver’s value proposition became mathematically impossible to defend. Even if the wagon driver worked for free, he couldn’t compete with the railroad’s speed. Today, the “losers” are likely those who sell routine cognitive output—basic copywriting, standard coding, or data entry. If your value is simply “producing 500 words on a topic,” “basic contracts,” or you have set up an expensive tollbooth tied to software, you are more than likely going to end up like the wagon driver. AI has made the unit cost of that specific output effectively zero.
On the flip side, the real wealth of the railroad era wasn’t made by people who simply “used a train instead of a wagon.” It was made by people who realized that speed and scale allowed for entirely new business models. Before the railroad, if you wanted to eat beef in New York, the cow had to be walked there. It was expensive, the meat was tough, and most of the cow was wasted. Entrepreneurs like Gustavus Swift, however, didn’t just “use the train.” He invented the refrigerated railcar. This allowed his business to slaughter cattle in Chicago and ship only the meat to the East Coast. In other words, he solved the “impossible” problem of providing fresh, affordable protein to an entire nation. He didn’t compete with wagon drivers; he created an industry that couldn’t exist in a “wagon world.”
As for agriculture, and how AI might eventually play out and make a major impact on a larger and wider scale, is going to take a bit of time. At the moment, AI just isn’t providing the in-the-field speed that is going to create major shifts or changes in production. But, I think it’s still way too early, i.e., meaning, we are at a similar point in time to the early 1800s, before all the rail tracks were laid, before all of the infrastructure along the railroads was built, and before the real speed of movement became a reality. Eventually, however, I suspect we will first start to see the agronomist and agronomy industry massively disrupted. Then we will start to see the ag retailer massively disrupted as equipment, inputs, and practices shift more dramatically. In technical terms, this shift represents the transition we have been hearing about for years, from “Systems of Record” to “Systems of Action,” where data moves beyond simple storage into proactive, intelligence-driven personalized execution. At the same time, ag retailers are already facing margin pressure and increasing complexity, forcing a move toward more integrated, high-value service models. We are already starting to see a trend toward “outcome-based financing” and crop warranties, where financial products are linked to the proven performance of specific agronomic recommendations.
The problem right now in agriculture is “speed,” meaning there is already a major labor shortage and lack of help on the farm. So, at the moment, most will argue that AI is only helping us make smart on-farm decisions; it’s not really helping us get a lot more done. Just like all the new “see-and-spray” technology. It helps us reduce input costs by only spraying where it needs to be sprayed, but it has yet to do anything to help with the labor shortage of getting the field sprayed in a quicker, more timely fashion. But just keep in mind, the railroad’s greatest gift was that it didn’t need to sleep or rest like a horse. Once all the tracks are laid and infrastructure built, I can almost guarantee that AI will bring that same “infinite endurance” and speed to the farm, i.e., Current tractors require a human operator to be present, even with GPS. New systems, like John Deere’s Autonomous 8R and 9RX, allow a single farmer to manage a fleet of machines from a phone. Instead of one worker tilling one field for 12 hours, one worker can monitor three machines running for 24 hours. This effectively triples the speed of operation per human hour.
We aren’t seeing a total reduction in “farm time” yet because the infrastructure (data connectivity and hardware) is still being laid—much like the tracks being laid for the railroad. Once 5G/Satellite connectivity and autonomous hardware reach “Standard Gauge” levels of adoption, the speed of agriculture won’t be limited by how fast a human can service a field, but by how fast a machine can recognize and react to data being processed out of a field. For the moment, the “labor shortage” in your area is acting as somehwat of a barrier that slows down the AI adoption (simply because many producers are too busy with daily chores and work, because they don’t have enough help, to use and implement the newest AI tools and technology). But eventually, this “shortage of labor” will become the very thing that forces many producers to adopt it faster, or become less and less competetive.
The railroad was completed despite “extreme loss of labor” from harsh conditions and dangerous work. Today, the “loss of labor” isn’t due to avalanches or extreme conditions, but to a shrinking, aging rural workforce. This is why the “labor shortage” will eventually be the primary “accelerant” for the AI revolution in agriculture, and you don’t want to miss this big pivot.
Historically, when labor is cheap and plentiful, businesses have little incentive to innovate. In other words, why buy an expensive machine when you can hire ten people? Think slavery. Once slavery went away, a lot of big things changed and happened with technology. As the cost of labor rises—or the labor simply disappears—the ROI calculation for technology, or in this case AI, changes overnight. A farmer may have balked at a $500,000 autonomous tractor two years ago. But if they can’t find a driver for a standard tractor at any price, that $500,000 becomes a “survival cost” rather than a “luxury upgrade.” The labor shortage turns AI from a “nice-to-have” into the only way to truly continue to compete at scale.
Agriculture is becoming increasingly technical, but the available labor pool often lacks specialized agronomic training. But AI acts as a “force multiplier,” making things simpler and easier. Remember, above, when I mentioned that once the railroads were in place, one no longer needed the survival skills of a frontiersman to move goods across the country. The “system” (the tracks and schedules) handled the complexity, allowing thousands of people to work in logistics with less specialized training. So at the moment, instead of needing five experienced crop-scouts to walk thousands of acres, a producer can use one worker to launch a fleet of AI-enabled drones. The AI does the “thinking” (identifying pests/diseases). As noted earlier, the railroad transformed America because it provided a “speed and ease of travel” that wagons couldn’t match; eventually, that will happen in agriculture. Right now, human workers are limited by fatigue, law, and daylight. In a labor shortage, a farmer can’t run a second or third shift. But eventually, autonomous systems that don’t have “shift limits” will take over. Many argue that simply removing the human from the cab, a farm can effectively double its productivity without adding a single person. This allows a shrinking workforce to manage a growing amount of land.
We all know that in production agriculture, timing is everything. There is a very small “window” for planting and harvesting, and the “labor shortage” is currently controlling and dictating what technology is being adopted. Eventually, however, large-scale producers are going to see that they can use AI to “de-risk” the calendar by automating the most labor-heavy parts of the season. This is when we will start to see savvy producers trading the variable risk of human labor and availability for the fixed cost of technology! In short, the labor shortage is the “last spike” being driven into the old way of farming. Eventually, it will force a transition where the “smart” farmer isn’t the one with the most workers, but the one who best manages the most “intelligent” machines. I’m not saying I like all of this change, but I am working to better understand it so I can help my friends and family better position themselves for the future. Lots to think about, sorry it was so long…
