This Challenge will support the ambitions of the AI Act, Apply AI Strategy. The AI powered robotics models developed under this Challenge are expected to comply with the EU concept for Trustworthy AI and the relevant ethical principles with due attention paid to data quality, transparency and accountability, privacy, security and safety. In the medium to longer term, it is expected to reduce European dependencies and support end users in leveraging advances in AI to enhance their products and develop new capabilities that will contribute to:

• Strengthened European leadership in physical AI, reducing reliance on non-European technology providers, and
• Contribution to EU priorities in sustainability, resilience, digital transformation, and industrial competitiveness.

Objective:

This is a two-stage Challenge competition of which the ultimate ambition is to deliver scalable prototypes, validated in a relevant real-world environment. Solutions supported under the Challenge are expected to demonstrate both the novelty and applicability of the proposed Physical AI technology, while mitigating associated risks.

Applicant to Stage 1 must also have a clear outlook on the potential longer-term impact(s).

Stage1- Solution validation and benchmarking

Applicants with disruptive physical AI systems that have achieved a prototype (i.e. at least TRL 4) will be invited to:

• Refine and validate the prototype in a relevant environment to demonstrate the core innovation (e.g., new perception (multimodal), actuation, continual learning, or autonomy capability, etc.). in the end use case
• Develop a methodology for performance assessment
• Assess ethical, data governance and scalability considerations
• Identify challenges to integration such as safety, certification or scalability with view to identifying risks and exploring options for overcoming barriers for adoption including in cases where products’ risks are not covered by the AI Act
• Conduct initial performance benchmarking, and
• Develop an initial market analysis, business plan and model.

The outputs of Stage 1 will be assessed for progress against the following milestones:

• Evidence of technical viability with initial performance data that benchmarks performance and disruptive potential against the state-of-the-art in at least one pilot case within the specified application area, and
• Clear commitment from end users and stakeholders to further develop the physical AI solution in Stage 2.
• To maximise impact, foster collaboration, and accelerate learning across funded teams, Progress to Stage 2 will be contingent on delivering a full proposal which will be assessed alongside the milestones achieved under Stage 1.

Stage 2 - Development and user testing

Applicants for Stage 2 will be expected to further develop and validate a Physical AI system through the following activities:

• Build and test (ensure at least two deployments loops) a robust, integrated solution capable of autonomous operation in complex or semi-structured environments (e.g., pilot deployment in real world environment, etc.) to reach TRL 6-7
• Address scalability and manufacturability: show that the solution can be scaled or adapted for broader integration and deployment
• Conduct necessary testing, documentation, and compliance activities towards certification or regulatory approval of the adapted solution
• Organisation of consultation meetings/workshops with industrial end users, regulators, standard setting bodies or other relevant stakeholders to gather further needs and feedback, and
• Develop a roadmap to commercialisation: refine market analysis, and validate business plan including stakeholder engagement, and identification of scaling strategies

At the end of Stage 2, projects will be expected to have:

• Delivered a robust, integrated solution capable of autonomous operation in complex or semi-structured environments i.e. capable of pilot deployment in real world environment
• Demonstrated reliability, adaptability, energy/resource efficiency, and safe human-AI interaction, and
• Provided evidence of necessary testing, documentation, and compliance activities towards certification or regulatory approval for the adapted solution

Scope:

Background and Scope

The integration of AI into physical systems is accelerating apace. Such integration will enable machines and systems to perceive, act, and interact autonomously in complex, real-world settings.

This Challenge aims to accelerate the development towards integration, deployment and commercialisation of breakthrough Physical AI solutions that will enhance Europe’s technological sovereignty, sustainability, and global competitiveness.

This Challenge seeks proposals for developing disruptive Physical AI innovations that demonstrate at least two of the following characteristics:

• Intelligent Perception and Cognition - Encompasses robust sensing, understanding, and real-time reasoning capabilities.
• Adaptive Learning and Optimisation - Focuses on continuous improvement, domain-specific adaptation, and intelligent resource management.
• Autonomous Decision-Making and Collective Intelligence - Combines real-time problem-solving with autonomous action and swarm coordination in dynamic environments.
• Human-AI Collaboration and Interaction - Encompasses safe, intuitive interfaces and collaborative robotics for human-AI teamwork.
• Physical Integration and Innovation - Focuses on developing novel sensors, actuators, and materials that advance embodied AI capabilities.

The solutions must also focus on addressing pressing needs in at least one of the following application areas:

• Disaster response and civil security physical agents: AI-powered robotic systems designed to act before, during, or immediately after disasters (natural or man-made), with the aim of saving lives, ensuring safety, and restoring normalcy in hazardous, unpredictable environments.
• Autonomous labs for Scientific Discovery: self-directed scientific environments where AI and robotics design, execute, and analyse experiments with minimal human intervention.
• Personal or professional robot assistants: AI-powered robots designed to support individuals or professionals by automating repetitive, menial, or complex tasks in homes, workplaces, or industrial settings.

The Physical AI solutions that will be designed and developed under this Challenge should be in accordance with principles of Responsible AI, Trustworthiness, Security and Human Centrism, ensuring transparency, accountability, privacy, safety, and ethical alignment.

Entry requirements

Applicants must demonstrate the interest of potential end users and integrators with a view to supporting the application and testing of the proposed solutions in a real-world environment. To enable this, they should have:

• an initial physical AI system prototype tested in the lab
• access to an appropriate infrastructure for data collection and testing, and
• a challenging use case formulated by a potential end-user, supported by a letter of intent or a letter of intent from a robotics integrator, confirming their interest aligned to the needs of the application domain.