Comment

Horticulture ยท Automation guide

Horticulture Automation: 7 Systems, Costs and Labor Data

Labor was 42% of production expenses on US greenhouse and nursery operations in 2022, against 12% for all farms (USDA ERS). Seven automated systems for greenhouses, nurseries and fruit and vegetable farms, from climate control to harvest robots, with costs and limits.

Authored by Farmonaut Research & Content Team
Reviewed and verified by Farmonaut Agricultural Experts Team
Updated 12 November 2026 ยท 15 min read
Sentinel-2 Satellite View Of Vegetable And Salad Fields In The Salinas Valley, California
Sentinel-2 view of Vegetable and salad fields in the Salinas Valley, California. Contains modified Copernicus Sentinel data, processed in Google Earth Engine.

Key takeaways

  • Wages and contract labor were 42% of production expenses for US greenhouse and nursery operations and 40% for fruit and tree nut farms, against 12% for all farms (USDA ERS, 2022 Census of Agriculture).
  • Real wages for US nonsupervisory crop and livestock workers rose from $12.05 to $18.12 an hour between 1990 and 2024, in 2024 dollars (USDA ERS).
  • A step controller holds temperature within about ยฝ to 1 ยฐF, against 4 to 8 ยฐF for a mechanical thermostat; each 1 ยฐF of over-heating adds about 3% to energy use (Farm Energy, Extension).
  • In 2019, about 61% of US controlled-environment tomatoes, 67% of cucumbers and 66% of lettuce were grown hydroponically (USDA ERS EIB-264).
  • Harvest and pollination robots work first in protected crops grown in a uniform way; climate and irrigation control are the usual first step.
In this article
  1. What horticulture automation covers
  2. Why growers automate: labor costs and wages
  3. Control systems in horticulture: thermostats to climate computers
  4. 7 automated systems in horticulture and what each does
  5. Automated quality control in horticulture
  6. Soilless growing: where controlled-environment automation is furthest along
  7. Costs and payback of horticulture automation
  8. How to choose a first automation project: 6 steps
  9. Challenges and limits of automation in horticulture
  10. Where horticulture automation is heading
  11. FAQ

Horticulture automation means using controllers, sensors, software and robots to run greenhouse, nursery, orchard and vegetable work that people once did by hand. The biggest driver is labor: wages and contract labor were 42% of production expenses on US greenhouse and nursery operations in the 2022 Census of Agriculture, against 12% for all farms, according to the USDA Economic Research Service (ERS). Most growers start with climate control and irrigation, not with robots.

This guide explains the seven main types of automated systems in horticulture, what each one does, who it suits and where the limits are. It adds USDA labor and wage charts, data on soilless growing, a payback calculator you can run with your own numbers, and a step-by-step plan for choosing a first project. The examples come from the US, the UK, the Netherlands and Australia.

What horticulture automation covers

Horticulture is the growing of fruit, vegetables, nuts, flowers, nursery stock and potted plants. These crops are high in value per acre and need a lot of hand work: planting, pruning, training, picking, grading and packing. Automation in horticulture is any system that takes over part of that work or makes a decision for the grower.

It helps to split the field into three layers:

  • Control. Devices that switch heating, vents, fans, shade screens, lights and valves to hold set points. This is the oldest and most common layer.
  • Sensing and data. Sensors, cameras, drones and satellite images that measure the crop and its environment, plus software that turns readings into alerts or decisions.
  • Robotics. Machines that move and handle plants or produce: weeders, harvesters, pollinators, tray movers and packing lines.

The layers build on each other. A harvest robot needs a crop grown in a uniform way. An AI climate strategy needs a climate computer that can carry out its decisions. That is why the order in which growers invest matters as much as the technology itself.

Automation also differs by setting. In a greenhouse or vertical farm the space is fixed, the rows are straight and the light and climate can be controlled. Outdoors, robots face mud, wind, sun glare and plants of uneven size. So most working systems today sit in protected cropping, nurseries and packhouses, while field robots are spreading more slowly, first in high-value vegetables.

