How a small Maharashtra town taught the world a lesson in smart farming

On July 3 at a United Nations Food and Agriculture Organisation (FAO) summit in Rome, global experts were discussing the future of farming. A project from Baramati, Maharashtra, stood out, claiming its agricultural AI was "superior to global benchmarks". By implementing a pilot across 10,000 sugarcane farms, yields doubled while water use decreased by 40-50 percent and fertilizer use fell by 30 percent. Soil health also improved.

What sets this apart is that most agricultural AI is created for large, uniform farms in the US or Brazil, but India's farms are small, fragmented and share water sources. To meet this reality, the Agricultural Development Trust (ADT) in Baramati built a bespoke AI system called Tiny AI. This lightweight system runs on ordinary smartphones and is affordable, needing only one automated weather station to cover 2.5 square kilometers of land, allowing around 25 farmers' plots to be monitored.

The system captures 12 different parameters every 30 minutes - soil moisture, temperature, humidity, rainfall, sunlight - and satellite imagery tracks canopy health. Field teams validate data with real soil samples. The AI predicts water stress, nutrient gaps, and fungal risks up to a week in advance. Farmers receive plain-language Marathi alerts on their phones, such as "Sugarcane leaves yellowing.

Reduce urea by 15 percent, delay irrigation by 2 days." This action-oriented approach, using only data India farmers can access, has proven successful.

The water savings achieved by shifting from calendar to evapotranspiration-based irrigation has been the biggest impact in drought-prone sugarcane belts. The system is expanding to other crops, including pigeon pea, turmeric, banana, grapes, cotton, and soyabean. The ADT is also piloting hydroponic lettuce and AI-guided banana cultivation.

They aim to integrate 'AI in Agriculture' into college curricula and train women and self-help groups to use the technology. Baramati's success shows that AI can work on Indian farms, raising questions about how to scale this model globally.