EU Frontier AI Initiative: Advanced Scaling Frameworks for High-Performance AI Models (RIA) - Apply AI
Overview
Horizon Europe Bridging actions 2027 topic HORIZON-BRIDGING (EU Frontier AI Initiative: Advanced Scaling Frameworks for High-Performance AI Models) is a single-stage Research and Innovation Action with an indicative budget of €15 million for one grant, planned to open 20 October 2026 with a submission deadline of 16 February 2027 at 17:00 Brussels time. The call funds research to develop predictive scaling laws and practical toolsets for model scaling, including automated scaling law derivation, architecture search, dataset mixture optimisation, advanced hyperparameter tuning, stability and cross-hardware validation to improve training and inference efficiency. Eligible applicants are legal entities established in EU Member States, Iceland, Norway and specified associated countries, subject to ownership and control restrictions and required declarations. Projects must demonstrate TRL 4–5 at completion, include quantitative KPIs and benchmarking, and allocate activities for cohesion with AI Factories, Data Labs, ADRA and the GenAI4EU hub.
Partner Search
Find collaboration partners for this call
What You Offer
Describe your expertise here...
You Are Looking For
Describe what you seek here...
Highlights
Call at a glance
Essential facts
What it funds:Research and Innovation Action to develop scientific foundations, predictive scaling laws, automated scaling toolsets and methodologies for efficient training and inference of frontier AI models. Projects should deliver an implementable toolset and validate methods with advanced AI labs and HPC centres; TRL expected 4-5 at project end.
Funding instrument and budget:Type of action: HORIZON-RIA (Research & Innovation Action). Indicative EU contribution: around €15 000 000 for one grant under this topic. Apply via the topic page Topic page. 1
- 1Automated scaling law derivation and predictive frameworks
- 2Exploratory architecture search for scalable, inference-efficient designs
- 3Dataset mixture optimisation and automated data quality assessment
- 4Advanced hyperparameter tuning and stable large-scale training
Who can apply
Eligible applicants:legal entities established in EU Member States, Iceland, Norway and the following associated countries: Canada, Israel, Republic of Korea, New Zealand, United Kingdom, Switzerland and Japan. Entities controlled (directly or indirectly) by ineligible third countries are restricted; entities controlled by China cannot provide guarantees to participate. See call conditions for ownership and control requirements and declaration obligations.
| Planned opening | 20 October 2026 |
|---|---|
| Deadline (Brussels time) | 16 February 2027, 17:00 |
| Expected grant size | Around €15 000 000 (one grant) |
Submission:single-stage. Proposals must include KPIs, benchmarks, linkages to other Apply AI topics and allocation for cohesion activities with ADRA and GenAI4EU hub. Read the topic page and General Annexes for admissibility, evaluation and ownership/control procedures.
Footnotes
- 1List of Participating Countries in Horizon Europe and details on association and participation rules are available at ec.europa.eu
Find a Consultant to Support You
Breakdown
Call and Administrative Overview
Call title Bridging actions 2027. Call identifier HORIZON-BRIDGING. Programme Horizon Europe (HORIZON). Type of action HORIZON-RIA Research and Innovation Actions. Type of grant HORIZON Action Grant Budget-Based (HORIZON-AG). Single-stage submission procedure. Planned opening date 20 October 2026. Deadline 16 February 2027, 17:00:00 Brussels time. Submission through the Funding & Tenders Portal. Indicative number of grants expected: 1. Expected EU contribution per project: around €15,000,000. Topic is part of the Frontier AI Initiative and complements related Apply AI topics. Activities expected to reach TRL 4-5 by the end of the project (see General Annex B).
Primary source:Topic description and administrative conditions published on the EU Funding & Tenders Portal topic page Topic details 1
Scope, Objectives and Expected Outcomes
This Research and Innovation Action aims to advance the scientific and technical foundations enabling predictable and efficient scaling of high-performance AI models for training and inference. The topic focuses on methodologies, theory, and toolsets to support European AI labs to compete at the global frontier while preserving cost-efficiency, reproducibility and hardware diversity compatibility.
- 1Expected outcomes include: establishment of robust scaling laws and predictive frameworks to reliably estimate model performance before large-scale compute investment; enhanced competitiveness of European advanced AI models through optimised dataset mixtures, refined architectures and improved hyperparameter configurations; and optimised development of frontier AI models via automated, scientifically grounded scaling methodologies.