Why growers automate: labor costs and wages

Labor is the single largest cost for many horticultural businesses. The ERS farm labor page, using the 2022 Census of Agriculture, puts wages, salaries and contract labor at 42% of production expenses for greenhouse and nursery operations and 40% for fruit and tree nut farms. For all US farms the figure is 12%. A grower who cuts labor hours by a tenth saves far more in a nursery than in a grain operation.

Chart

Labor’s share of production expenses

Wages, salaries and contract labor as a percent of production expenses, US farms, 2022

0%12%25%38%50%42%Greenhouse andnursery40%Fruit and treenuts12%All farmsSource: USDA Economic Research Service, Farm Labor topic page, using the 2022 Census of Agriculture.
0%12%25%38%50%Greenhouse and nursery42%Fruit and tree nuts40%All farms12%Source: USDA Economic Research Service, Farm Labor topic page,using the 2022 Census of Agriculture.
Labor weighs more than three times as heavily on greenhouse, nursery and fruit farms as on the average US farm.

Wages are also rising. ERS reports that real wages for nonsupervisory crop and livestock workers grew by 1.2% a year from 1990 to 2024, and by 1.9% a year over the last ten years of that period. The average real farm wage went from $12.05 to $18.12 an hour (in 2024 dollars), while the nonfarm wage went from $23.87 to $30.13. Farm pay is still lower, but the gap is slowly closing.

Finding workers is harder too. ERS notes that positions certified under the H-2A temporary farm worker program rose from just over 48,000 in fiscal year 2005 to around 385,000 in fiscal year 2024. Employers in that program must pay at least the Adverse Effect Wage Rate set by the US Department of Labor. For fiscal year 2025 that minimum ranged from $14.83 an hour in Arkansas, Louisiana and Mississippi to $19.97 in California, $20.08 in Hawaii and $22.23 in the District of Columbia, according to the same ERS page. Employers also pay for housing and travel. Rules change often, so check the current Department of Labor rates and get professional advice before you budget for seasonal labor.

ERS also notes that fewer young immigrants are entering US farm work, so the hired farm workforce is aging. Put together, hand work is expensive and harder to staff, and growers want machines for the most repetitive jobs.

Chart

Average real hourly wage, farm vs nonfarm

US$ per hour in 2024 dollars, private-sector nonsupervisory workers

19902024$0.00$10.00$20.00$30.00$40.00Farm workers$12.05$18.12Nonfarm workers$23.87$30.13Source: USDA Economic Research Service, Farm Labor topic page (NASS Farm Labor Survey and BLS data).
19902024$0.00$10.00$20.00$30.00$40.00Farm workers$12.05$18.12Nonfarm workers$23.87$30.13Source: USDA Economic Research Service, Farm Labor topic page(NASS Farm Labor Survey and BLS data).
The farm wage rose from about half of the nonfarm wage in 1990 to about 60% in 2024.

Control systems in horticulture: thermostats to climate computers

Most automation in horticulture is not a robot. It is the control system that runs a greenhouse or nursery. The textbook Greenhouse Management (Apex Publishers) describes control systems that manage humidity and temperature and can also run shade systems, extra lighting, carbon dioxide, irrigation valves, and the pH and electrical conductivity of the nutrient solution. Growers usually move through four levels:

  1. Thermostats and timers. Each heater, fan or valve has its own switch. Cheap, but devices can work against each other, for example a heater running while a vent is open.
  2. Step controllers. One electronic unit replaces several thermostats and runs heating and cooling in stages. It can also handle vents, alarms, lighting and irrigation.
  3. Zone computers. A controller that manages several growing zones, each with its own set points.
  4. Integrated climate computers. One system reads indoor and outdoor sensors and coordinates heating, vents, shade and thermal curtains, carbon dioxide, fogging and irrigation under shared logic.

The jump from a thermostat to a step controller is often the cheapest win in the whole field. Farm Energy (Extension) explains that a step controller holds temperature within about ยฝ to 1 ยฐF of the set point, against 4 to 8 ยฐF for a mechanical thermostat. It adds that each 1 ยฐF of heating above the set point raises energy use by about 3%, which is why the tighter control can pay for the unit. Most step controllers also have day, night and DIF settings, and the sensor can sit among the plants up to 100 feet from the control box.

