21 September 2026, New York, USA: Cornell University is leading a new US Department of Agriculture (USDA) funded research center that aims to bring autonomous robots into American apple and cherry orchards for nearly every task in the growing season, from pollinating blossoms to picking ripe fruit. Cornell’s College of Agriculture and Life Sciences announced the project on September 3, 2026, backed by a four year, $7.5 million grant from USDA’s Specialty Crop Research Initiative.
The new Center of Excellence for Orchard Robotics is co-led by Manoj Karkee, a professor in Cornell’s Department of Biological and Environmental Engineering, and Matthew Whiting, a tree fruit horticulture professor at Washington State University. It draws in five more universities, Oregon State, Penn State, Michigan State, Carnegie Mellon and Stanford, along with Israeli agricultural robotics maker Fresh Fruit Robotics and grower partner Washington Fruit and Produce Co. Cornell Cooperative Extension fruit specialist Mario Miranda Sazo is also part of the team, tasked with connecting the engineering work to growers through extension programs.
Most orchard robotics research to date has focused narrowly on one job, typically harvesting. This project is broader by design. The team wants machines that can handle pollination, fruit thinning, winter pruning, inter row weeding and harvesting, covering the labor intensive tasks that make up most of an orchard’s annual cost.
Building Machines That See and Feel
The robots depend on computer vision paired with neural networks, a form of artificial intelligence trained on large numbers of orchard images to recognize fruit against a cluttered background of leaves, branches and shifting light. Karkee said these neural network based tools have “revolutionized” what researchers can now do with image and data processing in agriculture. Picking or thinning fruit without bruising it requires more than good eyesight, so the team is also developing soft robotic end effectors, flexible gripping mechanisms built to handle delicate fruit the way a careful human hand would, rather than the rigid metal claws used in industrial robotics.
A separate part of the project is building digital twins, three dimensional computer models of real orchards that let researchers train and test robots and their AI systems virtually before sending machines into an actual field. This should cut down on the number of real world trial runs needed and let multiple university teams work from a shared, realistic simulation of the same orchard conditions.
Prototype versions of several components were already tested in orchards in the fall of 2025, and the new funding is meant to push those prototypes toward machines growers can eventually buy and operate day to day, rather than one off research demonstrations.
Rising labor costs are the immediate driver behind the project. At Washington Fruit and Produce Co, one of the grower partners, the share of total production costs going to labor climbed from about 45 percent to more than 60 percent over the past 15 years, according to Cornell’s announcement. Company representative Gilbert Plath said the rising costs are increasingly squeezing farmers rather than simply affecting workers’ pay. Apples contribute an estimated $21 billion a year to the US economy, and New York state, Cornell’s home base, ranks second nationally in apple production and first in the range of varieties grown, giving both researchers and growers there a direct stake in the outcome.
Avi Kahani, chief executive of Fresh Fruit Robotics, said one of the biggest obstacles in agricultural technology is not inventing a new capability but turning that invention into a dependable industrial product growers can rely on season after season. Pairing engineering schools with a commercial robotics company and a working orchard operator from the outset is meant to close that gap faster than a purely academic project would.
The grant runs for four years, so the center’s first full field seasons of testing are still ahead. Growers, investors and rival machinery makers will be watching how quickly the multi task approach, rather than the single purpose harvesters already sold by some companies, moves from Cornell’s test orchards into commercial sale.
Global Agriculture is an independent international media platform covering agri-business, policy, technology, and sustainability. For editorial collaborations, thought leadership, and strategic communications, write to pr@global-agriculture.com
