Bonsai Robotics

The Tractor That Doesn't Need a Driver

It is 2 a.m. in a California almond orchard. Dust hangs in the air from hours of sweeping. A Flory Super V machine runs another pass between the rows - no headlights needed, no operator in the cab. It knows exactly where the trees are, where the ground dips, and when the row ends. Not because a satellite told it so. Because cameras did.

That machine is running on Bonsai Robotics software. And that scene - the one that sounds like science fiction until you've stood in a dark orchard and watched it happen - is now a commercial reality.

Bonsai Robotics builds vision-based autonomy systems for off-road farm equipment. Their Bonsai Intelligence Platform combines embedded AI with retrofittable hardware kits and a cloud-based fleet management layer. It works in orchards, vineyards, and row crop fields. It works without GPS, without cell signal, and without a human hand on the wheel.

Agriculture Has a Labor Problem That Isn't Going Away

American agriculture runs on seasonal labor. It always has. Harvest crews follow the crops north as summer moves through California, Oregon, Washington. Pruning gangs work the winter vineyards. Thinning teams hit the stone fruit orchards in spring. The whole system depends on enough people showing up.

They are showing up less. The average age of a US farm worker is rising. Immigration policy has tightened. Labor costs have climbed faster than commodity prices for most specialty crops. Nut and fruit growers - who can't easily mechanize the same way row-crop farmers did - are caught in a squeeze that only gets worse with each season.

The obvious solution - robots - has been tried. GPS-guided equipment works fine in open fields. Add trees, uneven terrain, dense canopy, or a patch of dust from a harvester running ahead of you, and most autonomy systems either slow down or shut off. The farms that need automation most are the ones that standard tech handles least well.

Two John Deere Veterans Who Knew Where the Map Ended

Tyler Niday and Ugur Oezdemir did not build Bonsai Robotics because the agriculture robotics problem looked easy. They built it because they had spent years inside the institutions trying to solve it - and they knew exactly why those institutions kept running into walls.

Niday had been an engineering leader at Blue River Technology (acquired by John Deere in 2017) and then inside Deere itself. Oezdemir had worked on vision-based autonomy at Airbus before moving to Blue River and Deere. Both understood that agricultural environments don't behave like highways or warehouses. Rows aren't perfectly straight. Canopies block signal. Dust from the machine ahead of you erases visual landmarks in seconds.

Their bet: that the right approach to agricultural autonomy was vision-first, not GPS-first. Build a system that sees the world the way a skilled operator does - with cameras and learned context - rather than one that treats GPS coordinates as ground truth. Strip away the infrastructure dependencies. Then train it relentlessly on real field data until it handles the exceptions, not just the ideal runs.

They founded Bonsai in 2022. By October 2023, they had a $13.5 million Seed round and a handful of commercial deployments. By January 2025, they had a $15 million Series A. And by mid-2025, they had acquired farm-ng, a modular electric robot maker out of Santa Cruz, to get into the hardware game in earnest.

One Platform. No GPS Required. Works in the Dark.

The Bonsai Intelligence Platform is two things at once: an embedded autonomy layer that runs on the machine, and a cloud-based operations system that manages fleets across an entire farm. The onboard component, Bonsai Autonomy, handles navigation, path planning, and real-time decision-making. The cloud component, Bonsai Pilot, handles job planning, telemetry, real-time notifications, and emissions tracking.

VisionSteer, their retrofit hardware kit, is the physical entry point for OEM equipment manufacturers who want to add drive-by-wire capability to existing machines. It installs on legacy equipment, connects to the Bonsai software stack, and gives that equipment the ability to operate without a driver. It requires no cellular connectivity and no GPS - which is the whole point.

After the farm-ng acquisition in July 2025, Bonsai also builds their own robots. The Amiga lineup - Flex, Max, and Trax - are modular electric platforms suited for everything from vineyard scouting to full-scale weeding. The Amiga Max won the 2026 Top 10 New Product Award at World Ag Expo, which is either a sign of good product design or a very friendly judging panel. Probably both.

What Actually Happens When You Deploy This Thing

Bonsai's most documented commercial deployment is with Flory Industries, the California equipment manufacturer that makes sweepers for nut orchards. Sweeping - running a large brush head down the rows to windrow fallen nuts for pickup - is one of those jobs that requires sustained, precise movement through tight, uneven terrain. Do it wrong and you miss nuts. Do it too fast and you throw them out of the row.