An integrated system is where the other automation in this article connects. Irrigation, lighting and sensor data feed the same controller, so the grower sets targets rather than switching equipment by hand. For a small house, a step controller is often enough. Larger ranges with several crops gain more from an integrated computer, because each zone can run its own climate strategy.

7 automated systems in horticulture and what each does

The seven systems below are listed roughly from the most widely used to the most experimental. Each has a “Best for” line so you can see quickly whether it fits your operation.

1. Climate and fertigation control

Best for: any greenhouse, polytunnel range or nursery with heating, vents or drip irrigation.

This is the foundation described above. Climate computers hold temperature, humidity, light and carbon dioxide. Fertigation units mix fertilizer into the irrigation water and dose it by time, by light sum or by sensor readings, while checking pH and electrical conductivity. The payoff is steady growth, less energy waste and fewer late-night checks. A grower can see alarms on a phone instead of walking the houses.

The limit is that control is only as good as the sensors and the set points. A badly placed sensor in direct sun or next to a heater will give the controller wrong readings. Calibrate sensors on a schedule and keep a manual override for every key device.

2. Precision irrigation and soil moisture sensing

Best for: orchards, vineyards, berries, vegetables and container nurseries where water or pumping cost is high.

Drip and micro-sprinkler systems deliver water and nutrients near the roots. Adding soil moisture probes, weather data and automatic valves lets the system water when the crop needs it rather than on a fixed timer. Some systems also adjust for evaporation each day. For the design side, see our guide to irrigation system installation for agriculture.

Point sensors show one spot. On larger blocks, growers pair them with maps from drones or satellites that show which parts of a field are stressed, then check those spots on the ground. Satellite vegetation maps such as those from the Farmonaut Satellite + AI Farm Monitoring System can flag uneven blocks across many fields at once, which is useful for orchards and outdoor vegetables.

3. Automated weeding and thinning

Best for: large vegetable and salad growers with high hand-weeding costs.

Weeding robots use cameras and computer vision to tell crop plants from weeds, then remove weeds with blades, lasers, hot water or tiny doses of spray. Some also thin seedlings to the right spacing. These machines have moved from trials to commercial use in high-value vegetables.

The business side is still settling. FarmWise built the Vulcan intra-row weeder and thinner. Taylor Farms, a salad and fresh food producer, announced that it had acquired FarmWise’s business and would support Vulcan customers through the change. The lesson for buyers: check who will service a machine in five years, not only what it does today.

4. Robotic harvesting

Best for: protected crops grown on trellises or tables, such as sweet peppers, tomatoes, cucumbers and strawberries.

Harvest robots need to find the fruit, judge if it is ripe, reach it without hitting leaves and stems, and detach it without bruising. A field trial of a sweet pepper robot named Harvey, reported by Lehnert, McCool and Perez on arXiv (2017), combined vision for crop detection, a grasp selection method and a new end-effector. The team focused on protected cropping, where plants grow on flat trellises, to reduce blocked views and simplify movement. They listed detecting the fruit stalk and speeding up each picking cycle as the next problems to solve.

That focus shows where harvest robots work first: crops grown in a uniform way, under cover, with fruit that is easy to see. Crops trained for the robot, such as dwarf tomatoes, make the task easier. Robots that pick apples or table grapes in open orchards face more variation and are mostly still at the trial or early commercial stage.

5. Robotic pollination

Best for: greenhouses and vertical farms where bees or wind cannot reach flowers easily.

Closed growing spaces limit natural pollination. A review of robot-based pollinators in greenhouse farming (Singh, Seneviratne and Hussain, arXiv, 2024) notes that the decline of bee and wind pollination in controlled environments has raised interest in robotic options. These systems use air jets, vibration, brushes or sprays to move pollen, guided by cameras that find open flowers.

Most of this work is still at the research or pilot stage. Many greenhouse tomato growers today use bumblebee hives or hand-held vibrating tools. A robotic pollinator makes the most sense where bees perform poorly, such as vertical farms with no natural light, or where flowering is very uniform.

6. AI crop monitoring and decision support

Best for: growers with many blocks, sites or houses who want earlier warnings of stress, pests and disease.