- 2Scope emphasises provision of a toolset supporting scaling of AI models, covering automated derivation of scaling laws, exploratory architecture search, dataset mixture optimisation, advanced hyperparameter tuning and stability, and mechanisms to validate scaling across diverse hardware architectures.
- 3Proposals must demonstrate active support to Advanced AI labs and High-performance Computing Centres and describe benefits for strategic sectors such as Public Administration, Manufacturing, Energy, Healthcare, Cultural and Creative Industries, and Physical AI.
- 4Proposals must describe links to the Apply AI topic HORIZON-CL4-2027-04-DIGITAL-EMERGING-11 and allocate cohesion tasks with the European Partnership on AI, data and robotics (ADRA) and the GenAI4EU central Hub. Communicable results should be shared through the AI-on-demand platform and links established with AI Factories and Data Labs where relevant.
Recommended Research Areas (non-exhaustive)
- 1Automated Scaling Law Derivation: theory and empirical methods to predict performance across modalities and tasks, moving beyond simple power laws.
- 2Exploratory Architecture Search: automated discovery and refinement of architectures that preserve or improve performance when scaling for training and inference, including sparsity, memory efficiency and long-context handling.
- 3Dataset Mixture Optimisation: systematic selection, weighting and sequencing of datasets, including automated data quality assessment and metrics for mixture efficacy.
- 4Advanced Hyperparameter Tuning and Stability: methodologies for stable training at very large scales, automated and robust tuning of optimisers and learning rate schedules, and techniques to improve training stability as models grow.
- 5Cross-hardware Compatibility: ensure frameworks and methodologies are compatible with diverse hardware architectures and validated across different systems to avoid optimisation for single proprietary hardware.
Eligibility and Participation Conditions
Participation is limited to legal entities established in EU Member States, Iceland and Norway and the following associated countries: Canada, Israel, Republic of Korea, New Zealand, United Kingdom, Switzerland and Japan. Entities established in third countries that become associated to Horizon Europe during 2026–2027 may be eligible if identified in the List of Participating Countries at the time of submission. The association to the Programme must apply by the time of signature of the grant agreement. Entities established in eligible countries but directly or indirectly controlled by a non-eligible country or entity may be ineligible unless guarantees are provided and positively assessed by the eligible country of establishment. Entities established in an eligible country but controlled by China or by a legal entity established in China cannot submit guarantees for this topic.
Admissibility conditions (page limits, layout) are described in Annex A and Annex E of the Work Programme General Annexes and Part B of the Application Form in the Submission System. Financial and operational capacity, exclusion and evaluation award criteria are described in General Annexes C and D. Legal and financial set-up of grants in Annex G. Specific call-related templates and application forms are in the Submission System (standard HE RIA application form and standard evaluation forms). Ownership Control Declaration and other call-specific documents are mandatory where indicated.
Restrictions and Ownership/Control requirements:To protect the Union’s strategic assets and autonomy, the call includes ownership and control restrictions. Entities controlled directly or indirectly by non-eligible countries may not participate unless they can provide guarantees approved by competent national authorities demonstrating that their participation would not negatively impact EU strategic assets, interests, autonomy or security. Entities controlled by China cannot submit guarantees for this topic. Participants must complete the Ownership Control Declaration and, if requested, supply supporting documents to the Central Validation Service or the relevant EU service. If guarantees are permitted and used, they must be approved by the competent national authority and accepted by the EU service in charge of the grant.
Technical Readiness, Impact and Monitoring
Expected Technology Readiness Level of results:TRL 4-5 at project completion (see General Annex B). Proposals must include mechanisms for assessing and demonstrating progress, including qualitative and quantitative KPIs, benchmarking, and progress monitoring. Where possible proposals should build on and reuse public results from previous funded actions. Communicable results should be shared via the AI-on-demand platform. Projects should link to AI Factories, Data Labs and explore pre-competitive exchanges with European companies developing frontier AI models.
