Enhanced Learning Strategies for General Purpose AI: Advancing GenAI4EU (RIA) (AI/Data/Robotics Partnership)
Quick Facts
Email me updates on this grant
Get notified about:
- Deadline changes
- New FAQs & guidance
- Call reopened
- Q&A webinars
We'll only email you about this specific grant. Unsubscribe anytime. No spam.
Ready to Apply?
Get a personalized assessment of your eligibility and application strategy
See in 5 min if you're eligible for Enhanced Learning Strategies for General Purpose AI: Advancing GenAI4EU (RIA) (AI/Data/Robotics Partnership) offering max €30.0M funding💰 Funding Details
Enhanced Learning Strategies for General Purpose AI: Advancing GenAI4EU (RIA)
What the Grant Funds
* Purpose: Cutting-edge Research & Innovation Actions (RIA) that design, train and validate next-generation General Purpose AI (GPAI) models combining self-supervised learning with hybrid, reinforcement, continual, evolutionary, relational, active or physics-based paradigms.
* Activities Eligible for Funding
* Fundamental research delivering new algorithms, architectures, theoretical insights.
* Development, training and benchmarking of large-scale GPAI models.
* Creation or curation of datasets, simulators or testbeds required for the above (respecting EU data rules).
* Integration of explainability, safety, energy-efficiency and alignment with the AI Act.
* Demonstration pilots in priority EU sectors (robotics, health, mobility, manufacturing, energy, science, etc.).
* Open science assets: publication, open-sourcing components, contribution to AI-on-Demand platform.
* Coordination with ADRA partnership and the GenAI4EU Central Hub CSA.
Financial Framework
* Total EU contribution per project: up to €30 000 000 (100 % of eligible direct costs + 25 % flat-rate indirect costs).
* Type of Action: Horizon Europe RIA (budget-based, single grant agreement).
* Project Duration Guidance: 36–48 months typical for large RIA.
* Co-funding Requirements: None for eligible entities; third-country participants might need own funding unless listed as associated.
Key Dates
| Stage | Date |
|-------|------|
| Call opens | 10 Jun 2025 |
| Proposal deadline (single-stage) | 02 Oct 2025 – 17:00 Brussels |
| Evaluation results | ~Jan 2026 |
| GA signature | ~Apr 2026 |
| Earliest project start | Q2 2026 |
Eligibility Snapshot
* Consortium minimum: 3 independent legal entities from 3 different EU Member States/Associated Countries (MS/AC).
* Geographical restriction: Only entities established in MS, Iceland, Norway, Canada, Israel, Korea, New Zealand, Switzerland, United Kingdom (plus guarantees if under foreign control). High-risk 5G suppliers excluded.
* Target profile: Interdisciplinary mix of AI research labs, HPC/cloud providers, SME innovators, industrial end-users, social science & humanities (SSH) partners, standardisation/XAI specialists.
* Exclusion & capacity checks: Financial/operational capacity, ethics, security scrutiny (especially for GPU clusters, data transfer, dual-use concerns).
Mandatory Policy Requirements
* Alignment with AI Act, GDPR, open science and FAIR data.
* Contribution to ADRA partnership coordination tasks (work package & budget line expected).
* Results, code or models shared on AI-on-Demand platform unless duly justified.
* Energy-efficiency monitoring & carbon-footprint reporting.
* Gender equality plan for each public body, research organisation or HEI in the consortium.
Typical Budget Breakdown (indicative)
* 45 % – Personnel (researchers, engineers, ethics experts)
* 25 % – Computing infrastructure & energy costs (HPC/GPU cloud, optimisation for efficiency)
* 10 % – Data acquisition/annotation, testbeds, benchmarking campaigns
* 8 % – Pilots & demonstrators with end-users
* 5 % – Open science, dissemination, standardisation
* 4 % – ADRA & GenAI4EU hub liaison, policy alignment
* 3 % – Project management & quality assurance
Evaluation Criteria Highlights
1. Excellence (50 %) – Novelty of learning approach, scientific ambition, sound methodology.
2. Impact (30 %) – Industrial relevance, openness, contribution to EU autonomy & climate goals.
3. Quality of Implementation (20 %) – Work plan, resources, risk mitigation, consortium competence.
Thresholds: ≥4/5 on Excellence & Impact, ≥12/15 overall.
---
📊 At a Glance
Get Grant Updates
Get notified about:
- Deadline changes
- New FAQs & guidance
- Call reopened
- Q&A webinars
We'll only email you about this specific grant. Unsubscribe anytime. No spam.
🇪🇺 Strategic Advantages
EU-Wide Advantages & Opportunities for “Enhanced Learning Strategies for General Purpose AI: Advancing GenAI4EU” (HORIZON-CL4-2025-04-DIGITAL-EMERGING-07)
1. Why This Topic Is Intrinsically European
• Open Strategic Autonomy: Builds sovereign GPAI capabilities that reduce dependence on extra-EU foundation models, fully in line with Destination 4 and the *Apply AI* strategy.
