AI-BOOST launches a European competition to advance generative AI

The European AI-BOOST project, coordinated by Zabala Innovation and focused on promoting open innovation in artificial intelligence, is launching the AI Challenge Competition, a competition aimed at startups, SMEs, researchers and other innovators seeking to develop generative AI solutions to address specific technology challenges.

The challenges in AI-BOOST’s AI Challenge Competition

The new competition is structured around four challenges.

Challenge 1 – GenAI-based natural language mission generator for autonomous robots in agriculture

The first challenge, led by Consorzio Intellimech and JOiINT LAB, aims to develop a natural language mission generator for autonomous robots in agriculture. The objective is to democratise robotic programming through generative AI, so that instructions expressed in natural language can be translated into executable commands. The proposal is based on the use of advanced Vision-Language-Action (VLA) models to create a modular Proof of Concept that helps bridge the gap between human intent and robotic execution. Although the initial focus is on enhancing productivity in vineyards, the challenge has a broader ambition: to facilitate intelligent automation in industrial maintenance and field operations.

Application deadline: 25th August.

Challenge 2 – Agentic AI for automated CAD generation and autonomous simulation

The second challenge, led by SIAD Group, focuses on automated CAD generation and autonomous simulation through agentic AI. In this case, the challenge is to streamline the design and simulation of industrial piping, in particular for routing compressor tubes within skids. The solution will need to translate natural language requirements and existing CAD data into editable parametric models, with the aim of automating complex tasks such as extracting geometric constraints, mesh generation and convergence analysis. The expected result is a significant reduction in manual engineering effort, with more compact designs, lower costs and validation against stringent industrial standards.

Application deadline: 25th August.

Challenge 3 – Generative AI for enhancement of clinical datasets

The third challenge, led by the European Federation for Cancer Images (EUCAIM), focuses on using generative AI to enhance the quality and representativeness of clinical imaging datasets, which are often incomplete or imbalanced. By creating realistic synthetic patient cohorts, the solution aims to fill critical data gaps and harmonise information across different imaging conditions. The system will need to identify key demographic and clinical characteristics to ensure that synthetic data remains realistic and consistent with real-world statistics. Ultimately, the project aims to improve fairness and reduce bias, enabling the development of trustworthy models that support medical research and clinical decision-making more accurately.

Application deadline: 8th September.

Challenge 4 – Generative AI for automatic test case generation from crash databases and standards

The fourth challenge, led by Siemens Industry Software NV in collaboration with the EU RobustifAI project, aims to enhance the safety and validation of autonomous driving systems by using generative AI to automate the creation of simulation scenarios. By transforming accident reports, visual data and international safety standards into structured, simulation-ready formats, the solution replaces slow manual processes with a more scalable and consistent workflow. This approach strengthens the link between real-world data and regulatory frameworks, helping teams identify critical safety gaps and complex edge cases. Ultimately, the project will provide technical and regulatory experts with comprehensive scenario sets, significantly improving the efficiency and accessibility of safety assessments for autonomous vehicles.

Application deadline: 8th September.

The four challenges will be explained in greater detail in the webinar organised by Zabala Innovation on 22nd July at 11:00.

Key features of AI-BOOST’s AI Challenge Competition

Phases and funding

The competition will take place in two phases. In the initial phase, known as the Spark Phase, participants will work for two months on a concept note with a technology readiness level (TRL) of 1-2, followed by one month of evaluation. Five winners will be selected from this stage and will each receive €28,500. The second stage, the Advance Phase, will focus on algorithm development and validation, with the objective of reaching TRL 4-5 over five months.

The phase will conclude with a final live competition event, expected to take place in Brussels in February 2027. During this event, participants will publicly present and demonstrate their solutions before a panel of evaluators, Challenge Owners, representatives of the AI-BOOST consortium and other stakeholders. The live demonstration will allow participants to show the functionality, performance and practical relevance of their solution in relation to the challenge objectives. The final event will also include a public pitch session and an audience vote.

The final winner will receive €100,000. In addition, the competition includes Special Awards worth €25,000 per challenge, divided into an Innovation Excellence Award of €12,500 and a Responsible AI Award of a further €12,500.

Other benefits

Beyond funding, the AI Challenge Competition offers visibility and technical guidance from the project partners and the organisations that have defined the challenges. Its logic reflects the approach AI-BOOST has been building since 2023: turning specific needs into open innovation opportunities and connecting those developing solutions with companies, research centres and European infrastructure.

What AI-BOOST is

Launched in 2023, AI-BOOST aims to establish a challenge programme capable of serving as a benchmark for the European AI community and strengthening the continent’s competitiveness in this strategic field. The project is driven by a consortium made up of Zabala Innovation, F6S, Cineca, INESC TEC, Pavol Jozef Šafárik University in Košice, EIT Digital and Pompeu Fabra University.

Generative AI, which at that time was still presented as a disruptive novelty, is now part of the context in which companies, public administrations and research centres operate. Its opportunities and risks have already been identified: greater capacity to generate content, analyse information and support research processes, but also new requirements in transparency, bias, intellectual property, cybersecurity and human oversight.

In this sense, the project acts as a bridge between technological demand and the innovative capacity of the European ecosystem. The organisations participating in the challenges organised by AI-BOOST do not start from abstract problems, but from needs identified in areas such as industry, health, mobility, robotics and advanced research. From there, the project mobilises applications, provides technical support and steers solutions towards verifiable results.

This approach means that the competition does not merely reward ideas, but supports their evolution towards prototypes, algorithms and models with application potential. It also helps bring order to a rapidly expanding market in which many organisations are looking for specific use cases, validation criteria and partners capable of turning generative AI into useful tools.

“AI-BOOST does not approach this debate from a theoretical standpoint, but from a practical question: how to mobilise the European ecosystem to develop AI applications and models with real impact,” stresses Daniel Errea, head of the Digital area at Zabala Innovation and project coordinator. “That is why AI-BOOST is looking for innovators to solve the four high-impact challenges through generative AI,” Errea explains.

The strength of European supercomputing

One of the project’s distinguishing features is its connection with European supercomputing. The development of advanced AI models requires very high computing capacity, and AI-BOOST has placed this issue at the heart of some of its challenges. In the Large AI Grand Challenge, launched in 2023, the selected startups each received €250,000 and gained access to EuroHPC JU resources, including LUMI and LEONARDO, to develop large-scale AI models.

The award ceremony held in Brussels, with the participation of then European Commissioner Thierry Breton, was one of the project’s most visible moments. For AI-BOOST, it represented the launch of a European mechanism to identify promising projects, fund them and provide computing infrastructure under competitive conditions.

Europe looks to the future of AI

Since then, the project has continued to evolve. In 2026, the Frontier AI Grand Challenge has once again focused on Europe’s capacity to develop advanced models supported by its own infrastructure. The European Commission selected the EUROPA consortium as the winner, led by the Italian company Domyn together with Fraunhofer IAIS and Fraunhofer IIS, with the aim of developing an open frontier model covering the 24 official languages of the European Union. For Leyre Ansorena, senior consultant at Zabala Innovation, “this type of progress shows that AI-BOOST is part of a broader debate on European technological sovereignty and strengthens collaboration with the artificial intelligence community”.

Looking ahead to 2027, “the challenge is to consolidate the ecosystem created and demonstrate that open AI competitions are an effective tool to accelerate innovation and transform research into solutions with real impact,” says Sara López, consultant in Zabala Innovation’s Entrepreneurship area. “The project is therefore entering its final stretch not as a technology showcase, but as an instrument to strengthen European competitiveness in artificial intelligence,” she concludes.