EU Frontier AI Initiative: Developing frontier AI solutions that are safe and computationally efficient within Apply AI (RIA)
Horizon Europe · Grant topic
Summary
A Polish company can use this call for R&D.
Funds research and training of European frontier AI models combining advanced capabilities, computational efficiency and safety by design.
- Who can apply
- Companies can lead: SME, large company, startup, research org, university.
- Amount
- €44M per projectTotal budget €44,000,000
- Deadline
- 18 Mar 2027 01:00 CETin 145 days
- Status
- Forthcomingopens 17 Nov 2026
Machine-classified from the call text; check eligibility at the source.
Description
Expected Outcome:
The Apply AI Strategy[1] also seeks to bolster EU capabilities and achieve excellence in AI to support the development of European frontier models. As part of the Frontier AI Initiative, which brings together Europe’s leading actors in the field, this topic will support the development of sovereign frontier AI ensuring safety by design. This topic directly contributes to the Apply AI Strategy. Project results are expected to contribute to all of the following expected outcomes:
• Strengthened European capabilities in the development of frontier AI models.
• Improved computational efficiency of frontier AI models, resulting in reduced computational costs.
• Enhanced safety of advanced AI systems based on frontier AI models through the development and implementation of safe-by-design principles and/or AI agents acting as safety evaluators.
Scope:
To advance developments of frontier AI models towards highest-level performance, while ensuring energy efficiency, addressing computational constraints, and strengthening safety. The approach of this topic is twofold. First, it aims to advance the AI field through the development and training of a frontier AI model. The AI model should demonstrate state-of-the-art performance, have multimodal capabilities, and be optimized for agentic AI capabilities such as tool use, reasoning, and autonomous problem-solving. Second, this topic supports research on comprehensive methods to reduce the computational demands of frontier AI models and to ensure their safety, including technical methodologies such as automated testing and interpretability.
The primary drivers behind computational efficient AI systems are the urgent challenges posed by the growing energy footprint of AI and current computational limitations. Modern AI models, especially frontier AI models, require substantial computational resources, with a significant impact in the environment. Additionally, they create barriers to entry to those interested in advancing the AI field. Key research areas include compression and distillation techniques aimed at reducing the complexity of large AI models. Innovations in AI architectures are also relevant, with a focus on innovative models that significantly lower computational demands for training and inference. Further, algorithmic approaches aimed at minimizing computational load during pre-training, post-training, and inference can also be considered.
Ensuring the safety of AI systems is essential, especially as AI models become increasingly sophisticated and pervasive. Potential research areas to be considered include addressing misalignment, particularly the unintentional misalignment of large AI models. Work in this area could explore methods to detect and mitigate sophisticated misbehaviour, such as alignment faking, reward hacking of human oversight, and encoded reasoning in chain-of-thought (CoT). Additionally, research could focus on enhancing robustness against adversarial attacks, jailbreaks, and backdoors. Further potential areas for innovation include advancing AI models transparency and interpretability. Safety research could also consider risks that may arise when embedding frontier models within agentic AI frameworks, significantly contributing to the trust and safe adoption of powerful AI solutions.
This topic contributes to the EU Frontier AI initiative. The project should establish strong links with the Resource for AI Science in Europe (RAISE), ensuring that its priorities inform the research topics addressed. Activities are expected to involve the European AI research community and attract and retain top AI talent working on frontier models and related areas.
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, it could also establish links with European companies developing frontier AI models.
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 start at TRL 2 and achieve TRL 4 by the end of the project – see General Annex B.
[1] COM(2025)723 Apply AI Strategy
Who can apply, in detail
- Applicant types
- SME, Large company, Startup, Research org, University
- Purpose
- Research and innovation
- R&D relevance
- 3 of 3
- Poland eligible
- Yes
- Country limit
- —
- Region
- Any
- Consortium
- Required
- Readiness level (TRL)
- —
- Topics
- Frontier AI models
- Multimodal AI
- Agentic AI
- Computational efficiency
- AI safety by design
Machine-classified from the call text; check eligibility at the source. 23 Oct 2026
Conditions
Typed facts extracted from the source documents by Scindu's grant-data review. Each value keeps its own conditions. Unknown means unstated, not zero or unlimited.
- Eligible project scope
- The topic supports frontier AI model development and training with multimodal and agentic capabilities, computational efficiency and safety-by-design.
- Other condition
- Horizon Europe General Annexes describe personnel as actual costs unless declared as a unit cost; they separately list SME-owner/natural-person personnel unit costs, subcontracting as actual costs, purchases as actual costs unless unit cost, and FSTP only if provided by the specific call. Therefore these personnel categories do not turn an external service purchase or subcontract into a hire, and no specific FSTP entitlement is asserted for this topic.