• Single Digital Market Leverage: 450 million citizens and 25 million enterprises give unrivalled diversity of languages, use-cases and edge conditions—ideal for testing adaptability, reasoning and robustness.
• Regulatory First-Mover Advantage: Alignment with the new AI Act allows consortia to pilot *trust-by-design* GPAI systems, creating exportable compliance know-how.
2. Unique EU-Scale Assets to Exploit
1. Pan-European Data Spaces
• Health, Manufacturing, Mobility, Green Deal, Cultural Heritage—provide legally interoperable, high-quality and multilingual datasets for self-/active learning.
• Synergies with DEP Large Language Models and Common European Data Spaces maximise data availability while respecting GDPR.
2. EuroHPC & Edge-to-Cloud Fabric
• Pre-exascale machines (LUMI, LEONARDO) and upcoming exascale JUPITER offer subsidised compute hours; combined with sovereign edge/cloud stacks (GAIA-X, Eclipse EUCLID) to test continual & evolutionary learning in federated settings.
3. AI-on-Demand Platform & ADRA Partnership
• Ready-made dissemination, benchmarking and asset-sharing channels; guarantees EU brain circulation and community uptake.
4. Testing & Experimentation Facilities (TEFs)
• Robotics, Manufacturing and Health TEFs provide realistic, regulatory-grade sandboxes for reinforcement learning and XAI validation.
5. Living Labs & EDIHs
• 200+ European Digital Innovation Hubs connect SMEs/regions, broadening pilot domains and boosting technology diffusion.
3. Cross-Border Impact Opportunities
3.1 Industrial Leadership
• Manufacturing 4.0: GPAI-driven adaptive control and predictive maintenance across EU supply chains; links to *Made in Europe* partnership.
• Advanced Robotics: Neuro-symbolic planners for collaborative robots, leveraging ELIZA TEF and EuroC-XYZ robotics testbeds.
3.2 Societal Value
• Health & Care: Continual-learning GPAI for personalised treatment pathways tested on federated hospital data (e.g., TEF-Health).
• Climate Neutrality: Physics-informed GPAI accelerates material discovery and energy-efficient process optimisation; aligns with EERA & Green Deal missions.
3.3 Scientific Excellence
• Open Science Leadership: Mandatory FAIR data/code release via AI-on-Demand raises EU citation share and attracts global talent.
4. Consortium Design Advantages
• Geographic Coverage: Min. 3 MS/AC; maximise impact by including cohesion countries for “widening” bonus points.
• Interdisciplinarity Mandate: Combine AI labs, HPC centres, ethics scholars, TEF operators, SMEs and standardisation bodies (CEN-CENELEC, ISO/IEC JTC 1/SC 42).
• Security & Autonomy Compliance: EU-controlled entities avoid dependency clauses; eligibility restriction is an opportunity to structure an EU-centred IP portfolio.
5. Programme & Funding Synergies
• Digital Europe Programme (DEP): Re-use compute vouchers, datasets and Trustworthy AI testing services.
• Chips Act & KDT JU: Collaborate on neuromorphic or in-memory accelerators for energy-efficient GPAI.
• EIT Digital & EIT Manufacturing: Commercialisation pathways and scale-ups.
6. EU-Added-Value KPIs (sample)
• ≥ 6 cross-border pilots covering ≥ 4 official EU languages.
• ≥ 50 % energy-efficiency gain vs. state-of-the-art GPAI on EuroHPC benchmarks.
• Compliance toolkit reducing AI-Act conformity assessment effort by ≥ 30 % for adopters.
• Open-sourced models/datasets/tools downloaded ≥ 100 k times via AI-on-Demand within 12 months post-project.
7. Risk Mitigation & Ethical Leadership at EU Level
• Federated Governance: GDPR-compliant federated learning avoids data transfer issues.
• Human-Centric Co-Creation: Involve European social partners and civil society via *Conference on the Future of Europe* methodologies.
• Standardisation Track: Feed results into CEN-CENELEC/ETSI working groups to shape global norms.
8. Competitive Edge for Applicants
1. Access to Non-Dilutive Funding: Up to ~€4–6 m per project with 100 % direct cost reimbursement + 25 % overhead—more generous than national schemes.
2. Brand Value: Horizon Europe label facilitates later EIC Accel, InvestEU and VC leverage.
3. Early Market Creation: Trusted EU-compliant GPAI positions consortium for forthcoming public procurement waves under AI liability directives.
Bottom Line: Operating at EU scale unlocks unparalleled datasets, compute, regulatory alignment and market reach, enabling consortia to tackle GPAI’s grand challenges while cementing Europe’s technological sovereignty and ethical leadership in the next wave of artificial intelligence.
🏷️ Keywords
Ready to Apply?
Get a personalized assessment of your eligibility and application strategy
See in 5 min if you're eligible for Enhanced Learning Strategies for General Purpose AI: Advancing GenAI4EU (RIA) (AI/Data/Robotics Partnership) offering max €30.0M funding