Tomatoes tailored for robots in greenhouses

Automated greenhouse work hinges on tomato variety.

Eternal, a robotics firm, and seed breeder Rijk Zwaan have joined forces to explore which tomato varieties best support robotic crop work in greenhouse environments. The collaboration is purpose-built: pair plant genetics with autonomous systems so future greenhouse tasks can be performed with less manual intervention. In practical terms, the effort aims to identify plant traits that make pruning, scouting, shaping, and potentially selective harvest more reliable when robots are allowed to operate in controlled greenhouse spaces.

The partners are pursuing a knowledge-driven approach, sharing insights to understand how different plant characteristics affect robotic performance. The work centers on traits that influence how a plant presents itself to a robot, including how the plant grows up a support system and how leaves and fruit are arranged in the robot's working envelope. While the exact traits to be prioritized will emerge from hands-on testing, the project is built around the premise that genetics and robotics must be aligned from the outset if fully automated crop work is ever to move from prototype to routine practice.

Integration is the second front of the initiative. Robots in greenhouses rely on a blend of perception, reach, grip, and environmental sensing, all of which must be compatible with plant architecture. The project will evaluate how different tomato varieties affect a robot’s line of sight, grasp approach, and end effector design, as well as how plant structure interacts with trellising, staking, and irrigation around the robot's work path. In short, this is not a purely genetic exercise; it is a systems exercise that tests how plant form feeds into robotic control loops, vision algorithms, and maintenance cycles.

The work is also a reminder that automation is a capital and process decision, not a single technology purchase. For plant managers and CFOs, the initiative signals a longer, staged path to ROI that begins with pilot testbeds, moves toward measurable cycle times and throughput improvements, and ends with standardized plant material that minimizes retooling and manual interventions. The collaboration's ultimate value will be judged in concrete, plant-by-plant demonstrations: how quickly a robot can perform a routine crop-care task on a given variety, how often it requires recalibration, and how the broader greenhouse automation stack, climate, irrigation, and data systems, stays in sync with the breeding choices.

Two practitioner-level realities emerge from this type of work. First, integration requirements matter as much as genetic traits. Cycle times and throughput depend on how well the plant traits align with robot reach, gripper geometry, and sensor resolution; without harmony between plant form and machine design, even strong varieties can cause bottlenecks. Second, this effort reframes what skilled trades will do in a highly automated greenhouse. Automation is destined to augment technicians, inspectors, and crop specialists who supervise, maintain, and fine-tune the robots rather than replace them outright. Expect the human role to shift toward monitoring robot performance, validating results, and steering breeding programs to yield still more automation-friendly plant types.

Looking ahead, the alliance will proceed in phased pilots across greenhouse environments to capture actionable metrics. The project will track how trait-driven changes influence the effectiveness of autonomous tasks, what integration gaps appear between plant material and robotics, and how long it takes to scale from a test plot to a full-operation setting. If the approach proves durable, it could redefine how tomato varieties are selected for automated greenhouses, creating a feedback loop where breeders tune traits specifically to maximize robot reliability and plant health at scale.