Back

John Deere buys Orchard AI, whose cameras ride on farm vehicles and count the fruit on every tree

Orchard AI's FruitScope Vision camera unit, with two lamps and a pair of camera modules, mounted on an arm beside the bare vines of a vineyard
Image: Orchard AI

John Deere has bought Orchard AI, a San Francisco startup whose AI cameras ride on tractors and other farm vehicles and record each tree, vine or plant they pass: how much fruit it carries, and the size and color of that fruit. Deere says pairing that data with its own equipment can help fruit, nut and grape growers decide where to spray, thin and harvest. Deere has not said what it paid.

John Deere said on Nov. 19 that it has bought Orchard Robotics, a San Francisco company founded in 2022 that does business as Orchard AI. The startup makes tools that scan crops plant by plant for growers of high-value crops such as apples, wine grapes, cherries, almonds and citrus. Deere says the deal joins Orchard AI's detailed data on each plant with Deere's own farm equipment and technology.

Orchard AI's cameras mount on tractors, ATVs and other farm vehicles that growers already drive, and capture every tree, vine or plant as the vehicle rolls past at 1 to 12 miles an hour, Orchard says. On each pass, Deere says, the system looks at hundreds of images of every tree. Its AI software turns them into fruit counts, fruit size and color, measurements of the leafy canopy and estimates of how big the harvest will be. Orchard says results are ready the next morning, and that its estimates of harvest size and fruit size are at least 95% accurate, a figure backed so far only by Orchard's own results and the customer accounts it publishes. Orchard says it has put hundreds of its systems to work with growers in the U.S. and beyond, and it sells the service as a yearly subscription priced by the acre, with no hardware to buy.

Deere's pitch is timing. For these growers, it says, a season's work can hang on decisions made in narrow windows, and a missed spray or a harvest at the wrong moment can hurt the fruit's quality, the size of the harvest and the season's profits. In Deere's example, a grower can find the parts of an orchard with more blossoms and get a plan for where to spray, sent to a phone or to a display in the tractor's cab. Orchard says its maps have already helped one large fruit grower thin blossoms more efficiently.

Justin Rose, who leads Deere's small agriculture and turf business, said Deere wants to keep the startup's speed and its close ties with growers while giving it Deere's reach. Orchard AI's chief executive, Charlie Wu, said that together with Deere, the aim is to bring the benefits of its AI crop insights to every grower in the world. Orchard's own site looks further ahead, to a day when self-driving farm machines act on its recommendations.

The price was not disclosed. Neither company has said whether Orchard AI will keep offering its service to growers who use other brands of equipment, or how far its scans will be built into Deere's machines beyond that example.