Extreme summer droughts have had a severe impact on harvests across Europe. In Austria, yields have fallen by as much as 70%. 

As farmers look for ways to adapt, more are turning to artificial intelligence – from satellite-guided tractors to automated milking systems – to make farming more efficient. Some test fields are already showing significant yield increases.

Gerhard Zimmermann and his Lower Austrian company are developing smart farming technology, including tractor-mounted spot sprayers that can distinguish crops from weeds, directing pesticides only at unwanted plants.

"We detect with AI the plants and the weeds," he tells CGTN. "We process this information in real time on high-performance GPUs and then act with high speed nozzles directly and apply the crop-protection products really precise."

Until now, conventional camera systems could only distinguish green plants from brown soil. But this AI technology is much more clever.

Studies have found that AI-assisted spot sprayers cut pesticide costs by more than half, potentially paying for themselves in less than three years. Zimmermann says the technology also prevents crop damage, particularly to delicate vegetables such as onions.

"If you do not hit the onion with the crop protection product, you get a yield increase of maybe 15 to up to 20 or 25%," he says.

His family business developed the technology two years ago and has doubled its sales year-on-year ever since. But it's not alone. AI applications are increasingly finding their way onto farms of all sizes, from large agricultural operations to small shepherds and vegetable growers.

According to Martin Hirt of the Austrian Chamber of Agriculture, many small and medium-sized farms are even developing their own software using AI agents.

"There are calculation programs for feeding rations, there are planning instruments for fertilization, plant protection, seeding," he tells CGTN.

"Only 10% of farmers are skeptical about AI and digital farming – 90% say if it's useful, and the costs are lower than the expected gain, then why not?"