Funding, Budget and Financial Modalities
Indicative budget for this topic:around €15,000,000 for one grant. The grant will follow the Horizon Europe Rules and the HORIZON Action Grant Budget-Based Model Grant Agreement (HORIZON-AG). Applicants must consult General Annexes for financial and budget rules, including admissible costs, personnel, subcontracting, financial support to third parties where applicable, and reporting procedures.
| Budget item | Detail |
|---|---|
| Indicative EU contribution | Around €15,000,000 (one grant expected) |
| Number of expected grants | 1 |
| Type of funding | Grant (Horizon Europe Research and Innovation Action) |
| Grant model | HORIZON Action Grant Budget-Based (HORIZON-AG) |
Applicant and Consortium Requirements
Eligible applicant types:legal entities such as universities and higher education establishments, research institutes, public bodies, industry players including SMEs and large enterprises, non-profit organisations, HPC centres and advanced AI labs. The topic expects consortia that include advanced AI labs and High-Performance Computing Centres for validation in high-stakes environments. The call text encourages industry-led consortia, participation of SMEs, and clear allocation of cohesion activities with ADRA and the GenAI4EU hub. Public bodies validated in the Participant Register are automatically considered controlled by their country and are exempt from the Ownership Control Declaration requirement.
- 1Eligible Applicant Types: university, research institute, SME, large enterprise, nonprofit, public body, HPC centre, AI labs, consortiums including mixed academia and industry. Individual applicants are not applicable as beneficiaries; legal entities only.
- 2Consortium Requirement: single-stage submission but consortium-based proposals are expected and recommended. The topic expects a consortium; a single legal entity may apply where permitted by the general rules, but the scope and ambition suggest multi-partner consortia involving labs and HPC centres.
- 3Geographic Eligibility: legal entities established in EU Member States, Iceland and Norway and the associated countries explicitly listed: Canada, Israel, Republic of Korea, New Zealand, United Kingdom, Switzerland and Japan; plus additional third countries if associated to Horizon Europe during 2026–2027 and listed at the time of submission.
- 4Target Sector: AI, high-performance computing, machine learning, software services, advanced algorithms, DeepTech with applications across Public Administration, Manufacturing, Energy, Healthcare, Cultural and Creative Industries, Physical AI and multi-modal scientific domains.
- 5Project Stage: research and innovation activities expected to reach TRL 4-5 (development, validation and early demonstration of tools/frameworks).
- 6Funding Type: grant (Research and Innovation Action).
- 7Nature of Support: monetary funding in the form of Horizon grants; projects must produce public and communicable research outputs, tooling and frameworks.
- 8Application Type: open single-stage call (planned opening and fixed deadline); submission via Funding & Tenders Portal open call.
- 9Application Stages: single-stage submission; however evaluation and grant preparation include admissibility, eligibility, evaluation and possible ownership control assessment steps during grant preparation.
Evaluation, Award and Administrative Process
Evaluation criteria, scoring and thresholds follow General Annex D of the Horizon Europe Work Programme. Submission and evaluation processes are described in Annex F and the Funding & Tenders Portal Online Manual. The topic uses the standard HE RIA evaluation form with necessary adaptations. Indicative evaluation timeline and grant agreement procedures follow Annex F. Applicants must provide all required Annexes and supporting documents; failure to provide mandatory ownership/control documents where required may render the proposal inadmissible or ineligible.
Ownership Control, Guarantees and Restricted Participation:The call includes ownership/control restrictions to protect EU strategic interests. Entities controlled by non-eligible countries may be declared ineligible unless approved guarantees are provided and accepted. Entities established in eligible countries but controlled by China cannot provide guarantees for this call. Applicants must submit the Ownership Control Declaration as part of the proposal and, if requested during grant preparation, upload supporting documents to the Central Validation Service for a detailed ownership control assessment. If a participant is found non-eligible and guarantees are allowed, they must obtain national authority approval of the guarantee and submit it for EU assessment during grant preparation. The guarantee must show concrete measures preventing strategic risks and ensuring IP and results remain within eligible countries and not exported without approval.
Application Templates and Structure
Application form templates:the standard HE RIA application form (Part A and Part B) is available in the Submission System. Part B contains the technical proposal sections (excellence, impact, implementation) with page limits described in Annex A and Part B of the Application Form. For this Destination some Innovation topics had extended page limits for business plan/market strategy; for this RIA normal Part B limits apply but applicants should consult the call-specific Part B instructions in the Submission System. Mandatory annex: Ownership Control Declaration (OWNERSHIP CONTROL DECLARATION annex) must be completed by all beneficiaries and affiliated entities where required by the call text. Other templates and guidance: HE Programme Guide, Model Grant Agreement (HE MGA), Guidance on participation in EU calls with ownership and control restrictions, Rules for Legal Entity Validation and LEAR appointment, Funding & Tenders Portal Online Manual.
