On China's farms, the drone now flies itself

The farmer used to wait for a pilot to show up with a backpack sprayer. Now he taps a screen, walks to the field edge, and watches the rotor do the work. The machine plans the path, dodges the tree line, and comes home when the tank is empty.

That quiet change — from "hire someone to fly it" to "I fly it myself" — is the real story of agricultural drones (农业无人机) in China. The headlines talk about AI (人工智能) and low-altitude economy (低空经济). On the ground, what matters is that the drone now flies itself, and a lot of fields depend on it.

From "hire a pilot" to "own the drone"

Five years ago, almost nobody who owned the land operated the drone. A 2025 interview with a senior solutions engineer at DJI Agriculture (大疆农业) laid out the shift in plain numbers: in 2016, only about 10% of users flew their own machines, while professional pilots (飞手) made up 90%. By 2025, the split had flipped — roughly 60% of users now own and operate their own drone, and pilots account for about 40%.

The people behind the controls are also getting younger. DJI's data shows that users under 40 now make up about 60% of the base, with the 16–29 group alone at 25%. Agricultural drones have become a way for returning youth to earn a living without inheriting a tractor.

The numbers behind the rotors

DJI's 2025 domestic operation was large enough to surprise people outside the sector. According to figures the company gave Farmer's Daily (农民日报) and corroborated by national coverage:

• More than 335,000 drones worked Chinese fields under DJI equipment in 2025.
• Annual operation passed 3.3 billion mu-times (33亿亩次) — roughly 220 million hectares if you count every spray pass, since one field is often treated several times a season.
• Automatic route flying rose from 72% of operations in 2020 to 88.6% in 2025, meaning most flights now run without a human steering.
• In lifting work alone — hauling fertilizer, seedlings and harvested crops that used to move by hand or cart — DJI drones moved 6.5 million tons of material in 2025.

Globally, DJI says over 300,000 drones operate under its agriculture brand, treating more than 500 million hectares of farmland, with cumulative global sales above 600,000 units as of early 2026. Nationally, China's agricultural-drone stock now exceeds 300,000 units and annual operation area tops 460 million mu (4.6亿亩), per the Ministry of Agriculture and Rural Affairs.

XAG built the other pole

The second name to know is XAG (极飞科技), founded in Guangzhou in 2007. Where DJI came from consumer drones, XAG started by aiming at the farm. By 31 December 2025, according to figures the company gave Guangzhou Daily (广州日报), XAG's autonomous equipment had logged:

• Over 2.5 billion mu-times (25亿亩次) of cumulative operation.
• A yield gain attributed to its systems of 26.5 million tons of crops.
• Training for 230,000 agricultural-technology workers.
• Savings of 180 million tons of water and a 2.75-million-ton cut in carbon emissions.

These are company-stated figures carried by state media, not independently audited — but they show the direction: the two firms have turned a hobbyist's toy into the most deployed robotics category in Chinese agriculture.

AI in the loop, not in the brochure

The "AI" here is not a chatbot. It is the stack that lets a drone fly a planned line without a person at the sticks:

• RTK positioning for centimeter-level accuracy, so the spray overlaps correctly and misses no strip.
• Terrain-following and obstacle avoidance, so the rotor climbs over a dike and skirts a power line on its own.
• Field mapping, where the drone scans the plot first, then flies the saved route on every later pass.

DJI's own description of its latest machines — maximum payload around 85 kg, handling spray, broadcast, lifting and surveying — points to why farmers accept them: one airframe covers most of the season's jobs.

The policy tailwind

This is no longer just a market story. The 2026 No. 1 Central Document (中央一号文件), released in February, explicitly called for "expanding the application scenarios of drones (无人机), the Internet of Things and robots (机器人)" — the first time drones and robots were named in that document. Provincial governments from Anhui to Heilongjiang have since written agricultural low-altitude economy (低空经济) into their own plans. When the top rural-policy document names a machine, the procurement and subsidy logic behind it shifts from experiment to expectation.

Honest limitations

• The DJI 2025 figures (335,000 units, 3.3 billion mu-times, 88.6% auto-route) come from a company engineer interviewed by Farmer's Daily (农民日报), a state agricultural newspaper. They are vendor-stated, not audited by a third party.
• XAG's cumulative totals are company figures reported by Guangzhou Daily (广州日报), again not independently verified, and "mu-times" (亩次) counts repeated passes, so it is not a simple land-area measure.
• Adoption is uneven by crop and region. Rice, wheat and corn lead; orchards and vegetables lag, and hilly terrain still complicates autonomous flight.
• Drone price competition has driven average selling prices down sharply, which helps adoption but squeezes the manufacturers' margins — a structural risk, not a headline.
• These are productivity tools, not labor replacements across the whole farm. Planting, harvesting and post-processing still need people.

What readers can do now

1. If you farm or advise farms in China, pilot one autonomous drone on a single crop — spray or broadcast — and measure input savings (water, pesticide, labor hours) over one season, not one field.
2. Read the original operator data, not the press release. Pull DJI Agriculture's and XAG's published 2025 operation figures and check the measurement date before trusting any "billions of mu" claim.
3. Watch the youth-return number. The shift to 60% self-operated, 60% under-40 users is the signal that this is infrastructure now, not a demo — track it year on year.