Over 48 hours, 12 teams turned real-world problem scenarios into prototypes and technology solutions addressing seven challenges in areas including artificial intelligence and machine learning, HealthTech, cybersecurity, AgriFood, Smart City and GeoTech, as well as an Open Challenge that allowed participants to develop their own ideas.
This year’s edition brought together more than 400 participants, mentors, jury members, organisers and partners, including specialists from Romania, Georgia, Armenia and Bosnia and Herzegovina, and delivered approximately 209 hours of mentoring.
The awards ceremony was attended by public officials, representatives of partner companies and external partners supporting the development of the national innovation and technology sector.
Dominique Waag, Ambassador of France to the Republic of Moldova:
“These days have been about innovation, creativity and cooperation, but above all about the courage to test new ideas and build together. I encourage you to keep going, believe in your ideas, remain curious and continue creating. What I see here shows me that the Republic of Moldova is increasingly becoming a place where innovation takes shape, and that is impressive.”
Stella Jemna, Country Representative of the Ukraine-Moldova American Enterprise Fund (UMAEF):
“A hackathon does not end with the awards ceremony. Its real value comes when an idea tested here continues to be developed, reaches users and becomes a product or a business. We want the teams to use this experience, the mentors and the resources available in the ecosystem to take the next step and build projects that can grow from here.”
Cristian Cerneanu, Startup Ecosystem Manager at Innovate Moldova:
“We saw some very strong ideas, and from one edition to the next, we can clearly see an improvement in the quality of the solutions being developed. Some of the projects presented here have real potential to move forward. I encourage participants not to stop at the prototype they built during these days, but to use the resources, programmes and connections available within the ecosystem to turn their ideas into products and companies.”
At the same time, Technovator, the organiser of the competition, announced the launch of a programme dedicated to selected teams, providing them with support, mentoring and resources to further develop their solutions and turn the prototypes created during the competition into tested and validated products.
Solutions developed in 48 hours: AI for sales, digital agriculture and more
The challenges at Deeptech GigaHack 2026 were based on concrete problems faced by companies and institutions. Teams had 48 hours to develop a solution, test it and demonstrate how it could work in practice.
LeadRadar: Artificial intelligence for identifying business opportunities in public information
Team MacLovin developed a tool designed for sales teams. LeadRadar can turn the vast amount of publicly available information about companies into a clear list of potential customers, replacing hours of manual research.
The tool also provides references to the sources from which the information was collected, allowing it to be independently verified. Once a potential customer has been identified, LeadRadar can suggest a personalised message for email or LinkedIn based on the information available about the company.
Users can also subscribe to updates to be notified when new relevant information becomes available, while the data can be integrated into a CRM system.
Bloomsense: Drones and AI for faster plantation inspections
Bloomsense developed a solution that automatically turns drone imagery into a map of problems that need to be checked in a plantation. The technology analyses the images and can identify rows of vines and gaps between plants that require intervention.
Instead of requiring an inspector or farmer to cover the entire area looking for problems, the system shows exactly where they need to go and identifies the optimal route.
During the hackathon, the solution automatically processed 311 images in approximately seven minutes, analysing 53 blocks and 792 rows, equivalent to more than 50 kilometres of plantations. According to the results presented by the team, using the system could reduce the time required for a field inspection by approximately 72%.
To make the technology accessible to farmers who may not want to use a complex platform, the team also developed a Telegram bot. The user enters the plantation, and the system provides the map, route and specific points that need to be checked.
FOLDBACK: Packaging that folds after use and can re-enter the recycling stream
DreamTeam addressed a seemingly simple but challenging recycling problem: what happens to the plastic packaging used by retailers to sell freshly prepared food?
The solution, called FOLDBACK, focuses on packaging for rotisserie chicken, a product that is difficult to package using alternative materials because of its temperature and grease content.
