Agricultural Harvesting Robots: Which Crops, Which Robots, 2026

Agricultural robot deployment in 2026 does not look like a uniform wave of automation rolling across farming. It looks like a patchwork: some tasks fully automated at scale, others commercialising, others still years from reliable deployment, and some — grain harvesting — already mechanised decades ago in ways that make the "robot" framing inapplicable.

Spraying is done. DJI Agras drones operate at planetary scale across Asia, South America, and increasingly North America and Europe. Weeding is commercialising — Blue River See & Spray (John Deere) has treated millions of US acres; Naio Technologies is deployed across European market gardens. Selective fruit and vegetable harvesting is in early commercial deployment for specific crops: Agrobot has been commercially harvesting strawberries since 2019.

For the workforce context, see Will Robots Replace Farmers? and the Agricultural Field Worker profile on the Geppetto Jobs Index.

Where Agricultural Automation Actually Is in 2026

The taxonomy matters before the analysis:

Already mechanised (not a robot story): Wheat, corn, soy, rice — row crops harvested by combine harvesters. These crops have been mechanically harvested at scale since the mid-20th century. The combine harvester is not a robot; it is a mature agricultural machine. Automation in this segment is about precision guidance, yield mapping, and variable-rate input application — not harvesting robotics.

Deployed at scale: Aerial spraying (DJI Agras), precision weeding for row crops (Blue River See & Spray). These are not emerging technologies — they are operational commercial products generating measurable agronomic outcomes.

Early commercial deployment: Selective harvesting for specific crops (Agrobot for strawberries), automated transplanting and weeding for market garden operations (Naio Oz). Commercially deployed, limited scale, cost-versus-manual-labour still converging.

Demonstrator stage: Apple harvesting, grape harvesting, asparagus harvesting. Technically demonstrated by multiple vendors; not yet commercially deployed at scale with documented performance in production environments.

The structural driver across all categories is the same: seasonal agricultural labour shortage. Chronic underfilling of seasonal harvest positions is the primary commercial pressure driving investment in harvesting automation.

The Crop-by-Crop Automation Matrix

Not all crops automate at the same speed. Four variables determine whether a crop is a near-term or long-term automation target:

Delicacy of handling: Can the fruit or vegetable be grasped by a robot end-effector without bruising or damage? Strawberries and tomatoes are fragile. Wheat is not. Harder fruit and root vegetables tolerate more robust handling.

Uniformity of plant structure: Does the plant present the harvest target in a consistent, predictable location? Row crops are designed around machine harvesting geometry. Tree fruits grow in three-dimensional canopy structures with variable fruit positioning. Strawberries in raised-bed cultivation present fruit below the canopy in a relatively consistent plane.

Labour intensity: How many hours of hand labour does the crop require per hectare? High labour intensity crops — strawberries, asparagus, lettuce, picking tomatoes — face the greatest economic pressure from labour cost and shortage.

Crop value per hectare: Does the revenue per hectare justify the capital cost of robotic deployment? Strawberries at $30,000–60,000 per hectare revenue justify different capital investment than wheat at $1,000–2,000 per hectare.

Applying this matrix:

Spraying: The Done Deal

DJI Agras T50

The DJI Agras T50 is the current flagship of the world's dominant agricultural drone platform. DJI Agras drones have been operating at commercial scale in Asia since 2016. By 2024, DJI reported cumulative Agras fleet operations covering hundreds of millions of hectares globally.

The T50 carries a 40-litre spray tank and 50-kg payload capacity, covers approximately 20–40 hectares per hour depending on conditions, and uses terrain-following radar to maintain consistent spray height across uneven ground. It supports both spraying and broadcasting (solid material spreading). Operating autonomously from a flight plan, a single T50 with one operator can cover area in a day that would require a conventional ground sprayer days to complete on difficult terrain.

The ROI case for aerial spraying is documented and unambiguous in Asian markets. The penetration in China, where small-plot agriculture dominates, is particularly high — ground-based sprayers cannot economically service the fragmented plot structure that DJI's drones operate across efficiently.

Compare: DJI Agras T50 vs T25

Best for: Aerial application of pesticides, herbicides, fungicides, and foliar fertilisers across field crops, orchards, and difficult terrain. Any operation where ground-based sprayer access is limited or labour-intensive.

Weeding: Commercialising at Scale

Blue River Technology See & Spray

See & Spray is a precision weeding system developed by Blue River Technology, acquired by John Deere in 2017 for $305 million. The system uses computer vision to distinguish crop plants from weeds in real time, applying herbicide only to weed plants rather than broadcasting across the full field.