AI in horticulture takes data from sensors, cameras, drones and satellites and turns it into alerts and advice. Examples include yield forecasts from fruit counts, disease risk models that use leaf wetness and temperature, and vegetation maps that show weak areas in a field. The goal is to spend scouting time where the problems are.

The most striking tests come from the Netherlands. In the Autonomous Greenhouse Challenge run by Wageningen University & Research (WUR), teams let algorithms fully control their greenhouses, deciding on lighting, heating, carbon dioxide, water, plant density and harvest timing. The fourth edition grew dwarf tomatoes, and the winner was judged mainly on net profit. WUR chose that crop because it needs a single harvest, can be picked by robots and grows uniformly.

AI advice still needs a person to check it. Models trained on one crop, climate or variety may not transfer to another. Treat early outputs as a prompt to scout, not as an order to spray or irrigate.

7. Modular and open-source automation

Best for: small vertical farms, research growers and operations that want to build or extend their own systems.

Not all automation comes as a turnkey package. MACARONS, the Modular Automated Crop Array Online System, is an open hardware system that moves trays of plants between stations in a vertical farm. The paper in the University of Cambridge repository reports that it moves trays up to 1,060 mm by 630 mm and 12.5 kg, and that one grow unit cost US$144.96 to build, or US$128.85 per square meter of grow area. The designs are public, as described in the arXiv version of the MACARONS paper.

Open designs lower the cost of entry and let growers repair and change their own kit. The trade-off is that someone on the team must have the skills to build, program and maintain it.

Automated quality control in horticulture

Automated quality control means checking produce, plants or growing conditions with sensors and cameras instead of the human eye alone. It shows up at three points in the chain.

  • In the crop. Cameras, drones and satellite images track plant vigor, flowering and fruit set so growers can fix problems before they hit quality. Climate computers log every set point, which gives a record of how the crop was grown.
  • At harvest. Harvest robots judge ripeness before picking, which aims for more regular, selective harvests. The sweet pepper trial above lists better crop quality and scheduling as goals of robotic picking.
  • In the packhouse. Optical graders photograph each fruit or plant and sort by size, color and visible defects. Nurseries use similar camera systems to grade potted plants by height and fullness before dispatch.

Good quality control also needs records. Lot codes, harvest dates and block maps let a grower trace a problem back to its source. Pair any grading line with clear written grade standards and a person who checks a sample of each run, because cameras can drift out of calibration just like sensors.

A walk-through of monitoring fields with satellite crop health maps on the Farmonaut web app.

โ–ถ A walk-through of monitoring fields with satellite crop health maps on the Farmonaut web app. ยท 8:43

Soilless growing: where controlled-environment automation is furthest along

Automation is easiest where the whole growing system is engineered. Soilless or hydroponic growing is a good guide to how far that has gone. A USDA ERS report on controlled environment agriculture (EIB-264, January 2024) found that in 2019 about 61% of US tomatoes, 67% of cucumbers and 66% of lettuce grown in controlled environments came from hydroponic systems. The share was lower for other crops: 23% for fresh-cut herbs, 16% for strawberries and 3% for peppers.

Chart

Share of US controlled-environment output grown hydroponically

Percent of each crop’s controlled-environment production volume, 2019

0%20%40%60%80%Cucumbers67%Lettuce66%Tomatoes61%Fresh-cut herbs23%Strawberries16%Peppers3%Source: USDA ERS, Trends, Insights, and Future Prospects for Production in Controlled Environment Agriculture and AgrivoltaicsSystems (EIB-264, January 2024), from the 2019 Census of Horticultural Specialties.
0%20%40%60%80%Cucumbers67%Lettuce66%Tomatoes61%Fresh-cut herbs23%Strawberries16%Peppers3%Source: USDA ERS, Trends, Insights, and Future Prospects forProduction in Controlled Environment Agriculture andAgrivoltaics Systems (EIB-264, January 2024), from the 2019Census of Horticultural Specialties.
Tomatoes, cucumbers and lettuce are mostly soilless; peppers and strawberries much less so.

The same report says total food crop volume grown in hydroponic systems rose from about 3.7 million cwt in 2009 to 4.2 million cwt in 2019. Tomatoes, lettuce and cucumbers make up roughly 60% to 70% of crops grown in controlled environments. These crops are where automated climate, fertigation, transport and harvest tools have the most to work with.