- 1Application form structure (standard HE RIA Part B): Excellence (objectives, relation to topic, methodology, scaling research plan, TRL targets), Impact (expected outcomes, dissemination, exploitation, data management plan, sharing via AI-on-demand), Implementation (work packages, deliverables, milestones, management, resources, task allocation to AI labs and HPCs), Ethics and Security, Budget breakdown and justification.
- 2Mandatory annexes to include: Declaration on Ownership and Control (signed), Participant Register entries (PICs) for all participants, Declaration of Financial and Operational Capacity where required, CVs of key researchers, letters of commitment from HPC centres/industry partners where appropriate.
- 3If the proposal will include work with sensitive infrastructure/data or access to classified information, indicate security clearance arrangements and compliance measures in the proposal and guarantee where applicable.
Success Rate, Co-funding and Other Practicalities
Indicative success rates are not published for this specific topic. Historically, competition for single high-value Horizon grants is strong and success rates can be low; applicants should expect strong competition and prepare high-quality, well-justified proposals aligned with evaluation criteria. Co-funding: Horizon Europe grants typically finance a percentage of eligible costs depending on action type; Research and Innovation Actions generally fund up to 100% of eligible direct costs plus a contribution to indirect costs according to the MGA. No separate mandatory co-funding beyond eligible cost sharing is normally requested, but applicants must demonstrate financial and operational capacity to carry out the project and cover non-eligible costs as applicable.
- 1Success rates: not specified for topic; expect high competition for a single-grant call (low success probability overall).
- 2Co-funding requirement: applicants must demonstrate operational and financial capacity; HE RIA typically funds up to 100% of eligible direct costs plus indirect cost contributions as defined in the MGA. No separate mandatory cash co-funding unless otherwise stated; private partners may contribute additional investment as part of exploitation paths.
Application and Evaluation Stages
Submission is single-stage via the Funding & Tenders Portal. The sequence of stages from call to grant signature typically includes: proposal submission, admissibility and eligibility check, remote peer evaluation against the award criteria (excellence, impact, implementation), consensus and ranking, panel/selection review, notification of outcomes, grant preparation and signature including any ownership control assessment and guarantee approval where applicable. Grant preparation may include requests for supporting documents (legal and financial validation, ownership control documents) handled by the Central Validation Service and the EU service in charge of the grant. Applicants should anticipate administrative timeline for OCA and guarantees during grant preparation.
How this Opportunity Supports Strategic Priorities
This topic is part of the EU Frontier AI Initiative and aims to build Europe’s scientific and methodological leadership in scaling high-performance AI models, reducing wasted compute expenditure, increasing predictability and reproducibility of scaling experiments, and supporting Europe’s technological sovereignty in AI model development. It complements training-centric and deployment calls in Apply AI and is expected to feed validated methodologies to Advanced AI labs, AI Factories and Data Labs, and to be shared with the broader R&D community via the AI-on-demand platform and partnership activities.
Concise Answers to Specific Category Questions
- 1Eligible Applicant Types: university, research institute, higher education establishment, SME, large enterprise, nonprofit, public body, High-Performance Computing Centre, Advanced AI lab, consortiums combining those. Affiliated entities and associated partners as per call conditions.
- 2Funding Type: grant (Horizon Europe Research and Innovation Action, HORIZON-RIA).
- 3Consortium Requirement: consortium expected; single applicant technically possible under general rules but topic strongly anticipates multi-partner consortia including HPC centres and AI labs.
- 4Beneficiary Scope (Geographic Eligibility): legal entities established in EU Member States, Iceland and Norway, and associated countries specifically named: Canada, Israel, Republic of Korea, New Zealand, United Kingdom, Switzerland and Japan; additional third countries may be eligible if associated and listed at submission time. Entities controlled by China cannot submit guarantees for participation in this topic.
- 5Target Sector: Artificial Intelligence, High-Performance Computing, Machine Learning, Algorithms and computational infrastructure; cross-sector applications in Public Administration, Manufacturing, Energy, Healthcare, Cultural and Creative Industries, Physical AI.
- 6Mentioned Countries: EU Member States (unspecified individually), Iceland, Norway, Canada, Israel, Republic of Korea, New Zealand, United Kingdom, Switzerland, Japan. Reference made to List of Participating Countries in Horizon Europe for other eligible third countries 1.
