For a patent attorney, this transition is fascinating. Oversized Lamborghini-branded tractors and wonky tramlines are out. Ultra-precise AgBots and autonomous seeding and weeding FarmDroids are in. This article looks at the engineering behind these new agricultural co-stars and the intellectual property surrounding them.

The FarmDroid FD20: an all-in-one seeding solution

One major arrival at Diddly Squat is the FarmDroid FD20, a lightweight, solar-powered field robot that looks less like a traditional tractor and more like a high-tech lunar rover. The engineering problem the FarmDroid solves is elegant in its simplicity, yet incredibly complex in execution. Traditional weeding requires either manual labour or massive chemical applications. The FarmDroid bypasses both. Operating autonomously via high-precision GPS, the machine maps the field down to an 8mm accuracy. As it moves across the soil, it logs the exact coordinate of every single seed it drops into the earth.

Because the FarmDroid retains this digital coordinate database, the robot does not need to look for weeds visually once the crops begin to grow. It simply returns to the exact coordinates between the known seed placements and deploys a mechanical hoeing blade to destroy weeds before the actual crop ever breaks through the surface. The FarmDroid can even perform targeted spraying of weeds exactly where it is needed, which FarmDroid claim can reduce the use of chemicals by up to 94%.

The core pillars of the FarmDroid’s functionality represent distinct engineering feats that allow for separate layers of patent protection. This covers everything from the physical mechanics of the single-seed planting apparatus and specialized weeding rods to the algorithmic logic governing autonomous steering. Each of these innovations could potentially warrant its own patent, which can provide robust, wide-reaching commercial protection.

DUALEM 21HS: mapping the invisible landscape

Before any seeds can be sown with such absolute precision, however, the farmer must understand the unseen landscape beneath the topsoil. To achieve this, the show introduces proximal soil sensing via the DUALEM 21HS scanner, an instrument towed behind an all-terrain vehicle across Diddly Squat.

The DUALEM 21HS is entirely non-invasive; it does not dig a single borehole or take a single physical core sample. Instead, it relies on electromagnetic induction (EMI). The sensor consists of a long, sealed tube containing a single transmitter coil at one end and multiple receiver coils spaced precisely at the other. The transmitter passes an alternating current to generate a time-varying magnetic field. This field penetrates the earth, inducing tiny, localized eddy currents within the subsoil.

These underground currents produce their own secondary magnetic field, which is intercepted by the dual horizontal and perpendicular receiver arrays. By measuring the ratio between the primary and secondary fields, the device simultaneously records apparent soil electrical conductivity across multiple depths of exploration. This data identifies invisible variations in soil compaction, clay interfaces, moisture capacity, and organic matter content in real time.

The AgXeed AgBot: translating data into physical action

The data harvested by the soil scanner is a valuable asset, but it remains a passive map until it is translated into mechanical action. This is where the heaviest piece of precision kit arrives on the farm: the AgXeed AgBot.

The AgBot is a fully autonomous, cab-less crawler tractor powered by a 230-horsepower diesel-electric drivetrain. Unlike traditional machinery that relies on a human operator to constantly adjust speed and depth based on visual cues, the AgBot is an execution engine for digital data. It moves on soil-preserving rubber tracks, guided entirely by pre-planned path coordinates uploaded through cloud-based management systems.

When paired with a smart fertilizer spreader or sprayer, the AgBot executes what is known as selective, or variable-rate, fertilisation. Rather than blanket-spraying a uniform layer of nitrogen across an entire field, resulting in over-fertilizing nutrient-rich zones and wasting inputs on unproductive, gravel-heavy patches, the AgBot reads the digital prescription map derived from the DUALEM sensor data. As the tractor traverses the field, the tool control unit dynamically meters out the exact volume of fertilizer required by the specific square meter of soil beneath it.

The data-driven precision of the machine relies on two distinct layers of innovation. While the physical chassis relies on patentable mechanical milestones such as its soil-preserving track geometry and specialized high-torque suspension, it is the software architecture that truly drives the system. AgXeed has worked to secure patent protection for both the physical machine and the underlying algorithms governing real-time route planning, geolocation, and tool control, systematically bridging the gap between traditional agricultural machinery and the digital cloud ecosystems orchestrating modern field operations.

The value of intellectual property 

A multi-tiered agritech ecosystem of this scale has great potential in terms of patent strategy. While there are undoubted mechanical and electrical innovations in these systems which should be protected, without the underlying software these systems would not be capable of achieving the significant advantages promised.

The functionality of software can generally be protected by patents, provided the invention provides a technical effect beyond the software itself. Generally, in the context of controlling machinery, this is usually the case. For example, more precise control of an autonomous tractor, or autonomous seeding, weeding and spraying are all ‘real-world’ effects and thus confer technical character on the software inventions, making them patentable. Conversely, software that operates purely in the realm of data analysis, business management, or the presentation of information will struggle to clear the technical hurdle. For example, software which schedules deliveries based on wheat prices or which provides intuitive colour-coded user interfaces for farmers are generally not considered patentable by the European Patent Office (EPO).

That said, patents are only one piece of the digital IP puzzle. The literal software code itself is automatically protected by copyright, preventing competitors from directly copying or replicating the source text, though not the ideas that the software is expressing. Meanwhile, for the deeper layers of the software stack, particularly machine learning models, trade secrets offer a complementary strategy. While a patent is ideal for protecting the overarching functional utility of an AI-driven agricultural process, trade secrets are an option to protect the proprietary model architectures and optimized weights. Because these weights are often executed within secure cloud environments or encrypted modules directly on equipment, they are in principle shielded from public view, making robust operational secrecy a powerful way to maintain a competitive advantage without the need for public disclosure. However, concealing the weights of an AI model does not protect you from someone else using similar training for their own AI, and terms of use may be needed to prevent so-called distillation – training a new AI to emulate the outputs of your AI – so it can be useful to expressly protect training regimes and training datasets as well.

The future of farming, today

Precision farming has fundamentally changed the agricultural landscape. Innovation is no longer measured in raw horsepower or the sheer weight of structural steel, but in sensor frequencies, communication speeds, and algorithmic efficiency. For the engineers and tech startups developing these tools, establishing a proactive, sophisticated European patent strategy early is the only way to ensure that their underlying innovations are safely protected before the market matures.

The transformation observed at Diddly Squat Farm reflects a wider industrial reality. As agricultural businesses grapple with tighter margins and stricter environmental targets, the transition to autonomous, selective inputs is accelerating, providing farmers with unprecedented predictability and control over their crops. The one thing we still can’t control, however, is the weather.