Costs and payback of horticulture automation

Automation costs range widely, from a single step controller to a fully automated facility. ERS cites startup costs of $150 to $400 per square foot for a vertical farm, compared with $50 to $150 per square foot for a greenhouse. It adds that labor costs in advanced controlled environment systems are higher than in conventional fruit and vegetable production, because they need technical staff, and that robotics could help bring those costs down but is costly to fully integrate.

Field automation faces the same test. In a Purdue University Center for Commercial Agriculture podcast on autonomous farm equipment (February 2026), researchers said that under current performance and cost levels most corn and soybean farms in their model were not yet in the economic range for large autonomous machines. They named labor wages and constraints as the tipping point. That study covered row crops, not horticulture, but the logic carries over: automation pays first where labor is scarce or expensive.

A simple payback check helps you sort ideas before you talk to dealers. Use the calculator below with your own numbers. The defaults are examples set near break-even, not benchmarks.

Try the numbers for your own farm below.

Calculator

How long would an automation upgrade take to pay back?

Enter your own equipment cost, hours saved and wage cost.

US$

hours

US$/hour

US$/year

years
3.3years to pay back (a very large number means it never pays back)
  • Labor cost saved per year: $18,120 US$
  • Net saving per year: $12,120 US$
  • Net gain over working life: $8,480 US$

Example values only, set near break-even; they are not benchmarks or forecasts. The example wage of $18.12 is the 2024 average real wage for US nonsupervisory crop and livestock workers (USDA ERS); use your own full cost per hour including payroll taxes, housing or benefits. Running cost should include software fees, power, maintenance and any extra staff time. The calculator ignores financing, tax and resale value. Get professional advice before a large purchase.

How to choose a first automation project: 6 steps

  1. Log where the hours go. Track labor by task for a full season: planting, spacing, watering, scouting, picking, grading and packing. The biggest block of repetitive hours is your first target.
  2. Fix control before robots. If you still run on thermostats and timers, a step controller or climate computer is usually the lowest-risk start. Robots work best when the crop is uniform.
  3. Check the crop and layout fit. Ask vendors which crops, row widths, trellis types and container sizes their system has been tested on. Ask to see it working on a farm like yours.
  4. Run the numbers. Include purchase price, installation, software fees, power, maintenance, training and the hours you will still need. Compare with the hours saved at your real wage cost.
  5. Plan service and data. Find out who repairs the machine, how fast, and what happens if the maker is sold or closes. Read the data terms: who owns your farm data and can you export it.
  6. Start small and measure. Trial one house, block or line for a season. Record hours, yield, grade-out and energy use before and after, then decide on the next step.

Challenges and limits of automation in horticulture

  • High upfront cost. ERS names high startup and running costs among the main barriers for highly automated indoor farms, and robotics as costly to integrate. Smaller growers may find shared services or leasing easier than buying.
  • Crop and site variation. Robots struggle with uneven plants, hidden fruit and changing light. Outdoor systems face rain, dust and slopes.
  • Skills gap. Someone has to set up, calibrate and fix the equipment. Training is a real cost and should be in the budget.
  • Vendor risk. Young robotics firms are sold, merge or close. The FarmWise example shows why service plans matter.
  • Data and privacy. More sensors and cloud tools mean more farm data held by third parties. Check who can see and use it.
  • Safety. Moving machines and automatic valves need guards, stop buttons and clear rules for workers. Follow the maker’s manual and local workplace safety rules.

Where horticulture automation is heading

Three trends stand out in the sources above. First, control is moving from fixed set points to strategies set by software, as the WUR challenge shows. Second, crops and growing systems are being designed with machines in mind, such as dwarf tomatoes and trellised peppers. Third, open and modular hardware such as MACARONS is lowering the entry cost for small growers and researchers.

For most farm managers and consultants, the practical path is steady: tighter climate and irrigation control, better crop monitoring across blocks, then targeted robotics for the one or two tasks that eat the most hours. The data from each step makes the next one easier to judge.

How precision farming technology and AI are changing the way crops are monitored and managed.

