EU Frontier AI Initiative: Advanced Scaling Frameworks for High-Performance AI Models (RIA) - Apply AI

**A pályázat célja és várt eredménye** Expected Outcome: This topic aims to advance the scientific and technical foundations for the scaling of AI models, enhancing the performance of training and inference processes and providing European AI labs with the methodologies required to compete at the global frontier. Project results are expected to contribute to some of the following outcomes: Establishment of robust scaling laws and predictive frameworks that allow for the reliable estimation of model performance before large-scale compute investment. Enhanced competitiveness of European advanced AI models through optimised dataset mixtures, refined architectures, and superior hyperparameter configurations. Optimised development for frontier AI models, achieved via automated and scientifically grounded scaling methodologies. Scope: To maintain a leading position in AI, Europe must master the methodologies required to scale AI models’ performance, training and inference processes. Establishing a predictive methodology for AI model development ensures that limited computational resources are deployed with maximum efficiency, thereby bolstering the EU and AC’s technological sovereignty. While previous actions have focused on training High-Performance AI models, this topic addresses the research gap in the underlying theoretical basis, methodologies and tools for predictable scaling of AI models and inference. By providing the technical foundations for high-performance systems without the costs of unguided experimentation, this action complements training-centric efforts. In this light, proposals should deliver, as an outcome of the research, a toolset designed to support the scaling of AI models. Potential research areas include, but are not limited to: Automated Scaling Law Derivation: Research into scaling laws to predict performance across diverse modalities and tasks, moving beyond simple power laws. Exploratory Architecture Search: Developing automated methods to discover and refine model architectures that maintain performance while scaling both in training and inference, including innovations in sparsity, memory efficiency, and long-context handling. Dataset Mixture Optimisation: Research into the systematic selection, weighting, and sequencing of datasets. This includes developing tools for automated data quality assessment. Advanced Hyperparameter Tuning and Stability: Developing methodologies for stable training at large scales, including automated tuning of optimisers and learning rate schedules that remain effective as models grow. Proposals should ensure that the developed frameworks and scaling methodologies are compatible with diverse hardware architectures, demonstrating that scaling is studied across different systems to avoid optimisation for specific proprietary hardware. The selected project should actively support Advanced AI labs and High-performance Computing Centres already engaged in the development of frontier AI, ensuring that methodological advances are immediately validated in high-stakes environments. Proposals should describe how these scaling methodologies would benefit the development of models adapted to the specific requirements of particular strategic sectors, such as Public Administration, Manufacturing, Energy, Healthcare, Cultural and Creative Industries, or Physical AI. Part of the Frontier AI Initiative, this topic complements HORIZON-CL4-2027-04-DIGITAL-EMERGING-11 “EU Frontier AI Initiative: Developing frontier AI solutions that are safe and computationally efficient within Apply AI” by establishing the underlying methodologies required for the large-scale training and inference of High-Performance AI models. Proposals should, therefore, describe the link between their planned activities and the objectives of topic HORIZON-CL4-2027-04-DIGITAL-EMERGING-11. All proposals are expected to incorporate mechanisms for assessing and demonstrating progress, including qualitative and quantitative KPIs, benchmarking, and progress monitoring. When possible, proposals should build on and reuse public results from relevant previous funded actions. Communicable results should be shared with the European R&D community through the AI-on-demand platform. The project selected in this topic should link to the resources offered by the AI Factories and the Data Labs. Where relevant, links could be established to European companies developing frontier AI models to foster a pre-competitive exchange of expertise and explore ways to mobilise the wider developer community for the efficient reuse of results. All proposals are expected to allocate tasks for cohesion activities with the European Partnership on AI, data, and robotics (ADRA) and the CSA HORIZON-CL4-2025-03-HUMAN-18: GenAI4EU central Hub. Technology Readiness Level - Technology readiness level expected from completed projects Activities are expected to achieve TRL 4-5 by the end of the project – see General Annex B. **Jogosultsági és benyújtási feltételek** "> General conditions 1. Admissibility conditions: Proposal page limit and layout described in Annex A and Annex E of the Horizon Europe Work Programme General Annexes. Proposal page limits and layout: described in Part B of the Application Form available in the Submission System. 