Agro-Autonomy – Physical AI Is Making it Possible

Mechanized agriculture has been around for a while. The last 5 years have seen substantial progress in autonomous agriculture. Unlike autonomous cars which operate on public roads and uncontrolled environments, agriculture occurs on private land where everything is more controlled, which minimizes regulatory hurdles and potential for accidents.

CES 2026 in Las Vegas (a Consumer Electronics Show), had a full display of companies accelerating autonomy applications in all phases of agriculture – from seeding and transplanting to harvesting, cleaning and delivery – farm to table. At first glance, people may wonder – what has “Consumer” and “Electronics” got to do with these massive machines on the exhibition floor. Well, the answer is simple. Food (along with water and air)) is what humanity lives on. The ultimate consumer product. And electronics-enabled physical AI is a big part of this.

Driving this is something that was discussed in a previous article about construction autonomy. The main factors driving the autonomy thrust in agriculture are:

1. Addressing the gaps in skilled workforce availability that is willing to work in harsh and remote conditions, especially in targeted time frames that agriculture demands
2. Enhanced safety, lower accidents
3. Higher capital utilization, and the use of sensors and machine learning to predict downtimes and proactive machine maintenance
4. Continuous optimization of process flows based on data and experience-based learning, including intelligence for the next season
5. Higher profitability and productivity for end users of this equipment
6. Eliminating repetitive tasks for humans that lead to distraction, quality issues and safety incidents. Physical AI machines excel at this. They never get bored or distracted.

The rest of this article discusses innovation in agro-autonomy by two large multi-national companies (John Deere and Kubota) and a smaller, technology driven company based in Texas (Hylio).

John Deere – We Innovate on Behalf of Humanity

Everyone in the USA has seen a John Deere, at one time or another – working tirelessly on a farm, slowly making its way on a rural highway or local road, constructing a highway, or simply mowing your lawn. Founded in 1837, today it is a publicly listed company headquartered in Moline, Illinois. It has revenues of $52B/year and 75,000 employees, delivering agriculture, forestry, and roadbuilding equipment across the globe. At CES 2026, the company had a novel display of its efforts in all phases of agriculture – tillage, planting, growing and harvesting (Figure 1):

Melissa Neuendorf, a lifelong Iowa native with a passion for agriculture, and now Head of AI for Customer Experience at John Deere, explained this patiently to someone (me!) who’s never been on a farm:

1. Tilling: Chisel ploughs and vertical tillage applicators uses precision technologies to place seed at optimal depths, increasing yield. As this is done, residue, weed, and soil compaction operations along with bed preparation are performed.
2. Planting: planting equipment helps individual farmers seed their farms, based on the crops they grow, land topography, and farm size. This is typically done with semi-autonomous equipment moving at ~10 mph (hand free-driving) using cameras at the seeding arms to gauge exact depth at which seeds are planted.
3. Growing: irrigation and fertilizers are key elements for optimal growth. Another is weed pest control which John Deere delivers with it’s See & Spray™ precision sprayer (operating at speeds of 15 mph) that targets specific areas of the planted land. This reduces herbicide use by > 50%, reduces water use and crop stress, and improves costs efficiencies and profitability. Using cameras and image processing, the spray nozzle is activated at the edge (something that is difficult for humans to do repeatedly). Autonomy also helps shield humans from chemicals and related health impacts.
4. Harvesting: this is where the rubber meets the road and the results of a season’s labor of blood, sweat, tears and stress finally delivers. The fate of many farms depends on this final step, and doing it efficiently and in a timely way is critical for product quality and the logistics of getting it to the market. At CES 2026, this was one of the more dramatic exhibits on the floor. The X-9 Combine (Figure 2), harvests, processes, separates and delivers the final product:

The X9 is a diesel-powered “Factory on Wheels”, operating 18 hours/day and operating at speeds of up to 10 mph. It has 35 different versions of the front header tool (shown in Figure 2 for corn) depending on the crop type. It uses camera and satellite imagery to characterize crop density (based on which it modulates speed). In Figure 2, after the corn stalks are plucked, the cobs are separated and the chaff pulverized. The cobs are delivered to an on-board machine which strips the corn from the cob. An on-board grain feeder delivers the corn kernels into a delivery truck which drives it to a packaging center. All of this is done autonomously. There are humans in the cabin of the combine – mostly agronomists who study the operations and yield, and use this information to optimize seeding plans for next season. X9 has 18,000 parts and operates with an autonomy stack that has 70M lines of code (by comparison, autonomous driving companies like Waymo reportedly use > 100M lines of code).

Ms. Neuendorf explained that John Deere’s autonomy capability and intelligence relies heavily on connectivity and data from > 1M of its connected machines servicing > 0.5M acres globally. This allows the company to learns from experience and apply it to strategies going forward. In 2024, John Deere announced a strategic collaboration with Space X to secure Starlink connectivity for its machines. Per Ms. Neuendorf, “farmers today face mounting pressures from tight harvest windows and skilled labor shortages. John Deere’s X9 combine is designed to help farmers do more with less by integrating precision technology, automation, and data intelligence during one of the most important times of the year: harvest. The X9 shows how advanced technology can help boost productivity, enhance precision, and build more resilient operations for the future”.

