Short observations from inside the work. Why new agriculture stalls between proof and market, and what the products that make it through tend to have in common.
Most agtech doesn't die at the proof stage. It clears the science, clears regulatory review, and then spends two to five years trying to earn farmer adoption while competitors emerge and investor patience runs out. The gap that kills most products isn't between idea and proof. It's between proof and the grower's purchase decision.
A better number on a trial plot doesn't move a grower. Adoption happens when they trust the people recommending it, when it fits the financial risk of their operation, when their agronomist has seen it perform in conditions like theirs, and when the channel can actually deliver and support it. Those conditions don't come out of a field trial. They get built, deliberately and in order.
For a genuinely novel technology, the stewardship plan is often where adoption is won or lost. Treated as a compliance afterthought, it slows the launch and confuses the channel. Designed as part of the commercial story, it becomes a reason growers and regulators trust the product in the first place.
When a product is the first of its kind, the work isn't explaining what it does. It's building the frame of reference that lets a grower judge whether it matters. HarvXtra had no quality-driven precedent in the seed channel. ThryvOn had no category at all. In both cases the launch was really a category-creation problem wearing a product-launch costume.
Capital in agtech is more patient and more selective than it was at the peak. That isn't bad news for good technology. It rewards teams that can show a credible path to the grower, not just a promising trial. The companies that get funded and acquired from here will be the ones that treated commercialization as a discipline from the start, not a phase to figure out after the science worked.
Most agricultural regions have one company that dominates their patents, and the reflex is to discount the region's numbers because of it. The research points the other way: a large local research company makes a region better at using its own science. The useful question is what stands behind the anchor, so every reading should be taken twice, with it and without it. How that looks →
Patents are the best single record of where invention happens, and the worst record of whether anyone used it. The trail that matters runs from patent to company to capital to payroll, and it usually goes cold early. The other half of the story is the ground: how many soils and crops a new product can meet within a day's drive decides how fast it can earn trust. The maps →
More where these came from. Start a conversation →