- 7Project Stage: research and innovation targeting TRL 4-5 at project completion (development and validation of frameworks and prototype toolsets).
- 8Funding Amount: indicative EU contribution per project around €15,000,000; one grant expected for this topic.
- 9Application Type: open call single-stage via Funding & Tenders Portal (planned opening 20 Oct 2026; deadline 16 Feb 2027).
- 10Nature of Support: financial (monetary grants) to beneficiaries; results are expected to be shared publicly where appropriate and via AI-on-demand platform (non-monetary dissemination and community engagement are also required).
- 11Application Stages: 1 submission stage (single-stage call) followed by evaluation, then grant preparation which may include ownership control assessment and guarantee processing, then grant agreement signature — overall multi-step process but single-stage proposal submission.
- 12Success Rates: not specified; indicative low probability for single high-value grant calls due to strong competition — applicants should expect competitive selection.
- 13Co-funding Requirement: no specific additional co-funding percentage stated beyond standard Horizon eligible cost rules; applicants must demonstrate financial and operational capacity. Private partners may bring in additional in-kind or cash contributions for exploitation and scale-up.
- 14Templates: Standard HE RIA application form available in the Submission System (Part A and Part B). Ownership Control Declaration template and Annexes provided. Applicants should follow Part B structure: excellence, impact, implementation, ethics, security and budget. Additional templates: HE Programme Guide, HE MGA, Guidance on participation in restricted calls, Ownership Control Declaration annex (to be assembled by coordinator).
Key documents and guidance to consult:Horizon Europe Work Programme 2026-2027 Part 14 Horizontal Activities (Bridging actions 2027); General Annexes A to G; HE Programme Guide; Funding & Tenders Portal Online Manual; Ownership Control Declaration template; List of Participating Countries in Horizon Europe; Guidance on participation in EU calls with ownership and control restrictions.
Footnotes
- 1Topic details and official call documentation are available on the EU Funding & Tenders Portal: ec.europa.eu and the List of Participating Countries in Horizon Europe: ec.europa.eu
Short Summary
Impact Develop robust, practical toolsets and predictive methodologies that enable predictable, efficient scaling of high-performance AI models, reducing wasted compute, improving training and inference efficiency, and strengthening European technological sovereignty. | Impact | Develop robust, practical toolsets and predictive methodologies that enable predictable, efficient scaling of high-performance AI models, reducing wasted compute, improving training and inference efficiency, and strengthening European technological sovereignty. |
Applicant Teams with expertise in frontier AI research, scaling theory, large-scale model training and validation, dataset curation/mixture optimisation, architecture search, hyperparameter tuning, and access to high-performance computing for cross-hardware validation. | Applicant | Teams with expertise in frontier AI research, scaling theory, large-scale model training and validation, dataset curation/mixture optimisation, architecture search, hyperparameter tuning, and access to high-performance computing for cross-hardware validation. |
Developments Research and development of automated scaling-law derivation, exploratory architecture search, dataset mixture optimisation, advanced hyperparameter tuning and stability methods, and cross-hardware validation frameworks for AI model scaling. | Developments | Research and development of automated scaling-law derivation, exploratory architecture search, dataset mixture optimisation, advanced hyperparameter tuning and stability methods, and cross-hardware validation frameworks for AI model scaling. |
Applicant Type Researchers and research organisations, large corporations and SMEs active in AI and high-performance computing, and public bodies or non-profits involved in AI infrastructure and applied AI research. | Applicant Type | Researchers and research organisations, large corporations and SMEs active in AI and high-performance computing, and public bodies or non-profits involved in AI infrastructure and applied AI research. |
Consortium The topic expects multi-partner consortia (not a single individual) that combine advanced AI labs and HPC centres, though a single legal entity may apply under general rules; consortium proposals are strongly anticipated. | Consortium | The topic expects multi-partner consortia (not a single individual) that combine advanced AI labs and HPC centres, though a single legal entity may apply under general rules; consortium proposals are strongly anticipated. |
Funding Amount Indicative EU contribution of around €15,000,000 for one grant (one project expected). | Funding Amount | Indicative EU contribution of around €15,000,000 for one grant (one project expected). |
Countries Eligible participants must be legal entities established in EU Member States, Iceland and Norway, and the associated countries Canada, Israel, Republic of Korea, New Zealand, United Kingdom, Switzerland and Japan, with restrictions on entities controlled by certain non-eligible countries. | Countries | Eligible participants must be legal entities established in EU Member States, Iceland and Norway, and the associated countries Canada, Israel, Republic of Korea, New Zealand, United Kingdom, Switzerland and Japan, with restrictions on entities controlled by certain non-eligible countries. |
Industry Horizon Europe Frontier AI Initiative (Apply AI) — targeted at the artificial intelligence and high-performance computing sector. | Industry | Horizon Europe Frontier AI Initiative (Apply AI) — targeted at the artificial intelligence and high-performance computing sector. |
Additional Web Data
This is a forthcoming Horizon Europe Research and Innovation Action under Bridging actions 2027, with a single-stage deadline on 16 February 2027 at 17:00 Brussels time and an indicative budget of €15 millioneuro for around one project15. The submission system is planned to open on 20 October 2026.