After integrating VisionSteer and the Bonsai Intelligence Platform into their Super V sweeper, Flory reported 45% lower operating expenses and 45% faster job completion. The system runs at night without headlights, navigates dust and debris, and handles elevation changes across orchard floors. It does this without a driver in the cab.

Beyond Flory, Bonsai has built integrations with Orchard Machinery Corporation for harvest equipment, Carbon Robotics for autonomous LaserWeeder deployment, OnTarget Spray Systems for electrostatic sprayers (which achieve up to 80% water savings and two to three times more acres per day), Interlink Sprayers for vineyard trellis crops, and WEED-IT for autonomous weeding. The pattern is consistent: integrate the Bonsai stack with specialized equipment, let the vision system handle navigation, and the specialized equipment does more of what it was designed to do.

Progress Belongs to Everyone, Not Just Farms That Can Afford a Full Robotics Team

Bonsai's stated mission is to make autonomy "simple, affordable, and ready for every off-road job." That second word is where the real ambition lives. Agricultural autonomy has existed in some form since GPS guidance systems became widespread in the 1990s. But those systems assumed flat terrain, open sky, and reasonable signal quality. They were built for the Iowa corn belt, not the Fresno almond grove.

The company built its platform to meet farmers where their fields actually are - not where a satellite signal is clearest. The VisionSteer retrofit approach matters here. Rather than requiring growers to buy entirely new equipment, Bonsai designed their system to integrate with existing machines. An OEM puts VisionSteer into a product they already manufacture, and a grower who trusted that OEM can now deploy autonomy without starting over.

After acquiring farm-ng, Bonsai expanded that logic into electric robot hardware. The Amiga lineup targets operations - particularly research institutions, specialty crop farms, and smaller commercial operations - that can't afford large-scale fleet contracts but still need capable autonomous platforms. The USDA and UC ANR both work with Amiga in research contexts, which gives Bonsai a pipeline of academic data to improve their models alongside commercial field data.

The Farm Labor Shortage Is Not Going to Fix Itself

The broader argument for what Bonsai is building is not really about robots. It is about what happens to food production as the workforce that has sustained specialty crop agriculture ages out and isn't replaced. The US Bureau of Labor Statistics has been tracking declining agricultural labor availability for over a decade. California's almond crop alone requires more than 100,000 seasonal workers during harvest. Those workers are harder to find and more expensive every year.

Autonomous equipment doesn't fully replace labor - at least not yet, and probably not in most operations for the next decade. What it does is extend the capacity of the workers who remain, reduce the hours those workers have to spend on repetitive machine-minding tasks, and allow operations to continue at night and in conditions that are too hazardous for unassisted human operators.

The same capabilities that reduce operating costs for an almond sweeper also improve worker safety for orchard spray operations, where chemical exposure is a documented occupational hazard. And the fleet management and telemetry data that Bonsai generates gives growers insight into their operations that they have never had before - every pass, every anomaly, every job logged and available for analysis.

Who Bet on Them and When

Back to That Dark Orchard

Back to that California almond orchard at 2 a.m. The Flory sweeper is still running its rows. What changed is not the machine - it's what the machine knows. It has seen this orchard before, in every light condition, at every speed, with every kind of ground variation. It has seen 750,000 acres worth of orchards in aggregate, distilled into a model that runs in real time on embedded hardware under the hood.

The farmer is not in the cab. They might be asleep. Or they might be checking the Bonsai Pilot dashboard on their phone - watching the telemetry, confirming the job is running clean, maybe looking at the emissions data that their sustainability report will require next quarter.

That is what Bonsai Robotics sells. Not a robot. Not a software subscription. The ability to run your operation at 2 a.m. without anyone running it. For an industry that has been asking for that capability for thirty years and getting incremental GPS updates as the answer, it is a meaningful step in a different direction.

Whether Bonsai scales that step into a platform that reaches the full diversity of American agriculture - from the specialty crop orchards they serve today to the bedded row crops they are targeting with the Amiga lineup - is the question the next few funding rounds will answer.

See It in Action