Farm Robots Are Leaving the Pilot Phase. Two Announcements Prove It.

In the same week that London hosted the World Agri-Tech Innovation Summit, two announcements captured how far field robotics has come. A Japanese machinery giant expanded its European rollout of a Norwegian weed-spraying robot. A British startup set a target of 300 disease-fighting robots in strawberry crops for next year. Neither story is about a research prototype. Both are about machines being sold, serviced, and put to work.

That is the real story of agricultural robotics in 2026. The prototypes worked years ago. The question was whether anyone would build the dealerships, parts warehouses, and business models to run them like ordinary farm equipment. This week answered: yes.

The sprayer that sees individual weeds

The Kilter AX-1 is not a tractor with a robot bolted on. It is a small autonomous machine, 284 kilograms, that crawls through vegetable fields at about half a meter per second, scanning soil and plants with cameras. Artificial intelligence distinguishes crop from weed in real time. Then, instead of broadcasting chemical across the field, spray modules deposit small droplets directly onto individual weeds through what Kilter calls Single Drop Technology.

The resolution is 6 by 6 millimeters, fine enough to treat weeds growing between and within crop rows. The manufacturer reports herbicide savings of up to 95 percent compared with conventional spraying. Kilter commercial figures add weight to the claim: between January 1 and August 16, 2026, 54 AX-1 robots treated 4,265 hectares across Norway, Sweden, Germany, the Netherlands, and Australia. Against a conventional broadcast reference rate of 250 liters per hectare, the company calculates that more than one million liters of spray mixture was never applied.

Two caveats belong with that number. It refers to spray mixture volume, not concentrated active herbicide, and it is a company calculation rather than an independent measurement. Actual savings depend on weed density and conditions. Still, the fleet count tells its own story: 66 AX-1 robots delivered by October 2026, up from a handful in the Nordics a year earlier.

Kubota has a role that may matter more than the resolution figure. The Japanese manufacturer, which led Kilter 6.5 million euro pre-Series B round in February 2026, will now sell and support the AX-1 across more of Europe through its Smart Farming Solutions division and existing dealer network. Spare parts are already stocked at Kubota European Parts Center in Thionville, France. Germany and the Netherlands are live; France and the United Kingdom are next. For small robotics companies, access to a service network is the difference between a clever demo and a product.

The farm robot industry is learning that selectivity does not have to come from chemistry. It can come from cameras and software.

Strawberries, under violet light

Antobot, a British startup founded in 2019, took a different route to the same idea. Its robots carry panels of UVC light that roll between strawberry rows at night. The light damages the DNA of pathogens such as powdery mildew so they cannot replicate. Nothing is sprayed, so there is no residue and workers can re-enter the field sooner.

Strawberries are among the most heavily sprayed fruit crops in the world. They grow close to the ground in damp air, which invites fungal disease. Antobot approach is preventive rather than rescue: short nightly passes keep disease pressure low before an outbreak starts. The company reports big reductions in chemical spraying and, interestingly, yield increases that it attributes partly to healthier plant canopies.

At the London summit, commercial chief Marc Jones announced the 2027 target: 300 autonomous robots in strawberry crops, which would rank among the largest fruit-industry robotics rollouts ever. Customers already include Driscoll, one of the world largest berry companies, and Tiptree. The robots operate in the UK, Spain, Portugal, the Netherlands, Canada, and parts of Asia. A smaller share of the fleet will handle scouting, spraying, and logistics tasks alongside the UVC units.

The more telling announcement was the business model. Until last year, Antobot operated exclusively through robots-as-a-service, which lowered the entry cost for growers but left the startup responsible for servicing and daily operations. From next year, most customers will buy the machines outright. That shift suggests growers now trust the technology enough to own it, and that a startup scaling toward 300 units cannot stay in the service business.

Why commercial scale is the real breakthrough

Put the two stories together and a pattern emerges. The farm robotics industry is solving its distribution problem. Kubota is lending Kilter its dealers and its parts warehouse. Antobot is signing distributors in the UK and Spain and setting up a joint venture in Asia. The technology challenge was plant-level precision; the business challenge is who shows up when a robot breaks during harvest week.

The economics explain why. Strawberries, vegetables, and herbs are high-value crops where weed control and disease pressure consume a large share of operating costs. Seasonal labor is hard to find. Chemical rules keep tightening. A machine that works through the night, treats only what needs treating, and leaves no residue answers all three pressures at once. For the growers doing the math, that matters more than whether the robot drives itself.

There is also a consolidation logic at work. Big machinery companies are not building every robot themselves. Kubota partnership strategy, invest in a specialist, distribute through the network you already own, lets it enter specialty automation without betting on a single internal design. Expect more of these pairings: the specialist provides the precision, the incumbent provides the parts counter.

The fine print

None of this means robots are replacing farm crews this season. The AX-1 crawls at about 1.1 miles per hour with a spray width under 1.6 meters, which suits intensive vegetable beds but not broad-acre grain. Kilter recommends relatively flat fields or consistent raised beds; uneven fields and ridge systems may be unsuitable. And the machine is not the battery-only future its silhouette suggests: a gasoline generator powers the onboard batteries and electric drive.

Antobot 300 figure is a target for 2027, not a deployment that has happened. Targets in farm technology slip; manufacturing, weather, and funding all bite. UVC treatment also demands repetition: miss too many nightly passes and disease catches up. Neither company has published independent, peer-reviewed trials of long-term efficacy; the performance numbers are company-reported and should be read that way.

Those caveats do not change the direction of travel. A decade ago, farm robots were keynote slides. In 2026, they are purchase orders: 66 sprayers delivered, 4,265 hectares treated, parts stocked in Thionville, and a 300-unit strawberry fleet on the horizon. The industry second act will be won or lost on unglamorous ground, dealer networks, uptime during harvest windows, and machines that earn their keep without a service contract. The fields are not just automated. They are becoming commercial.