The documented outcome is a 77% reduction in herbicide use in commercial deployments across millions of US acres. This is an agronomic and economic outcome: less herbicide cost, less chemical resistance development, less environmental load. The system mounts on standard John Deere sprayer equipment, which reduces the capital and operational transition cost versus an entirely new machine.

See & Spray's commercial scale — deployed across millions of acres in the US corn and soy belt — makes it one of the most commercially significant precision agriculture technologies deployed in 2026. It is not a niche research tool. It is production-scale infrastructure on American farms.

Best for: Large-scale row crop operations (corn, soy, cotton) in the US and other markets where John Deere equipment is standard. Operations facing herbicide resistance pressure or seeking to reduce input costs.

Naio Technologies Oz

Naio Oz is a compact autonomous weeding robot for vegetable market garden operations — a different scale and context from See & Spray's row-crop application. Oz navigates between crop rows, mechanically weeding inter-row and in-row positions without herbicide.

Naio Technologies is a French company with deployment concentrated in European market gardens: organic vegetable producers, small to medium farms growing salad crops, root vegetables, and brassicas. The organic segment is a natural fit — mechanical weeding with a robot is both the agronomic and economic alternative to herbicide in certified organic production.

Compare: Agrobot E-Series vs Naio Technologies Oz

Best for: Organic and conventional market garden operations in Europe, small-to-medium vegetable producers seeking to reduce hand-weeding labour, operations growing crops in row configurations suited to wheeled robot navigation.

Harvesting: The Hard Problem

Agrobot E-Series

Agrobot E-Series is a multi-arm strawberry harvesting robot that straddles raised-bed strawberry cultivation rows, using computer vision to identify ripe strawberries by colour and position, and robotic arms to harvest them without bruising.

Strawberries are the proving ground for selective harvesting robots. They are the most commercially significant delicate fresh fruit crop grown at large scale in the US, Spain, and other major production regions. They require harvest at a specific ripeness stage — determined by colour and give — by human pickers who assess each fruit individually. A strawberry picked too early has poor flavour and shelf life. One damaged in handling is unsaleable.

Agrobot has been commercially deploying the E-Series since 2019 in California and Spanish strawberry operations. The current performance gap versus skilled human pickers — in throughput and damage rate — is narrowing but not fully closed. Agrobot's commercial deployments represent the frontier of what selective harvesting robots can do in production environments.

The cost comparison is shifting: skilled strawberry pickers in California are among the most compensated agricultural workers in US agriculture, with piece-rate earnings that reflect the difficulty and skill of the task. As labour shortage drives wages higher and robot performance improves, the crossover point approaches.

Best for: Commercial strawberry operations on raised-bed cultivation systems in the US and Spain. Operations facing severe seasonal labour shortage and willing to operate hybrid human-robot harvesting teams.

Fendt Xaver

Fendt Xaver is a swarm-based autonomous field robot system for precision seeding and crop monitoring in large-scale arable farming. Xaver is not a harvesting robot — it operates in the planting and scouting phases. A fleet of small autonomous robots covers the field, each seeding individual positions with precision placement and mapping crop emergence.

Xaver represents a different automation paradigm from large single-machine equipment: many small autonomous units covering large areas with high positional precision, at a lower per-unit capital cost than large robotic equipment.

Best for: Large arable operations evaluating precision seeding and crop monitoring automation, research into swarm robotics in agricultural contexts.

The Labour Shortage: The Structural Driver

Agricultural automation is not being driven by technology push alone. The structural driver is labour.

The US Department of Agriculture and agricultural industry associations have documented chronic underfilling of seasonal agricultural labour positions. Estimates of unfilled seasonal positions in US agriculture consistently exceed 240,000 annually. In California — the largest US agricultural state by value — strawberry, lettuce, and wine grape operations report that they cannot fill seasonal positions reliably even at wages substantially above minimum.

The pattern is consistent across Western Europe and Japan: declining rural working-age populations, reduced cross-border seasonal labour movement, and increasing competition from other sectors for low-skill labour. The McKinsey Global Institute's analysis of automation adoption trajectories identifies agricultural labour shortage as one of the primary accelerants of agricultural automation investment.

This is the commercial logic behind every platform on this list. Agrobot is not competing against cheap labour. It is competing against positions that cannot be filled at any price in some markets and seasons.