| Field | Detail |
|---|---|
| Topic code | HORIZON-BRIDGING |
| Call title | Bridging actions 2027 |
| Programme | Horizon Europe |
| Type of action | HORIZON-RIA Research and Innovation Action |
| Submission model | Single-stage |
| Planned opening date | 20 October 2026 |
| Deadline | 16 February 2027 17:00 Brussels time1 |
| Indicative budget | €15 millioneuro15 |
| Indicative number of grants | 115 |
| Expected TRL at end | TRL 4 to 5 |
| Type of MGA | HORIZON Action Grant Budget-Based |
What the funding is about
The topic supports research that strengthens the scientific and technical foundations for scaling AI models, improving training and inference efficiency and giving European AI labs better predictive methods for frontier AI development.
Projects are expected to produce a practical toolset for scaling AI models, with emphasis on automated scaling laws, architecture search, dataset mixture optimisation, advanced hyperparameter tuning, stability, and compatibility across different hardware architectures.
- Automated scaling law derivation to predict performance before large compute investment
- Exploratory architecture search to improve sparsity, memory efficiency, and long context handling
- Dataset mixture optimisation including selection, weighting, sequencing, and data quality assessment
- Advanced hyperparameter tuning and stable large scale training
- Cross hardware validation to avoid optimisation for a single proprietary system
Who can apply
Eligibility is limited to legal entities established in EU Member States, Iceland, Norway, and the additional associated countries listed in the call conditions, namely Canada, Israel, Republic of Korea, New Zealand, United Kingdom, Switzerland, and Japan, with possible inclusion of third countries that become associated in 2026 or 2027 if they are eligible at submission and associated by grant signature.
Entities established in an eligible country but directly or indirectly controlled by a non eligible country or entity may only participate if they can provide positively assessed guarantees, and entities controlled by China or a Chinese legal entity cannot submit such guarantees.
- Advanced AI labs
- High performance computing centres
- Universities and research organisations
- Industrial and applied AI consortia with frontier model expertise
- Consortia able to work with strategic sectors such as public administration, manufacturing, energy, healthcare, cultural and creative industries, or physical AI
Funding and project expectations
The call expects a project that reaches TRL 4 to 5 by the end of implementation and that demonstrates progress through qualitative and quantitative KPIs, benchmarking, and monitoring.
The public excerpt does not state a special project duration or a non standard reimbursement rate; as a Horizon Europe RIA, applicants should plan under the standard RIA funding regime and build their budget around a single stage competitive proposal.
| Applicant issue | What to plan for |
|---|---|
| Budget scale | A project around €15 millioneuro is the benchmark15 |
| Delivery level | TRL 4 to 5 by project end |
| Validation environment | Advanced AI labs and high performance computing centres |
| Dissemination | Share communicable results through the AI on demand platform |
| Reuse | Build on public results from prior funded actions when possible |
Conditions applicants should plan for
- Provide clear qualitative and quantitative KPIs, benchmarking, and progress monitoring
- Describe the link to HORIZON-CL4-2027-04-DIGITAL-EMERGING-11 and the wider Frontier AI Initiative
- Show how the work will support the AI Factories and Data Labs ecosystem
- Allocate cohesion activities with ADRA and the GenAI4EU central Hub
- Ensure the methods are compatible with diverse hardware architectures
- Explain how results will help strategic sectors and pre competitive reuse by the wider developer community
Applicants should also be ready to use the standard Horizon Europe RIA application form, the call specific ownership control declaration where relevant, and the general evaluation and grant conditions from the Work Programme annexes.
