[Global "Agri-Talk"] Why U.S. Strawberry Farmers Invested Directly in Robot Development

[Global "Agri-Talk": Capturing the Future of Agriculture] Professor Choi Da-eun, University of Florida, U.S.
 

In Florida, U.S., strawberry farms and distribution companies invested money directly in the development of a harvesting robot. Gary Wishnatzki of the strawberry cultivation and distribution company Wish Farms judged that a shortage of harvest labor threatened the sustainability of the industry, and in 2012 founded Harvest CROO Robotics with engineer Bob Pitzer. Seven investors from the strawberry industry participated in the $1 million raised as initial development funding.[Source: 1, click the number to open the link]

Rather than purchasing a completed agricultural machine, the farmers themselves took on the development risk of a robot whose success was uncertain. If an agricultural machinery company developed it, farmers could simply buy it, so why did strawberry farmers step forward as investors? The reason is that the commercialization of agricultural robots is not determined solely by the severity of labor shortages.

A niche market for companies, a matter of survival for farmers
To understand why strawberry farmers set out to develop a robot themselves, it is first necessary to look at where agricultural robotics companies direct their investments. Companies prioritize crops to which a single technology can be repeatedly sold across large areas and to many farms, rather than crops suffering the most severe labor shortages. For farmers, a strawberry harvesting robot is a technology essential to continuing production, but for companies it is a niche product with low expected sales compared with development costs and risks. This gap turned farmers from consumers of technology into participants in development and investors.

The market conditions under which companies move first
In U.S. agriculture, a representative sector where autonomous driving and AI-based precision work technologies have been most widely commercialized is large-scale field crops such as corn and soybeans. In the United States, corn is grown on about 36.4 million hectares (ha) every year, and soybean acreage in 2024 also reached about 35.2 million hectares. In other words, a vast market already exists in which a single automation technology can be applied to tens of millions of hectares of farmland.[Source: 2 · 3]

Field-crop farms are wide and open, and crops are grown at regular intervals and in regular directions. The farm machinery and work methods used by different farms are also relatively similar. Because corn and soybeans are harvested all at once when mature, there is no need to judge the ripeness of individual crops or pick them one by one without causing damage. GPS, cameras, and artificial intelligence functions can be added to existing tractors, sprayers, and combines, so there is little need to design a new robot for each farm.

This is also the background behind the rapid spread of John Deere’s See & Spray. This system uses cameras and artificial intelligence mounted on existing sprayers to distinguish crops from weeds and applies herbicide only where needed. According to John Deere, it was used on more than 2.02 million hectares in 2025 and reduced herbicide use by an average of about 50% compared with conventional broadcast spraying.[Source: 4 · 5] 
The same technology can be applied to multiple crops such as corn, soybeans, and cotton and across broad areas, and the savings in herbicide costs can be calculated immediately, making the return on investment clear.

Even expensive equipment becomes less costly per hectare if it is used repeatedly across large areas every year. Manufacturers can also spread research and development costs, production facilities, and the costs of building service networks across more units, and can make use of existing sales networks and maintenance personnel. Because the market is large, products can be standardized, and after-sales support systems are already in place, companies have ample reason to invest first.

In fruit and vegetable crops, commercialization begins with repetitive tasks
Fruit and vegetable farming has difficulty meeting the same scale and standardization conditions as field crops. However, when looking at robots that have already entered the market, they share the common feature of automating clearly defined and repetitive tasks first, rather than handling the entire crop.

Carbon Robotics’ LaserWeeder finds weeds in vegetable fields with cameras and removes them with lasers, and the company says that since its launch in 2022, more than 100 farms in North America, Europe, and Australia have owned and operated the equipment.[Source: 6] 
GUSS, an autonomous sprayer for orchards, also automated the task of moving along predetermined rows and spraying chemicals first, rather than harvesting fruit. GUSS, which began customer deliveries in 2019, was acquired by John Deere in 2025.[Source: 7] 

The technological and market barriers blocking harvesting robots
However, the moment a robot harvests fruit directly, the problem becomes far more complex. It must find only ripe fruit among leaves and stems, detach it without damaging the product, and transfer it. Missed fruit must be harvested again by people, and damaged fruit loses its market value, so every process must be performed accurately within a short period of time.

What makes commercialization even more difficult is the small and fragmented market. A strawberry harvesting robot cannot be used as-is for apples or tomatoes, and even within strawberries, varieties and cultivation methods differ by region and farm, requiring adjustments to design and settings. Because it is used only during a specific harvest season, its annual operating time is short, and production regions are scattered, making field support difficult as well. Ultimately, for large agricultural machinery companies, it is a high-risk market with low expected sales compared with development costs and a need for a separate maintenance system. Farmers’ needs are urgent, but this is why it is difficult for companies to make large-scale investments first.

Robot development changed by farmers’ participation
At Harvest CROO, the role of farmers did not stop at providing early funding. Their participation also influenced the direction of the robot’s design. A representative example is that it was designed to be used without significantly changing existing beds and cultivation methods.

The current Harvest CROO harvester is designed so that 16 harvesting robots operate independently and use cameras and artificial intelligence to judge the ripeness and condition of strawberries.[Source: 8] The company announced that in 2025 commercial farm trials in Florida, it confirmed performance at a level comparable to human harvesting work.[Source: 9] However, because these are results announced by the company, additional verification is needed to determine whether the same performance can be repeated across diverse farms, varieties, and multiple harvest seasons.

The questions left by farmers’ investment
The direct investment by U.S. strawberry farmers is an unusual attempt to fill an “automation gap” created by a small market size with the urgency of those on the ground. They are not remaining mere consumers, but are directly testing the commercialization potential of robots for fruit and vegetable crops within the flow of agricultural automation that had been centered on large-scale field crops.

The reason the Harvest CROO case is special is that, among various options, farmers undertook the most active form of intervention by directly shouldering the early capital risk. It is also evidence that large agricultural machinery companies are hesitating to enter the fruit and vegetable (horticultural crop) market.

However, beyond prototype development, the market barriers remain high: mass production, building a continuous maintenance system, and forming robot prices that farmers can afford. If left solely to the logic of large capital and profitability, even this technology born from the urgency of the field will inevitably face limits in the end. For this bold investment not to remain a one-off experiment but to bear fruit, the demand confirmed in the field must be connected with more active intervention from a larger industrial ecosystem, such as the production and maintenance networks of major agricultural machinery companies, and public R&D support from universities and governments.

This experiment that began in Florida strawberry fields raises important questions about how horticultural crop agriculture around the world, facing a labor cliff, should secure technology and bring it to market. Now that farmers who had been waiting for robots have become developers themselves, who indeed is responsible for completing this technology as a sustainable industry?

[Reference materials / Sources]
1) https://wishfarms.com/newsroom-item/harvest-croo-robotics-develops-strawberry-picker-latest-solution-agricultural-robotics/
2) https://www.ers.usda.gov/topics/crops/corn-and-other-feed-grains/feed-grains-sector-at-a-glance?
3) https://www.ers.usda.gov/topics/crops/soybeans-and-oil-crops/oil-crops-sector-at-a-glance?
4) John Deere Investor Day, NYSE 8 December 2025, Transcript
5)  https://investor.deere.com/home/default.aspx
6) https://carbonrobotics.com/laserweeder
7) https://www.deere.com/en-us/john-deere-news/john-deere-acquires-guss-automation
8) https://www.harvestcroorobotics.com/technology
9) https://www.harvestcroorobotics.com/press/harvest-croo-demonstrates-commercially-viable-automated-robotic-strawberry-harvesting