2. Eligible countries described in Annex B of the Work Programme General Annexes. A number of non-EU/non-Associated Countries that are not automatically eligible for funding have made specific provisions for making funding available for their participants in Horizon Europe projects. See the information in the Horizon Europe Programme Guide . 3. Other Eligibility Conditions In order to achieve the expected outcomes, and safeguard the Union’s strategic assets, interests, autonomy, and security, it is important to avoid a situation of technological dependency on a non-EU source, in a global context that requires the EU to take action to build on its strengths, and to carefully assess and address any strategic weaknesses, vulnerabilities and high-risk dependencies which put at risk the attainment of its ambitions. For this reason, participation is limited to legal entities established in Member States, Iceland and Norway and the following additional associated countries : Canada, Israel, Republic of Korea, New-Zealand, United Kingdom, Switzerland and Japan. In addition, entities established in third countries which may become associated to Horizon Europe during 2026 and 2027 may be eligible to participate in this topic if the third country is identified for this topic as an eligible country in the List of Participating Countries in Horizon Europe at the time of submission of the application[[See the List of Participating Countries in Horizon Europe available at https://ec.europa.eu/info/funding-tenders/opportunities/docs/2021-2027/common/guidance/list-3rd-country-participation_horizon-euratom_en.pdf .]]. In any case, the association agreement to the Programme must apply by the time of the signature of the grant agreement. For the duly justified and exceptional reasons listed in the paragraph above, in order to guarantee the protection of the strategic interests of the Union and its Member States, entities established in an eligible country listed above, but which are directly or indirectly controlled by a non-eligible country or by a non-eligible country entity, may not participate in the action unless it can be demonstrated, by means of guarantees positively assessed by their eligible country of establishment, that their participation to the action would not negatively impact the Union’s strategic assets, interests, autonomy, or security . Entities established in an eligible country, but which are directly or indirectly controlled by China or by a legal entity established in China cannot submit guarantees. [[ The guarantees shall in particular substantiate that, for the purpose of the action, measures are in place to ensure that: a) control over the applicant legal entity is not exercised in a manner that retrains or restricts its ability to carry out the action and to deliver results, that imposes restrictions concerning its infrastructure, facilities, assets, resources, intellectual property or know-how needed for the purpose of the action, or that undermines its capabilities and standards necessary to carry out the action; b) access by a non-eligible country or by a non-eligible country entity to sensitive information relating to the action is prevented; and the employees or other persons involved in the action have a national security clearance issued by an eligible country, where appropriate; c) ownership of the intellectual property arising from, and the results of, the action remain within the recipient during and after completion of the action, are not subject to control or restrictions by non-eligible countries or non-eligible country entity, and are not exported outside the eligible countries, nor is access to them from outside the eligible countries granted, without the approval of the eligible country in which the legal entity is established.]] described in Annex B of the Work Programme General Annexes. 4. Financial and operational capacity and exclusion described in Annex C of the Work Programme General Annexes. 5a. Evaluation and award: Award criteria, scoring and thresholds described in Annex D of the Work Programme General Annexes. 5b. Evaluation and award: Submission and evaluation processes described in Annex F of the Work Programme General Annexes and the Online Manual . 5c. Evaluation and award: Indicative timeline for evaluation and grant agreement described in Annex F of the Work Programme General Annexes. 6. Legal and financial set-up of the grants described in Annex G of the Work Programme General Annexes. Specific conditions described in the specific topic of the Work Programme. Application and evaluation forms and model grant agreement (MGA): Application form templates — the application form specific to this call is available in the Submission System Standard application form (HE RIA, IA) Evaluation form templates — will be used with the necessary adaptations Standard evaluation form (HE RIA, IA) Guidance HE Programme Guide Model Grant Agreements (MGA) HE MGA Call-specific instructions Ownership Control Declaration Annex Additional documents: HE Main Work Programme 2026-2027 – 1. General Introduction HE Main Work Programme 2026-2027 – 14. Horizontal Activities HE Main Work Programme 2026-2027 – 15. General Annexes HE Programme Guide HE Framework Programme and Rules for Participation Regulation 2021/695 HE Specific Programme Decision 2021/764 EU Financial Regulation 2024/2509 Rules for Legal Entity Validation, LEAR Appointment and Financial Capacity Assessment Guidance on participation in EU calls with ownership and control restrictions EU Grants AGA — Annotated Model Grant Agreement Funding & Tenders Portal Online Manual Funding & Tenders Portal Terms and Conditions Funding & Tenders Portal Privacy Statement