Looking back on the GTPB Café smart greenhouse with drones, data and AI

On 11 September 2025, growers, tech companies and education came together at Van der Avoird Trayplants in Molenschot for the GTPB Café. In this recap you can read what the event delivered. The core is clear. When we test together, organise data properly and stack functions, technology becomes deployable in the greenhouse faster and the threshold to invest goes down. NXTGEN connects the parties and translates questions from the greenhouse into projects, so risks are shared and results become measurable. This is how the path grows from first applications to broad use of drones, data and AI in daily cultivation.

The different speakers each highlighted the subject from their own perspective, creating a complete picture of what it takes to move towards a smart greenhouse. It started with a shared future vision and the choice to learn together through pilots, which creates direction and speed. From that base it became clear that innovation only counts when it truly simplifies the grower’s work, and that networks and projects form the bridge to implementation. This requires reliable, consistently collected data, so that AI and imaging deliver usable signals and a consortium can bundle its energy. Autonomous drones already show the value of early detection, while adoption also demands clarity on costs and returns, especially since the Netherlands is still lagging in application. The practical route is to start with one application that delivers value quickly and to use the generated data for the next step. In this way evidence grows and risk decreases. Multispectral images make this tangible, because you can intervene earlier when growers and suppliers jointly test what works. In this interaction a learning path emerges in which vision, practice and data reinforce each other.

Data is the engine, but also sensitive. Growers rightly want certainty about what happens to their data. Clear agreements on ownership, use and compensation are needed so knowledge can flow without giving away competitive advantage. Initiatives such as an open data model from the lily sector show that ownership and access can be carefully separated, accelerating collaboration. When this is arranged properly, pilots become repeatable, results comparable and suppliers can improve more quickly.

The greenhouse is not yet designed everywhere for autonomous flights between gutters and wires, but most obstacles can be solved when you design together at an early stage. At the same time the business case is often the bottleneck, because manual scouting still feels cheaper. That comparison shifts once you stack functions, reuse data and adapt processes around the drone. As datasets grow and multiple use cases run on the same platform, costs per application go down and value becomes visible in labour, crop security and input use.