The Cricket's Assessment

> The strawberry is the proving ground for robot harvesting. It is the most delicate commercially grown fruit, harvested at a specific ripeness stage, by touch. If robots can harvest strawberries reliably, they can harvest almost anything. Agrobot has been commercially deploying strawberry harvesting robots since 2019. The industry has been "almost there" on selective harvesting for a decade. Agrobot is past almost. > > The more important near-term story is not strawberries — it is Blue River. See & Spray is treating millions of acres of US row crops with a 77% herbicide reduction. That is a deployed, scaled, documented agronomic outcome, not a demonstration. The fact that it does not look like a "robot" in the popular imagination — it looks like a modified John Deere sprayer — is part of why it does not get the attention it deserves. > > The swarm approach that Fendt Xaver represents is the long-game bet in large-scale arable automation. Replace one large expensive machine with twenty small cheap ones. The coordination and reliability challenges are real, but the capital and mechanical simplicity advantages are also real. Watch that architecture over the next five years.

Frequently Asked Questions

What agricultural tasks are robots currently doing at scale?

Aerial spraying is the most widely deployed agricultural robot task globally. DJI Agras drones cover hundreds of millions of hectares annually for pesticide, herbicide, and fertiliser application. Precision weeding is deployed at scale in US row crops (Blue River See & Spray, millions of acres) and European market gardens (Naio Technologies). Selective fruit and vegetable harvesting is in early commercial deployment for strawberries (Agrobot) and some lettuce operations. Grain harvesting — wheat, corn, soy — is handled by conventional combine harvesters, not robots; this mechanisation occurred decades ago.

Why haven't robots replaced fruit pickers yet?

Selective fruit harvesting requires solving three simultaneous hard problems: identifying ripe fruit among unripe fruit on a variable plant structure (vision), reaching and grasping the fruit without damaging it (manipulation), and doing this at a speed that is economically competitive with human pickers (throughput). Human pickers are remarkably capable at all three. Robots have solved the vision problem reasonably well for uniform crops like strawberries on raised beds. The manipulation and throughput problems, particularly for irregular fruit in three-dimensional canopy structures, remain partially solved. Strawberries and lettuce are the closest to full commercial viability; apples, grapes, and asparagus are further behind.

What is Blue River See & Spray and how does it work?

See & Spray is a precision herbicide application system that uses machine vision cameras to distinguish crop plants from weed plants in real time as the equipment moves across a field. Solenoid-controlled nozzles apply herbicide to identified weed plants only, not to crop plants or bare soil between them. The result is approximately 77% less total herbicide applied versus conventional blanket spraying, documented across commercial deployments on millions of US acres. The technology was developed by Blue River Technology, acquired by John Deere in 2017 for $305 million, and is deployed on John Deere-compatible equipment.

How many seasonal agricultural positions go unfilled in the US?

Estimates of unfilled seasonal agricultural positions in the US exceed 240,000 annually, based on USDA and agricultural industry association data. The shortage is most severe in California for specialty crops — strawberries, table grapes, lettuce, wine grapes — and in other high-wage agricultural states. The shortage is structural: it reflects declining availability of seasonal agricultural labour, not a temporary cyclical phenomenon. This labour gap is the primary commercial driver of investment in selective harvesting robotics.

What crops will robots harvest first?

Crops most likely to achieve reliable robot harvesting first share common characteristics: high labour intensity, high crop value per hectare, plant structures that present the harvest target in consistent positions, and sufficient tolerance for some harvest loss or quality variation during the technology maturation period. Strawberries on raised-bed systems, protected environment lettuce, and processing tomatoes are the nearest-term targets. Apples, wine grapes, and asparagus are medium-term targets with harder robotics problems. Root vegetables and field tomatoes are longer-term given field variability and plant structure complexity.

What is the ROI case for agricultural drones like DJI Agras?

The ROI case for aerial spraying drones is documented across Asian markets and is increasingly demonstrated in North and South America. Key drivers: labour replacement for spraying on difficult terrain, treatment of small fragmented plots that ground equipment cannot economically service, faster application timelines reducing crop loss windows for disease and pest events, and precision application reducing input cost versus broadcast methods. At 20–40 hectares per hour coverage rate, a DJI Agras T50 with one operator replaces multiple ground sprayer passes and significantly reduces the man-hours per hectare for chemical application.

How does Naio Technologies Oz work?

Naio Oz is a small wheeled autonomous robot that navigates between crop rows in vegetable fields, using mechanical implements to weed inter-row and in-row positions. It uses GPS and camera-based navigation to follow crop rows without damaging plants. The robot works mechanically rather than with herbicide, making it suitable for organic production systems. Naio Oz is designed for market garden scale operations — small to medium vegetable farms — rather than large arable operations. It is deployed primarily in France and other European markets for salad crops, root vegetables, and brassicas.

The Geppetto directory tracks agricultural robots across spraying, weeding, harvesting, and field monitoring categories with full specs and deployment data.