The team proposed a polypropylene container that, after the food has been consumed, can be folded into a compact cylinder and returned through existing infrastructure in EU stores designed to collect plastic packaging.
This would prevent the container from being mixed with other plastic waste and allow it to form a separate recyclable material stream. DreamTeam’s proposed objective is to reduce the amount of plastic used per serving by approximately 50%.
FIR: A single point of contact between citizens and the municipality
NOV Team created FIR (Find, Inform, Resolve) – Fiecare Întrebare Rezolvată, a digital assistant designed to simplify interaction between citizens and municipal authorities.
The system searches exclusively through official sources and identifies the exact document and passage from which an answer is taken. If the required information is not available, it directs the user to the responsible institution.
The prototype, available in Romanian and Russian, was built using 266 citizen questions and 110 types of issues as its basis. For petitions, the solution can identify the responsible department, incorporate the relevant legal deadline and allow users to track the request through the municipal administration.
The solution was designed for a system covering municipal departments, five sectors, 18 suburban mayors’ offices and hundreds of subordinate institutions.
Notavra: Less paperwork for doctors, more time for patients
The Horalix team from Bosnia and Herzegovina started from a problem doctors encounter every day: a significant share of their time is spent not talking to patients, but transcribing information and completing documentation.
Their solution, Notavra, can automatically transcribe medical conversations and, more importantly, understand what was decided during the discussion: what the next steps are, who is responsible and what needs to happen afterwards.
One of the challenges addressed by the team is distinguishing between what was merely proposed and what was ultimately decided. If one date for a procedure is suggested during a conversation and a different proposal is made later, the system preserves the context and attempts to distinguish between the alternatives and the final decision.
Another important feature is privacy: processing can take place locally, without medical data being transmitted to the cloud.
During the demonstration at the hackathon, an 11-minute and 42-second recording was processed in 6.2 seconds, including in a conversation that switched between languages.
VOX: Affordable sensors that can “hear” and locate a drone
Brainstormers presented VOX, a solution that transforms affordable acoustic sensors into an early-warning system for drone detection.
The concept is based on sound: the sensors “listen” to the environment, identify the acoustic signature of a drone and attempt to determine its position and trajectory.
The system can then alert a person in the risk area and provide an indication of the distance to the drone and the estimated time until it approaches.
The solution was designed to operate with simple infrastructure: a single server, no cloud and no costly specialised equipment. The project therefore explores the possibility of building an early-warning system that could be deployed more easily and at a lower cost than existing complex solutions.
Cybersecurity: Critical infrastructure tested in a scenario close to a real-world attack
The cybersecurity challenge was structured differently from the other competitions. Instead of developing a conventional product, participants took part in a Red Team vs. Blue Team exercise designed to simulate attacks on critical infrastructure.
The Red Teams attempted to identify and exploit vulnerabilities in the infrastructure made available through the Cyber Range, while the Blue Team specialists were responsible for detecting and mitigating the attacks.
One of the participating teams was F-Society, which documented the actions carried out against the infrastructure and the vulnerabilities identified.
The challenge focused not only on identifying security gaps, but also on testing participants’ ability to work in a scenario close to the real-world situations faced by cybersecurity professionals.
For some of the projects created at Deeptech GigaHack, the competition is only the first step. The organisers announced that 15 teams will continue to receive mentoring and support to further develop and test their solutions and connect with potential partners. Other teams will have access to programmes that can help them turn the prototypes developed during the hackathon into viable products.
Deeptech GigaHack 2026 was organised by Technovator with the support of EU4Innovation East, a regional project funded by the European Union, co-funded by the Government of France and implemented by Expertise France; the Ukraine-Moldova American Enterprise Fund (UMAEF); and the Innovate Moldova Programme, funded by Sweden and the United Kingdom, as well as the Central European Initiative (CEI) through the CEI Know-how Exchange Programme 2025, funded by the Ministry of Foreign Affairs and International Cooperation of the Italian Republic.
