Funder: European Commission
Due Dates: March 18, 2027 (Full application, single-stage)
Funding Amounts: €44 million total; one grant expected per call; single-stage RIA; project budgets expected to match scope and ambition.
Summary: Supports development of safe, computationally efficient frontier AI models in Europe, prioritizing both advanced capabilities and robust safety by design.
Key Information: Participation restricted to EU, select associated countries, and entities not controlled by non-eligible countries.
Description
This opportunity funds research and development of sovereign frontier AI models in Europe, focusing on both safety by design and computational efficiency. The aim is to advance state-of-the-art, multimodal frontier AI models optimized for agentic capabilities (such as tool use, reasoning, and autonomous problem-solving), while also addressing the energy and computational demands of large AI systems. Projects are expected to develop robust safety mechanisms, including automated testing, interpretability, detection and mitigation of misalignment, adversarial robustness, and transparency. Proposals should foster collaboration within the European AI research community, leverage existing EU resources (e.g., RAISE, AI Factories, Data Labs), and share results with the broader R&D community. The initiative is part of the EU’s Apply AI Strategy and the Frontier AI Initiative.
