The Harvesting Robots Being Built Today
Robotic harvesting is concentrated in high-value fruit, where hand-picking is expensive and labour is short. Agrobot of Spain builds the E-Series strawberry harvester with multiple independent picking arms and colour-vision ripeness sensing. Tortuga AgTech targets tabletop strawberries and table grapes, and Harvest CROO Robotics builds a field-strawberry machine that reported commercial-viability field trials in 2025.
Tree and soft fruit have their own builders. FFRobotics of Israel develops a multi-arm apple harvester, the UK company Dogtooth picks tabletop strawberries, Fieldwork Robotics works on raspberries and other soft fruit, and Organifarms of Germany targets greenhouse strawberries.
Two patterns are worth noting. Almost every machine is crop-specific rather than general-purpose, and most are at the field-trial or early-commercial stage rather than mass deployment, so a grower is usually joining a programme rather than buying off a shelf.
| Robot / maker | Crop | Setting | Status |
|---|---|---|---|
| Agrobot E-Series | Strawberries | Field and tabletop | Commercial, Spain-based |
| Tortuga AgTech | Strawberries, table grapes | Tabletop and greenhouse | Commercial deployments |
| Harvest CROO | Strawberries | Open field | Field trials to commercial |
| FFRobotics | Apples | Orchard | Development and trials |
| Dogtooth | Strawberries | Tabletop | Commercial trials, UK |
| Fieldwork Robotics | Raspberries, soft fruit | Field and tunnel | Development, UK |
How to Judge a Harvesting Robot
The number every vendor quotes is pick rate. It is the least useful figure, because a fast picker that bruises fruit or misses ripe berries destroys more value than the labour it replaced. Judge two things first: what fraction of ripe fruit the machine correctly identifies, and what fraction of what it picks arrives undamaged and saleable.
Ask for that damage-rate and detection evidence on your own variety and at your real canopy density, not on a clean demonstration row. Crop feasibility varies enormously: firm, uniform, single-plane fruit is tractable, while soft, hidden, or continuously-ripening crops remain hard because ripeness detection, reach, and gentle gripping compound.
The picking window is the constraint no factory robot faces. A machine down for three days during the ripe week can cost the crop, so peak-season response time, regional spares, and a retained manual fallback matter as much as capability, which is the season framing in the agricultural robots guide. Model that against your current labour cost and season length using the robotics ROI guide before comparing vendor quotes.
- Ask for damage rate and detection accuracy on your own variety.
- Require evidence at your real canopy density, not a demo row.
- Confirm peak-season response time and regional spares.
- Keep a manual or contract fallback for the first seasons.
- Model the benefit against a measured baseline before buying.
Field Versus Tabletop, and Assisted Picking
The clearest split among these robots is field versus controlled environment. Tabletop and greenhouse strawberries present fruit at a consistent height with even lighting, which is why Tortuga AgTech, Dogtooth, and Organifarms concentrate there. Open-field machines such as Harvest CROO face variable terrain, weather, and canopy, which is a harder problem.
Assisted harvesting is the underrated middle ground. A platform that carries, sorts, or transports produce while people pick can pay off even where fully autonomous picking cannot, and it is available sooner and at lower risk than a full harvester.
The International Federation of Robotics counts field robots separately from industrial installations, a reminder that this is a distinct and less mature market than factory automation, so treat vendor timelines with the caution the stage warrants.
Bottom Line
Harvesting robots today are crop-specific machines from a handful of named builders, mostly for strawberries and apples and mostly at trial or early-commercial stage. Judge them on damage rate against a fixed window.
Match a named builder to your crop, then ask for damage-rate evidence on your own variety before any pick-rate claim.
FAQs
What companies make harvesting robots?
Named builders include Agrobot, Tortuga AgTech, and Harvest CROO for strawberries, FFRobotics for apples, and Dogtooth, Fieldwork Robotics, and Organifarms for soft and greenhouse fruit. Most machines are crop-specific.
Can you buy a harvesting robot today?
For some crops, yes, though outright purchase is rare this early. Most machines are at field-trial or early-commercial stage, so a grower typically joins a vendor deployment or service programme rather than buying equipment outright. Financing structures such as a lease, a per-acre service fee, or a purchase option vary by builder, so ask each company directly which one they offer.
What is the biggest risk with a harvesting robot?
Damage rate and the fixed picking window. A machine that bruises produce or is down during the ripe week can cost more than the labour it replaced, so pick rate is the wrong headline.
Are field or greenhouse harvesting robots more advanced?
Controlled-environment machines are further along, because tabletop and greenhouse fruit is presented at consistent height and lighting. Open-field harvesting faces variable terrain, weather, and canopy.
What ongoing maintenance does a harvesting robot need during the season?
A harvesting robot needs routine in-season upkeep: cleaning camera and vision sensors after each shift, checking picking arms and end effectors for wear, and reverifying ripeness-detection calibration between fields or varieties. Ask each builder for their season service plan and regional parts support before the picking window opens.
Is it safe for farmworkers to work alongside a harvesting robot in the field?
Farmworkers can generally share a row with a harvesting robot, but its picking arms move automatically to reach fruit, so keep the manufacturer's specified clearance and follow its posted safety procedures. OSHA notes that robotics incidents often happen during setup, testing, and maintenance, which is exactly when workers are closest to the machine, so confirm the builder's lockout and clearance rules before the first shift.
Does farm insurance cover a harvesting robot if it damages crops or injures a worker?
Coverage depends on your policy and the vendor's liability terms, not one fixed rule, so confirm crop-damage, equipment-damage, and worker-injury coverage with your insurer before deployment. Get the manufacturer's liability terms in writing as part of the deployment agreement, since coverage for a machine this new is not automatic.
How much does a harvesting robot cost to buy or lease, and can a small farm afford one?
There is no fixed sticker price: cost is quote-driven and scales with farm size, crop type, and which financing structure a builder offers, whether that is a lease, a per-acre service fee, or a purchase option. A small farm should ask each vendor directly which structure they support, since a service or per-acre model can lower the entry cost compared with buying a machine outright.
How accurate is ripeness or fruit detection in practice?
There is no single industry-wide accuracy benchmark, because detection performance depends on crop, variety, canopy density, and lighting conditions in each field. That is why buyers should ask a vendor for damage-rate and detection evidence measured on their own variety and canopy, not a published percentage, before comparing machines.
How does a harvesting robot's cost compare to hiring seasonal labor?
The comparison depends on your crop, farm size, and current labour cost, and it should be modelled against a measured baseline rather than a vendor's pick-rate claim. Use the robot automation ROI guide to run that comparison against your own season length and labour cost.