Practical assessment
This opportunity is best suited to consortia that can combine frontier AI theory, model scaling experimentation, access to compute, and strong dissemination capability, because the call is not just about research outputs but about usable scaling methodologies for Europe’s advanced AI ecosystem.
Update Log
No updates recorded yet.
Documents
Discover with AI
Let our intelligent agent help you find the perfect funding opportunities tailored to your needs.
EU Grant Database
Explore European funding opportunities in our comprehensive, up-to-date collection.
Stay Informed
Get notified when grants change, deadlines approach, or new opportunities match your interests.
Track Your Favorites
Follow grants you're interested in and keep them organized in one place. Get updates on changes and deadlines.
EU Frontier AI Initiative: Developing frontier AI solutions that are safe and computationally efficient within Apply AI (RIA)
The EU Frontier AI Initiative focuses on developing safe and computationally efficient frontier AI solutions under the Horizon Europe program, specifically in the DIGITAL sector. The primary funding mechanism is a grant, categorized as a...
Large scale operational validation and upscaling of state-of-the-art (Generative) AI tools and models powering a next generation digital energy system
The Horizon Europe funding opportunity identified as HORIZON-CL5-2027-02-D3-24 focuses on substantial investments aimed at the large-scale operational validation and upscaling of state-of-the-art generative AI tools and models within the...
Researching and applying the potential of Artificial Intelligence to foster climate resilience at the regional and local levels
Horizon Europe Research and Innovation Action HORIZON-MISS-2027-01-CLIMA-02 funds AI solutions to strengthen climate resilience at regional and local levels under the Adaptation to Climate Change Mission. The topic has an indicative budg...
AI4Creatives Support Platform: embracing a fair AI revolution
The grant opportunity titled "AI4Creatives Support Platform embracing a fair AI revolution" falls under the Horizon Europe framework, specifically aimed at establishing a European virtual AI support platform for the cultural and creative...
Autonomous Drive Ambition Cities (IA) - Apply AI
Horizon Europe Bridging actions 2027 Innovation Action 'Autonomous Drive Ambition Cities (IA) - Apply AI' will fund one large-scale demonstrator with an indicative EU contribution of around EUR 15,000,000 (single-stage call; deadline 16...
International cooperation in AI (IA)
This EU funding opportunity is a part of the Horizon Europe program, specifically under the Digital call (HORIZON-CL4-2027-04). It is classified as an Innovation Action (HORIZON-IA) aimed at fostering international cooperation in Artific...
EU AI Media Platform – modular multilingual public-interest infrastructure for European news and audiovisual media (IA) - Apply AI
Horizon Europe Bridging Actions 2027 funds an Innovation Action (topic HORIZON-BRIDGING-2027-01-09) to build and validate a modular, trustworthy, AI-enabled multilingual public-interest media infrastructure. The single-stage call opens 2...
Energy efficiency and sustainability of AI data processing in Data Centres (IA)
The grant opportunity is titled HORIZON-CL4-2027-04-DATA-09, focusing on "Energy efficiency and sustainability of AI data processing in Data Centres" within the Horizon Europe Innovation Actions framework. The status of this call is fort...
Testing and optimising models of co-creation of advanced research infrastructure technologies
The Horizon Europe initiative, specifically the HORIZON-INFRA-2027-TECH-01-01, focuses on the advancement of research infrastructure technologies through a collaborative grant mechanism. Eligible applicants mainly include research infras...
New approaches for decentralized, federated and sustainable AI data processing (RIA)
This EU funding opportunity, under the Horizon Europe program, specifically the call DIGITAL (HORIZON-CL4-2027-04), seeks to advance decentralized, federated, and sustainable AI data processing. The project type is classified as a HORIZO...
Online harms detection and investigation tools using a short development cycle model
The Horizon Europe call, specifically under the Civil Security for Society 2027 program, focuses on developing online harms detection and investigation tools. This initiative is structured as a Horizon Innovation Action, classified as a...
Community of practice - Data-Driven Decision-Making in Energy
The funding opportunity titled "Community of Practice - Data-Driven Decision-Making in Energy" is part of Horizon Europe, specifically under the Coordination and Support Actions category. The call identifier is HORIZON-CL5-2027-07-D3-28....