High-precision viticulture: this is the future envisaged by the start-up Vitidrone, hosted by Inria Startup Studio. The company is developing autonomous drones combined with AI algorithms to identify deficient or diseased vines in a vineyard. This is an ideal solution for limiting the use of plant protection products while maintaining yields.
"In recent years, pressure on vineyards has intensified significantly. Climate change is causing more frequent droughts, promoting the emergence of diseases and pests, and increasing the number of late frosts, which often have devastating consequences. In this context, the need for accurate and regular information on the state of vineyards has never been more crucial. However, a winegrower cannot spend their days walking around their plots," explains Titien Cubilier, CEO of the Bordeaux-based start-up Vitidrone. When you consider that one vine produces approximately one bottle of wine, and that some bottles sell for several hundred pounds, it is easy to understand that preserving these plants, which take four to five years to become productive, is a major financial challenge for winegrowers.
So how can yields be preserved while limiting the use of plant protection products? Vitidrone offers a solution: an autonomous drone capable of detecting the first signs of deficiencies and diseases in vines, down to the individual vine. This enables precision farming, with treatments being used in a highly targeted manner, before the pathogen spreads, thereby limiting yield losses. ‘An environmentally and economically sound solution,’ sums up Titien Cubilier.
A childhood spent between viticulture and IT
The idea of using drones to monitor vineyards and their operators came quite naturally to the young scientist, as the logical continuation of a childhood spent immersed in the world of viticulture – his family has worked for generations in the great châteaux of Pomerol – and IT (his father was a passionate self-taught computer scientist who passed on his enthusiasm to his son from an early age).
After choosing to study the latter, he obtained a master's degree in ‘Autonomous Systems, Perception, Interaction and Control’ from the University of Bordeaux, then a position as a research engineer for Thales and the Bordeaux Computer Science Research Laboratory (LaBRI), where he specialised in autonomous drone swarms. He also teaches robotics and autonomous drone design at the University of Bordeaux.
The meeting of a drone specialist and a bioinformatician
Then, in 2024, he took the plunge and presented his Vitidrone project to the experts at Inria. He was accepted into the Inria Startup Studio programme... but he still needed to find a partner to help him develop it. Théo Gauvrit Introduced himself, and once again, it was a perfect fit. ‘We met nine years ago, in our second year of university, through an optional course on new agriculture offered by the university,’ recalls the new partner and now CTO of Vitidrone. ‘It was already a sign! We remained friends, and when I heard that Titien was looking for a partner for his project, I offered my services.’
The expertise of the two researchers is indeed complementary, since Théo Gauvrit has a PhD in bioinformatics and, during his previous studies in biology, became interested in plants and their pathogens. ‘The goal of contributing to the viticulture of tomorrow, while taking advantage of an opportunity to discover the world of start-ups, immediately appealed to me,’ he explains.
AI to the rescue
When they began working together in March 2025, the two partners set themselves the goal of having a prototype ready within six months. It was an ambitious challenge: ‘Other companies had tried to deploy a solution like the one we had in mind in the past, but without success,’ says Titien Cubilier. However, we had an extra advantage: in the meantime, the two technologies we needed, namely drones and artificial intelligence, had made enormous progress and their costs had fallen dramatically."
The researchers therefore set to work. On the one hand, they made progress on the algorithmic programming of the drone to enable autonomous flight. On the other hand, with the help of Iban Oyharcabal, an expert in computer vision, they developed georeferencing of vine stocks, as well as AI analysis of high-resolution photos to detect the first two diseases, flavescence dorée and Esca.
Already a first working prototype
Less than six months later, they had a working prototype: it was able to identify the vines, locate them and use NDVI (normalised difference vegetation index) and, above all, high-resolution photos to assess the health of the vines. Titien Cubilier explains the process: "The drone makes an initial pass over the plots to assign a GPS position to each vine. This step only needs to be carried out once, or when the crop is being replanted. The drone then carries out its routine surveillance: it takes high-resolution photos of each vine and uses AI to identify the early signs of disease.
Verbatim
A smartphone or tablet app will allow winegrowers to view the areas requiring attention, down to the individual vine, on each of their plots, obtain recommendations on the treatment to be implemented and check for themselves, using the photos, that it seems appropriate.
Auteur
Titien Cubilier
Poste
CEO of the startup Vitidrone
To test their technology, the researchers are already benefiting from technical partnerships with several wine estates, including Château Le Bon Pasteur in Pomerol and Château l'Apogée in Lalande de Pomerol.
The contributions of Inria Startup Studio
At the same time, the young entrepreneurs received training from Inria Startup Studio on the business aspects of their start-up. In particular, they exhibited at the Vivatech trade fair and benefited from internal events where entrepreneurs shared their experiences, as well as a comprehensive course at EM Lyon. ‘It was an excellent training programme, which helped us make significant progress,’ says Titien Cubilier.
In addition, their place within Inria Startup Studio gave them access to the Institute's equipment. ‘This allowed us to use the PlaFrim supercomputer to train our AI and save valuable time,’ says Théo Gauvrit. Immersion in the Inria ecosystem also provides an opportunity to meet the best research teams in computer science and exchange knowledge and ideas. ‘David James Sherman, from the Pleiade project team, helped us a lot with our project by guiding us towards key people for our development,’ explains Titien Cubilier.
Dual support from the Bernard Magrez Start-Up Win incubator and Unitec
Other partnerships will be launched in 2026, as Vitidrone will reach a new milestone at the beginning of the year: it will join the Primeur programme, jointly run by the Bernard Magrez Start-up Win incubator, which specialises in viticulture and winemaking, and the UNITEC incubator, the leading incubator in the Bordeaux region. The latter, with its extensive expertise in agri-food and wine-growing businesses, will complement and consolidate the structure provided by Inria Startup Studio.
In addition to premises, these two incubators will provide support for the development of the company and, above all, a network of more than 40 Bernard Magrez châteaux for future experiments. ‘This is a wonderful opportunity to forge new technical partnerships!’ enthuses Théo Gauvrit. ‘We will be able to benefit from a wealth of diverse feedback, while taking advantage of the tremendous publicity that comes with joining the incubator.’
Coming soon to market
Over the coming year, the two partners will continue to add features to their technology, for example to try to detect mildew at a very early stage. They will also work with INRAE in Bordeaux to identify other areas where drones could help winegrowers. The ambitious start-up Vitidrone hopes to bring its solution to market by the end of the year.
