SPRIND Challenge: Recoding Medicine

Funding for interdisciplinary teams that develop AI-ready health datasets and clinically relevant AI applications that unlock the value of European health data.

Bundesagentur für Sprunginnovationen (SPRIND – German Federal Agency for Breakthrough Innovation) acts as a forum between the German government, industry, and research representatives. SPRIND finances international research and development projects for ground-breaking technologies and topics that have the potential to fundamentally improve people’s lives.

The aim of the SPRIND Challenge: Recoding Medicine is to unlock Europe’s fragmented health data for artificial intelligence (AI) innovation by supporting the creation of high-quality, AI-ready health datasets and demonstrating their value through clinically relevant AI applications. Teams are expected to curate at least one unique health dataset of high clinical or societal relevance and use it to develop AI solutions such as biomarkers, diagnostics, clinical decision support systems or drug repurposing tools. Projects must also generate a documented public contribution, such as curated datasets, benchmark suites, annotation tools, methods pipelines or other reusable resources, while ensuring compliance with data protection and governance requirements.

Funded projects should create high-quality, machine-readable datasets, establish reproducible evaluation pipelines and demonstrate the translational or commercial value of their AI applications. In addition to developing innovative AI tools, projects are expected to lay the foundations for sustainable exploitation through product development, regulatory planning, industry collaboration or venture creation.

Projects should combine expertise across disciplines, bringing together data holders and AI specialists from universities, research organisations, healthcare institutions, startups or industry to address clinically relevant challenges using previously inaccessible health data.

The programme is divided into three stages:

• Stage 1: Data curation and model training.
• Stage 2: Validation of specific clinical or translational use cases.
• Stage 3: Transfer, commercialisation and venture building.

After each stage, projects are evaluated against predefined milestones, with only the highest-performing teams progressing to the next stage.

The programme takes the form of a competitive challenge, with the number of participating teams decreasing at each stage as follows:

• Up to 12 teams in Stage 1.
• Up to 7 teams in Stage 2.
• Up to 4 teams in Stage 3.

At the end of Stage 3, the winning team or teams will be announced.