- Other condition
- For A.1 employees/equivalent, the HE MGA states annual personnel costs divided by 215 as the daily-rate basis; declared day-equivalents must be identifiable and verifiable. The total declared day-equivalents per person across EU grants in a year cannot exceed 215, minus parental leave time where applicable.
- Other condition
- Activities are expected to involve the European AI research community and attract and retain top AI talent working on frontier models and related areas; the call does not name a job title, qualification threshold, headcount, FTE or salary scale.
- Other condition
- Activities are expected to start at TRL 2 and achieve TRL 4 by the end of the project.
- Other condition
- Exact source identifies budget-based HORIZON-AG; General Annexes describe actual-cost grants. Eligibility/reimbursement terms depend on applicable MGA and grant agreement.
Amounts and costs
- Indicative budget
- €44,000,000
Topic budget as stated in the portal's budget table. Lump-sum and per-project limits are in the call conditions.
- Eligible cost
- Directly contracted natural persons (under a non-employment contract) and paid seconded persons may be personnel costs when assigned to the action and meeting general eligibility conditions; they must work under conditions similar to employees, results belong to the beneficiary unless agreed otherwise, and actual costs must not be significantly different from employees performing similar tasks.
- Eligible cost
- Employees/equivalent may be charged as personnel costs only if general eligibility conditions are met, the person works for the beneficiary under an employment contract or equivalent appointing act, is assigned to the action, and remuneration-linked amounts arise under national law or the employment terms; this category is based on costs actually incurred.
- Eligible cost
- The HE MGA allows nonsalaried SME owners and natural-person beneficiaries to claim personnel costs as unit costs under Annex 2a, subject to general eligibility conditions; the AGA gives a daily-rate formula of EUR 8,745.4 divided by 18 days, multiplied by the applicable country-specific correction coefficient for calls opened on or after 30 July 2024. This is a unit-cost calculation, not an employee salary cap or rate guaranteed for every country.
- Funding intensity, maximum
- 100
General maximum for exact HORIZON-RIA code; topic and GA control; not a staff pay rate.
- Expected project funding, up to
- €44,000,000
Approximate expected contribution per project, not an award floor/cap; the topic expressly permits different requested amounts.
- Expected project funding, from
- €44,000,000
Approximate expected contribution per project, not an award floor/cap; the topic expressly permits different requested amounts.
- Topic allocation
- €44,000,000
Indicative topic allocation, distinct from expected EU contribution per project and staff costs.
Personnel and services
Other role
Staff cost the grant can fund
Potential fundable personnel category only when the grant conditions and Article 6.2.A.2 apply; assigned to the action, work under conditions similar to an employee, result belongs to beneficiary unless agreed otherwise, and actual direct-contract costs are not significantly different from comparable employees. It is a contractor engagement category, not employee payroll; no qualification, headcount, FTE, duration or rate is set by this topic.
Other role
Staff cost the grant can fund
Potential fundable personnel category under the applicable Horizon Europe General MGA, conditional on actual action assignment and all eligibility rules. This is not a call-specific required team, named job, or promise of award. Qualification, headcount, FTE, duration and salary are not stated by the topic. Employee remuneration must be actual and linked to national law/employment terms; daily-rate and reporting rules are described in the cited MGA.
Other role
Staff cost the grant can fund
Potential personnel unit-cost category where the applicable grant agreement includes Article 6.2.A.4. Applies only to an SME owner not receiving a salary or a natural-person beneficiary not receiving a salary, with work assigned to the action and general unit-cost conditions met. The daily unit rate is country-adjusted; it is not a salary and this packet does not assert a per-person EUR rate without the recipient country coefficient. No required headcount, FTE, duration or qualification is named in this topic.
Other role
Staff cost the grant can fund
Potential fundable personnel category only when the grant conditions and Article 6.2.A.3 apply; assigned to the action, works under conditions similar to an employee and actual costs are not significantly different from comparable beneficiary employees. This is a paid secondment, not a new employment contract with the beneficiary. No call-specific qualification, headcount, FTE, duration or pay rate is stated.
Source documents
Call facts
- Opens
- 17 Nov 2026
- Deadline model
- Single-stage
- Programme
- Horizon Europe · 2021 - 2027
- Call
- HORIZON-CL4-2027-04 · DIGITAL
- Types of action
- HORIZON Research and Innovation Actions
- Keywords
- Frugal AI
- Generative AI
- Safe AI
- Source
- ec.europa.eu
- Checked
- 24 Oct 2026
- Portal updated
- 19 Oct 2026
- Review status
- Partial review
- Reviewed
- 23 Oct 2026
This is a summary of the source page, not the official text. Eligibility, conditions and the submission system are at the source.