Bonsai Robotics AI aims to slash farm costs

California-based Bonsai Robotics is deploying an artificial intelligence vision system designed to work across multiple specialty crops, from almonds to strawberries. The company's cofounder and CEO, Tyler Niday, told AgFunderNews the goal is to drive down the overall cost of farm operations by combining autonomy with lower-cost, more flexible machinery.

Bonsai's core technology is a vision-based autonomy stack that it claims can turn standard 2D camera images into a scaled 3D understanding of a farm environment. This allows the same underlying software system to operate in orchards, vineyards, and berry fields without developers rewriting thousands of lines of code for each new crop or machine.

From 2D images to a learned 3D world

Historically, agricultural robotics required writing vast amounts of specific code for each task. Niday described this as hundreds of thousands of "if-then" statements tailored to a single crop. Bonsai's approach uses a learned AI model trained on extensive farm data.

The company has deployed 400 units and collected data across a million acres of specialty crops. It uses this data to train what it calls a foundation model. This model builds multiple 3D understandings from a 2D camera feed, such as elevation maps and semantic occupancy maps that identify trees, ground, or machinery.

"This one model can really go to all these different crops very quickly just from a 2D camera," Niday said. He contrasted this with traditional methods that rely on hardware like lidar, which can be blinded by dust. Because Bonsai's system is a learned model, its camera system can infer what lies beyond visual obstructions.

Expanding scope with new machinery

Bonsai started by focusing on challenging environments like almond orchards in Australia, where GPS is unreliable. Having proven its vision stack there, the company has expanded. Its AI model now works in strawberries, vineyards, apples, table grapes, and citrus.

Following its acquisition of Farm-ng, Bonsai is building its own robots alongside retrofitting existing equipment. Its new, larger Amiga platforms are designed for tasks like spraying, hauling, and lifting. Revenue is currently split roughly half and half between retrofitting original equipment manufacturer (OEM) machines and selling its own Amiga robots.

The company has sold out of its new hybrid-electric Amiga Max platforms for this year. Of its total 400 units sold, about 75 are OEM retrofits, with the rest being various Amiga models.

The economics of flexibility

Niday argues that flexibility is important for making the economics of autonomous farm machinery viable. The cost of an autonomy subscription on a standard tractor alone is often prohibitive. The value comes from creating multi-purpose machines that are not idle for most of the year.

He gave the example of an Orchard Machinery Corporation AR-500 shuttle truck used in almonds. The $200,000 machine historically sat parked for much of the year. By turning it into a flexible, autonomous tractor capable of multiple tasks, its utility and value increase dramatically.

"You need to go past just a labor replacement and find other ways to drop the cost," Niday stated. He highlighted that Bonsai's own Amiga sprayer can burn about three gallons of diesel per day compared to 30 gallons for a traditional machine, representing significant operational savings.

Testing and community connections

Bonsai has use Reservoir Farms, a testing and demonstration hub in Salinas, to accelerate development. Niday said this access was invaluable, as it initially took him seven months to find a single orchard for early testing. Reservoir provided immediate testing capabilities across various crops and help rapid connections with growers for demonstrations.

This environment helped the company ensure it was building usable products. The ability to quickly gather decision-makers from local agricultural valleys to watch demos has been a key enabler for commercial progress and product refinement.