Sometime in 2025, a company called FarmWise quietly turned off the lights.
Its robot weeder was the real deal — a machine that could run a row and kill weeds with a laser and a blade (no chemical, no hoe crew), exactly what a superweed-plagued, herbicide-litigated industry has been begging for.
It made Time magazine’s list of best inventions. It won awards at the World Ag Expo. It raised something north of $65 million. And it’s gone.
FarmWise didn’t fail because the technology didn’t work. It failed because the machine cost more than the farm could absorb — and that is a different kind of failure than the one we usually talk about.
It wasn’t alone. Researchers at the University of Nebraska-Lincoln counted roughly 18 ag-tech companies that shut down or wound down in 2025 and found the same fingerprint on most of them: The science worked, but the farm-level economics didn’t. They gave it a name: the Cost-Adoption Mismatch Effect. The tool was too expensive, too clunky or too disruptive for the operation it was built to serve.
I’ve been waiting on the white paper to render that verdict for more than a decade.
A Room in Silicon Valley
Years ago I spent time at The Mixing Bowl, a Silicon Valley forum the late Rob Trice built in 2013 to marry information technology to food and agriculture. Founders took the podium certain their app was going to fix the problem on the farm. Build enough apps, the thinking went, and you’d solve everything.
I thought most of them had a fringe view of what a farmer actually wrestles with daily and across a season. They had a technology and were out hunting for a problem to hang it on. The problem? They had it backwards.
Trice, who died in February and knew this business as well as anyone alive, spent his last years saying the quiet thing out loud. In one of his final interviews he called the ag-tech market “constipated,” said artificial intelligence (AI) had become “the belle of the ball, sucking oxygen away” from everything else, and admitted the whole project was harder than he’d bargained for.
A decade and a mountain of apps later, a lot of those same farm problems are still sitting in the same fields. The man who built the front door to ag-tech came to call the room constipated. He wasn’t wrong.
The reason isn’t that farmers are slow. It’s that every tool has to clear two gates to survive on a farm. Most die at the first. The best ones — the FarmWises — die at the second. And most of the people selling to agriculture still don’t know the second gate is there.
The $4 Fence Line
The first gate is cost, and it’s a narrower opening than anybody outside farming thinks.
It’s not, “Does it pay off eventually?” The real question is whether a grower can clear the price this season, at scale, on margins that in 2026 range from thin to underwater. There is no slack in the checkbook to absorb a technology that doesn’t pencil the first year.
Ankit Chandra, one of the Nebraska researchers, figures a row crop grower can afford to spend maybe $4 an acre on a new technology before the math breaks. That’s the width of the first gate on corn and soybean ground — his estimate, but the truest $4 in this conversation.
FarmWise’s weeder cleared every engineering bar and never came close to that one. The survivors of 2025, the Nebraska crew found, were consistently the low-cost, easy-to-run tools.
But clearing $4 an acre only gets you through the first gate.
The Way We’ve Always Done It
The second gate is the one nobody budgets for: habit, workflow, trust — culture, if you want the $10 word.
“The way we’ve always done it” gets mocked as stubbornness. It isn’t. It’s arithmetic. Howard Buffett, a farmer himself, built a book around it: “40 Chances.” The idea is that a farmer gets about 40 growing seasons in a working life. Farming offers one validation cycle a year. Switch seed, switch program, switch practice; if you’re wrong, you don’t find out until fall, and you’ve used one of your 40 on a bet you couldn’t afford to lose.
When being wrong is that expensive, that slow to show up and that scarce, caution is survival.
We already know which tools clear the second gate, because we’re running them. Auto-steer. Guidance. Planter clutches cutting the seed bill as you cross that invisible headrow in the field. They scaled because they live inside iron and routines the farmer already owns.
The tool didn’t demand a new habit; it hid inside an old one.
Cheap Isn’t the Same as Adopted
Watch this play out now with biologicals, one of the hottest categories in ag and a useful test of the idea.
Biologicals sail through the first gate. A jug bought by the acre slips inside that $4 fence line far easier than a $200,000 machine ever could. Cheap gets them funded. Cheap doesn’t get them adopted.
Then they hit the second gate head-on. A biological’s payoff swings with soil, weather and the season, so the farmer often can’t see it work in a single cycle — and a benefit you can’t see, you won’t trust or rebuy.
Which is the whole point: Cheap isn’t the same as adopted. You have to clear both gates.
Two Roads at the Gate
Let me be fair to that room in Silicon Valley. I owe those founders a better verdict than I gave them.
The issue was never that farmers won’t change. We change constantly — mechanization, hybrid seed, Roundup Ready, GPS. We just don’t change on a technologist’s calendar; we don’t change for a tool built from the tech backward instead of from the dirt forward.
Even hybrid corn, arguably one of the best returns this industry ever produced, took the better part of 13 years to cross the Corn Belt — and it traveled neighbor to neighbor, not spreadsheet to spreadsheet.
Which leaves everyone who is selling to agriculture at a crossroads.
One road keeps building tools that clear a demo and flunk the farm — priced past $4 an acre, bolted onto a routine no one has time to relearn, sold on an ROI the grower can’t see before Christmas. That road runs straight to the 2027 shutdown list.
The other builds to the $4 and to the workflow — intelligence tucked inside the iron, inputs and paperwork a farm already runs, solving a problem from the ground up instead of asking the ground to change. Trice spent a decade finding which road reaches the field; the market just spent another hard year proving him right.
The tool that finally moves the needle on your operation probably will show up the way auto-steer did: Nobody signed up for a “guidance platform.” They just bought a tractor that drove straighter than they could.
It’s something already inside the machine, already in the jug and already in the workflow, quietly clearing both gates like they were never there.
Steve Cubbage is a precision ag consultant and farmer from Nevada, Mo. He is the founder of Longitude 94, an agriculture sustainability and technology consulting business.