โ–ถ How precision farming technology and AI are changing the way crops are monitored and managed. ยท 23:42

FAQ

What is horticulture automation?

It is the use of controllers, sensors, software and robots to run tasks in greenhouses, nurseries, orchards and vegetable farms. It covers climate and irrigation control, crop monitoring, weeding, harvesting, pollination, plant transport and grading.

Which horticultural tasks are easiest to automate?

Climate control, irrigation and fertigation are the most established. Plant transport, grading and weeding in vegetables are next. Harvesting and pollination by robot are still mostly limited to protected crops and trials.

Why is automation growing in horticulture?

Labor is a large and rising cost. ERS reports that wages and contract labor were 42% of production expenses for US greenhouse and nursery operations in 2022, and that H-2A positions certified rose to around 385,000 in fiscal year 2024.

How much does greenhouse automation cost?

It ranges widely. A step controller is a small purchase compared with a full climate computer or robot. For whole facilities, ERS cites startup costs of $50 to $150 per square foot for a greenhouse and $150 to $400 for a vertical farm. Get quotes for your own site.

Can small growers use automation?

Yes, usually by starting with a step controller, timed or sensor-driven irrigation and crop monitoring. Open hardware such as MACARONS shows lower-cost options for small vertical farms, though it needs in-house technical skills.

Related pages on our site: ornamental and sustainable horticulture trends, greenhouse technology in Canada, AI for fertilizer planning, farm carbon footprinting, produce traceability, farm fleet management, crop loan and insurance checks, satellite and weather API documentation, satellite data API, web and mobile app, large-farm admin tools.

Authored by: Farmonaut Research & Content Team

Reviewed and verified by: Farmonaut Agricultural Experts Team

Reviewed November 2026 against the USDA ERS Farm Labor page and EIB-264 report, Greenhouse Management (Apex), Farm Energy Extension, Wageningen University & Research, the University of Cambridge repository, Purdue University’s Center for Commercial Agriculture, Taylor Farms, and research papers on arXiv.

Farmonaut crop monitoring

See your own fields from space

Draw your blocks once and Farmonaut shows satellite vegetation maps for them through the season, on web, Android or iOS.

Open The Farmonaut Crop Monitoring Web AppGet The Farmonaut App On Google PlayGet The Farmonaut App On The App Store
Farmonaut Web App With A Satellite Crop Health Map

Sources

  1. USDA Economic Research Service, Farm Labor (topic page): ers.usda.gov
  2. USDA ERS, Trends, Insights, and Future Prospects for Production in Controlled Environment Agriculture and Agrivoltaics Systems, EIB-264 (January 2024): ers.usda.gov
  3. Apex Publishers, Greenhouse Management: Greenhouse Environmental Control Systems: greenhouse-management.com
  4. Farm Energy (Extension), Step Controllers Optimize Greenhouse Energy Use: farm-energy.extension.org
  5. Taylor Farms, Taylor Farms Announces Acquisition of FarmWise’s Business: taylorfarms.com
  6. Lehnert, McCool and Perez, Lessons Learnt from Field Trials of a Robotic Sweet Pepper Harvester (arXiv, 2017): arxiv.org
  7. Singh, Seneviratne and Hussain, A Comprehensive Review of Current Robot-Based Pollinators in Greenhouse Farming (arXiv, 2024): arxiv.org
  8. Wageningen University & Research, Winner of the 4th Autonomous Greenhouse Challenge announced (January 2025): wur.nl
  9. University of Cambridge repository, MACARONS: A Modular and Open-Sourced Automation System for Vertical Farming: repository.cam.ac.uk
  10. Wichitwechkarn and Fox, MACARONS (arXiv, 2022): arxiv.org
  11. Purdue University Center for Commercial Agriculture, When Will Autonomous Farm Equipment Actually Pay? (February 2026): ag.purdue.edu


Discussion

Questions, corrections or field experience? Share them below. First-time comments are reviewed before they appear, and your email is never published.

No comments yet. Be the first to add a question or your experience.

Leave a comment

Farmonaut Farmonaut Trusted by 200,000+ users and 100+ businesses 200,000+ users trust us Get started