Modelling & simulation-supported AI framework for military decision-making and training

Fill out the form

Official call identifier: EDF-2026-RA-SIMTRAIN-MSAI

Objective of the call

This topic funds a feasibility assessment and initial development work towards a unified AI-system framework for defence modelling & simulation (M&S). The framework is intended to support tactical intelligence and decision-making, connect with command-and-control (C2) systems and synthetic training environments, and enable faster experimentation, validation, and learning support in mission planning and execution.

Scope of activities

Proposals must focus on simulation-area solutions and:

• Cover two or more use cases spanning forces readiness, operational support, and capability development.

• Identify challenges and preconditions for AI use in a multi-national defence context, including state of the art, research gaps, needs, and requirements.

• Research a subset of challenges and increase maturity of applying AI in a defence context for at least two selected use cases.

• Ensure compliance with NATO standards and accommodate coalition situations for interoperability across domains and systems.

• Demonstrate capabilities to integrate AI for simulation purposes.

Mandatory work (within the eligible activity types) includes:

• Studies on research gaps, needs and requirements, including gaps in training technologies and standardisation, state-of-the-art architectures across domains and countries, impact of disruptive technologies, AI brittleness and bias identification/mitigation (in AI outputs and AI-influenced human decision-making), interoperability challenges for live-system integration, and early concepts for AI use.

• Coverage of AI core components such as data ingestion/processing and common defence ontologies; ML/analytics for predictive modelling and decision support; human-AI collaboration interfaces; cybersecurity and resilience including adversarial robustness; simulation and training modules; operational needs/scenarios/use cases; and evaluation methods for model training and performance.

• Identification and handling of data sources (simulation, real-world, synthetic streams, historical data, benchmarks, doctrinal models, human performance data) and support for (re)playing data streams to test AI solutions.

• Research on military data models/ontologies/standards for interoperability between AI services, and standards for exporting AI outputs consumable by end-user C2 and simulation systems (including human-machine interface concepts).

• Design and demonstration of technological novelties improving realism and efficiency, example customised solutions using the framework, and demonstration/evaluation of a representative multi-national simulation combat scenario.

• Address at least two of these four use cases:

  • Case 1: End-to-end MDO system-of-systems performance assessment integrating real C2, M&S-based AI synthetic force-on-force engagement, and human-in-the-loop simulation enhanced with AI (preparation, execution, analysis stages).

  • Case 2: Human-machine teaming with adaptive levels of automation, role/authority transitions, workload management, training for taskwork and teamwork, exposure to AI failure modes/adversarial actions, and practical metrics for team performance and usability.

  • Case 3: AI-supported COA development/comparison using reinforcement learning agents to generate COAs from the common operational picture, assess COAs via simulation, and generate corresponding orders.

  • Case 4: Simulation-generated data to support AI training where real-world capture is not feasible; includes a sample use case on AI recognition of asymmetric threats at sea under varied conditions.

Optional elements (where applicable):

• Use of financial support to third parties (FSTP) (must be described if used).

• Synergies/complementarity with NATO and EDF simulation activities (planned, ongoing, or completed).

Eligible applicants 

• Innovative SMEs, startups and enterprises aiming to scale and strengthen competitiveness 
• Public-sector organisations driving transformation and societal impact 
• Research and academic institutions commercialising knowledge 
• Non-profit organisations delivering purpose-driven innovation 
• Organisations established in EU Member States (including OCTs) or listed EEA / countries associated to the EDF Programme, with executive management in eligible countries and not controlled by a non-associated third country/entity
• Minimum 3 independent partners from 3 eligible countries 

Specific consortium we are looking for:

• Defence modelling & simulation and wargaming/battlespace simulation expertise (including high-fidelity, fast-executing simulation environments).

• AI/ML engineering for defence use (including reinforcement learning baselines/agents, multi-agent systems, explainable AI, data fusion, analytics).

• C2 systems integration and interoperability engineering (including APIs/interfaces, modular service architectures, export/consumption of AI outputs by end-user systems).

• Military data expertise: data models, ontologies, doctrinal models, benchmark datasets, (re)playable data streams, and data governance in multinational settings.

• Cybersecurity and resilience expertise for adversarial robustness, secure deployment and update pipelines (MLOps for defence), and protection against dataset poisoning.

• Human factors / human-AI teaming expertise: workload, trust, usability for non-technical users, HITL/HOTL concepts, and metrics for team performance.

• Standards and compliance capability, including NATO standards and coalition interoperability considerations, and security handling where classified information is involved.

Eligible costs 

Eligible cost categories and conditions include:

• Personnel costs (employees, natural persons under direct contract, seconded persons; unit costs for SME owners/natural person beneficiaries; average personnel costs unit cost allowed)

• Subcontracting costs (country restrictions apply: subcontracted work must be performed in eligible countries)

• Purchase costs

  • Travel and subsistence (actual costs only)

  • Equipment (depreciation only)

  • Other goods, works and services

• Other cost categories

  • Financial support to third parties (FSTP) (allowed; max EUR 60,000 per third party)

  • Internally invoiced goods and services (unit cost allowed)

• Indirect costs

  • Flat-rate 25% of eligible direct costs (with stated exclusions) or actual indirect costs (method fixed for the project)
    Additional eligibility notes in the call text include:

• Only costs for activities carried out in eligible countries are eligible.

• Use of assets/infrastructure/facilities/resources located or held outside eligible countries is ineligible (even if exceptional use is authorised).

• Non-deductible/non-refundable VAT is eligible (with stated limitation for public bodies acting as public authority).

• Costs for separate project websites are not eligible; communication on participants’ websites/social media is eligible.

Funding conditions 

Type of action: 
Research Action

Funding rate: 
Up to 100% of total eligible costs 

EU contribution per project: 
Around EUR 8,000,000 

Number of funded projects: 
Up to 2 projects 

Project duration: 
Typically 36 months

Deadline for submission

29 September 2026