Data and AI skills for agriculture students in 2026
600,000 DJI farm drones were flying by the end of 2025

More than 600,000 DJI farm drones were in use by the end of 2025.
With over 600,000 trained operators, across 100+ countries (DJI).
Farming is already a data job. The question for agriculture and food students is whether your degree has caught up with that... and if not, which AI skills to build yourself.
What's actually happening with AI and data in farming
- Drones went mainstream. DJI unveiled its latest industry report at Agrishow in Brazil last month. It also claims its drones have saved 410 million tonnes of water and cut 51 million tonnes of carbon, but the announcement doesn't explain how those were calculated. (Company numbers, so treat them as claims.)
- Investors are backing the farm end of the chain. Agrifoodtech funding was flat at $16.2B in 2025, but startups working on farms and in food production drew $9B, up 7%. Farm robotics held steady whilst most categories saw fewer deals (AgFunderNews).
- Satellite yield prediction is getting field-level. A March review of research on Sentinel-2 satellite data found a shift from regional estimates to individual fields, using machine learning like random forests and neural networks (arXiv).
- The data exists. The skills often don't. A UK report from The Institute for Agriculture and Horticulture (TIAH) found demand moving towards roles like data analysts and robotics technicians. Some people it surveyed said as much as 90% of the data collected by combine harvesters goes unused (AgNavigator).
That last one is the opportunity. Farms are collecting data nobody reads.
What uni won't teach you about farm data
Agriculture degrees are strong on soils, crops and animals. TIAH itself recommends building digital skills into agricultural education "at all levels", which tells you it isn't there yet.
What often gets skipped:
→ cleaning messy data. Real yield maps have gaps, sensor faults and wrong dates → remote sensing basics. Reading satellite imagery is now an agronomy skill, not just a geography one → model limits. The Sentinel-2 review flags limited ground truth (real measured yields to check against), cloud gaps and models that don't transfer between years or places
The agriculture student's data and AI toolkit
- Satellite imagery: Copernicus Browser. The Copernicus Data Space gives free access to Sentinel data. Pick a farm you know and watch a field change through a season.
- Spreadsheets first, then Python. Most farm forecasting starts with a clean table. Learn pivot tables, then move to pandas (a Python data library) when the spreadsheet gets slow.
- Drone basics. Learn what a spray drone and a mapping drone actually do, and check your country's operator rules before you fly anything.
- A chatbot, carefully. Good for explaining a statistics method or debugging code. Bad as an agronomist. Check every crop answer against a trusted source.
The critical-thinking move: 5 questions for any farm AI claim
- Who measured the result? A company press release, a field trial or a peer-reviewed study? DJI's water figure is the first kind.
- Is there ground truth? A yield model with no real harvest data to check against is a guess with nice colours.
- Will it work on THIS farm? Different crop, soil, climate or year can break a model trained elsewhere.
- What's the money? Turn the claim into cost per hectare. If it can't pay back, farmers won't buy it.
- Is the data already there? Before buying a new sensor, check what the combine is already recording.
What to do this month
- Ask a farmer for one season of yield data. Then just clean it. That alone teaches you more than a lecture on precision farming.
- Track one field from space. Use the Copernicus Browser to compare the same field in spring and summer. Write down what changed and why.
- Build a tiny forecast. Use any public crop yield dataset, forecast next season in a spreadsheet, then try a simple model in Python and compare.
- Follow the price side too. Food prices are getting algorithmic further down the chain, as we covered in AI dynamic pricing in retail. And check every green claim the way we did in AI and climate change.
More on the field in our agriculture and food hub.
so TLDR: farms already collect huge amounts of data. Learn to clean it, read it from space and question every model built on it.
If a farm handed you all its unused harvester data tomorrow, what's the first question you'd try to answer?
Sources
- DJI Agriculture reveals global adoption of agricultural drones cuts 51MT in carbon emissions and saves 410MTs of water for farmers globallyDJI · 29 APR 2026
- New report: Agrifoodtech funding is flat but the story is shiftingAgFunderNews · 18 MAR 2026
- Upskilling workers 'essential' to realise benefits of AI-driven farming, report saysAgNavigator · 20 JUN 2025
- Sentinel-2 for Crop Yield Estimation: A Systematic ReviewarXiv · 24 MAR 2026
- Copernicus Data Space EcosystemEuropean Commission / ESA · 19 MAY 2026
Get the next issue by email
theVANTAGEd is written by students to bring critical thinking to AI across real industries. One issue every two weeks.