Kubota – For Earth, For Life

Established as a casting foundry in Japan ~150 years ago, Kubota today is a global corporation with revenues of ~$20B/year and ~50K employees, 15% based in the United States. Today, the company makes machines to support construction, agriculture and other blue collar industries. The focus is on solving global challenges like skilled labor shortages, an aging farming population, and the need to feed a growing world. Kubota’s focus is on smart, right-sized, user-centric equipment with real-time intelligence to give customers the freedom to work in the field or through a smartphone app.

Per Brett McMickell, Chief Technology Officer for Kubota North America. “Physical AI is a key inflection point for our industry and for Kubota. Decision-making, obstacle detection, and voice recognition capabilities mean AI based real-time insights will now inform tasking, labor assignments, and efficiency improvements. We can now facilitate more complexity with more certainty and more simplicity than ever before”.

Kubota recently announced the commercialization of a smart, integrated autonomous solution developed in partnership with Agtonomy for the specialty crop market, fully built into the 105.7-horsepower diesel Kubota M5 Narrow tractor with advanced sensing and artificial intelligence. One application is for grape crops in wineries where compact equipment is essential to navigate narrow spaces between vineyard rows. Figure 3 shows examples of Kubota’s autonomous equipment for rice crops.

Cameras, LiDAR, radar and ultrasonic sensors are used for perception and mapping, while GNSS sensors are used for localization and path planning.

Kubota is passionate about water, food and the environment. According to Todd Stucke, General Manager of Agri Solutions Headquarters, Kubota Japan (KBT), and President of Kubota North America, the company is investing heavily in the “Farm of the Future”, an R&D initiative being prototyped in Thailand. Per Mr. Stucke, “big challenges don’t just need big machines; they need smarter solutions that make life easier. Our go-to-market solutions are the culmination of our customer-driven innovation cycle, where the goal is not simply to automate what you’ve always done, but to rethink how work can be done more intuitively and efficiently”.

Hylio – Your Crops, Our Mission

A Texas-based agdrone company, established more than 10 years ago, and led by CEO and founder, Arthur Erickson, an aeronautics engineer, passionate about agriculture. The focus of the company is to build drones (in the United States), specifically for agricultural autonomy. It builds various types of models with different payload capacities, primarily for precision delivery of herbicides, fertilizers, insecticides and water. Its technology can support individual drones or drone swarms that work in a collaborative autonomy mode to accomplish a mission.

In the past 5 years, it has sold ~1000 of these drone units (revenues of ~$35M). The drones typically fly very close to land (10 ft), and deliver precision sprays of appropriate liquids on the crop. A human is in the loop, since FAA regulations still do not allow completely autonomous operation. Figure 5 shows examples of the drones supplied by Hylio.

1. HYL-30 Pegasus: 4 gallon tank, 20 lb. payload, up to 13 acres/hour (max) @ 2 gallons/acre spray capacity. Applications include spot-spraying of weeds, application of nutrients to specific plant zones, and treating disease outbreaks with minimal chemical usage and zero overspray.
2. HYL-150 Ares: battery powered, 30 gallon tank, 100 lb. payload, 70 acres/hour (max) @ 2 gallons/acre spray capacity. Equipped with obstacle avoidance sensors like radar, and RTK (real time kinematic) positioning capabilities (Figure 6).

Two other drones that are part of Hylio’s drone portfolio include:

1. HYL-300 Atlas: 50 gallon tanks, 250 lb payload, 150 acres/hour (max) @ 2 gallons/acre spray capacity
2. Photon Scout Drone: for ISR (Intelligence, Surveillance and Reconnaissance) for commercial and military applications. It is battery powered and has mission times of up to 55 minutes (with imaging payload like visible and thermal cameras.

These drones can be operated autonomously or under operator control via Agrosol, Hylio’s agricultural UAS ground control software (GCS), through an intuitive interface. An integrated Windows tablet and RC controller allow users to seamlessly switch between autonomous and manual controls for one or multiple drones. An array of sensors delivers real-time access to vitals such as GPS position, flowrate, altitude, and more. All flight and treatment data are stored within user’s accounts providing access to as-applied maps, maintenance data, and more within the Hylio interface.

Per CEO Arthur Erickson, key advantages of Hylio’s drone based agriculture solutions include coverage rate, precision and cost (a Hylio drone costs ~$40,000 vs a $300,000 tractor). “Hylio is aggressively pursuing a fully autonomous end-to-end tech stack for our drone family that requires no human intervention for day to day operations. This requires hardware such as automated recharging/payload reloading stations and robust AI software tools that allow the drones to plan, manage, and execute a wide variety of tasks completely on their own. Once the drones are capable of working independently, incredibly scalable business models become possible as there is no longer a need for human operators who require presence, training, and licensing. Millions of ‘drone in a box’ solutions can busily work in the background across the world, out of sight, out of mind, laboring away to automate many personal and industrial tasks”.

Farm-to-Table is a great marketing phrase, coined to get rich diners to restaurants and pay a premium. But < 2% of the U.S. population can afford such restaurants. The bulk of the United States, and certainly the world, is fed by more affordable options that agro-autonomy enables (all the produce we eat eventually comes from a farm). The toil and sweat that’s behind the scenes to make this happen is admirable, and the technology advances in autonomy are inspiring.