That picture is changing. On fields in Kraggenburg and Ens, a white machine glides silently through the furrows. Day and night. No coffee breaks, no toilet breaks. The Maverick by Odd.Bot is not a spectacular robot from a science-fiction film. It is a patient worker. One that pulls two million weeds from the ground in two weeks, with a precision of two millimetres. [1]
From delivery bike to AI-driven weeding arm
Martijn Lukaart did not start out in agriculture. In 2015 he founded Bun.Run, a platform for local bicycle deliveries. Urban logistics, last-mile delivery. The typical startup vocabulary of the time. But somewhere along the way his focus shifted. He saw a more urgent problem.
The organic sector was growing, but had hit a wall. No herbicides meant manual weeding. Costs of between one thousand and two thousand euros per hectare. And the people needed to do that work were becoming increasingly scarce. Seasonal workers from Eastern Europe more often chose jobs in their home countries, where wages were rising and unemployment was falling. [2]
In 2018, Lukaart began working on a prototype together with students from TU Delft and RoboHouse. Could an AI system recognise every individual weed plant among thousands of crops? And could a robot remove that weed without damaging the root next to it? The answer turned out to be yes, but the path there took years of training, iteration and fieldwork.
Artificial intelligence among the carrots
The Maverick now operating across Dutch, Belgian and German fields is the fourth generation. It does not navigate by GPS but by computer vision. Cameras track the crop rows, AI models analyse every plant in the image and distinguish crop from weed. Two mechanical arms then pull only the intruders from the ground. Root and all.
The figures are impressive. The robot can work roughly one hectare per day. Per second, it removes up to two weeds per arm. And the system operates 24 hours a day, seven days a week. The only human action required is swapping the batteries. [3]
"We got this carrot field clean without pulling a single weed by hand. Only the robot went in. No coffee breaks, no toilet breaks, it just keeps going. At night too.", Bert Benedictus, organic grower in Kraggenburg [1]
The price ranges from 90,000 to 120,000 euros depending on the configuration. That sounds like a significant investment. But do the maths. Digni van den Dries, an organic grower in Ens, would have needed twelve people to do the same work. And those people simply no longer exist. [1]
"Two weeks of driving, two million weeds gone. I would have needed twelve workers for that. And you can't find them anymore. This robot has saved me labour and given me peace of mind.", Digni van den Dries, organic grower in Ens [1]
Scarcity as strategy
This is where it gets interesting. The first production run of 2025, sixteen robots, sold out before assembly even began. Pre-orders for 2026 are nearly full. Demand is clear. And yet Odd.Bot is choosing not to accelerate.
It is a deliberate choice. Lukaart keeps production limited to guarantee quality and service, together with local partners. [4] Those partners are crucial. Machinery companies such as Weevers BV in Swifterbant and Doorgrond in the north handle delivery and support. An AI-driven robot that heads into the field without good service is an expensive paperweight. Odd.Bot builds trust first, volume later.
There is something else noteworthy. In the latest funding round of two million euros, led by Iconic Ventures, regional farmers also participated as investors. [4] Not anonymous venture capitalists, but growers who see the robot working in the field every day. They are not just buying a product. They are betting on the future of their own sector.
A lesson from Flevoland
Twenty Mavericks are now operating across fields in four countries. They work in carrots, onions, chicory and red beetroot. The technology works. The remaining question is whether Odd.Bot can build the capacity it needs without losing the quality that is currently convincing farmers.
For other startups, there is a lesson in this story. Funding does not always have to come from Silicon Valley or traditional VCs. Sometimes your best investors are right there in your customer base. Farmers putting their own money into a weeding robot because they see it perform every day. That is validation no pitch deck can match.