UK registered SME businesses are eligible to apply for a share of up to £1.66 million to deliver experimental validation of prototype AI systems for novel reusable components of efficient, structured and controllable learning systems. This listing outlines the key parameters, themes, eligibility conditions and support arrangements for the funding opportunity.
Competition Overview
• Funding body and investment: Innovate UK, part of UK Research and Innovation (UKRI), will invest up to £1.66 million on industrial research for novel reusable AI mechanisms, subject to a sufficient number of high quality applications being received.
• Aim: The competition aims to advance the development of novel reusable artificial intelligence (AI) mechanisms that will form the foundations for next-generation AI and machine learning (ML) systems.
• Strategic imperative: Future capabilities beyond the current state of the art require fundamentally different underlying architectures and learning systems. Resource adaptive learning mechanisms, intervention tested and counterfactual prediction, graph-native learning topologies and representation modelling for scalable primitive learning components are fundamental components for adaptive scalable learning systems.
• Funding purpose: Innovate UK will fund UK SMEs to experimentally validate novel AI mechanisms that could underpin future generations of efficient, structured, adaptive and controllable AI systems.
• Novel mechanistic AI definition: For this competition, novel mechanistic AI means an AI or ML system in which the principal technical advance is attributable to one or more of:
- a new learning or credit assignment mechanism
- an adaptive or dynamically reconfigurable architecture
- a learned causal or intervention grounded model
- a graph-native or relational computational architecture
- a compositional representation, skill or operator mechanism
- a mechanism, model or architecture exerting measurable control over an agent
• Frontier model condition: When an existing frontier model is used, applications must demonstrate that the proposed advance remains technically material above the base model.
• Competition themes: Proposals must fall within one or more of the following themes: resource and data adaptive learning; causal and intervention grounded models; learned modular computation and routing; geometric, equivariant and physics structured learning. See the Specific Themes section below for further details and priorities.
• Intended applicants: The competition is intended primarily for research intensive AI startups, deep tech SMEs, university spinouts, and pre-commercial companies developing proprietary AI technology.
• Technology Readiness Level: Projects must start at TRL 1, 2 or 3 and are expected to make an advance of 0.5 to 1 TRL during delivery.
• Required project deliverables: Projects must deliver a prototype algorithm architecture or subsystem; an experimental and validation environment; results against named baselines or a hypothesis; ablation and negative control evidence; and a plan for further development.
• End of project technical whitepaper: At the end of the project, applicants will be required to submit a technical whitepaper summarising technical progress and outcomes — a concise description of what was built and learned; evidence and validation — quantified results against pre-defined success criteria, test conditions and a clear statement of remaining technical risks and limitations; and an innovation defensibility plan.
• Potential Phase 2: Subject to Business Case approvals, there might be a second phase of this opportunity. If successful, applicants would be notified of the conditions and availability of Phase 2. There is no funding currently allocated for Phase 2.
• Estimated success rate: Experience from similar competitions suggests that applicants could have around a 4% chance of success.
Registration and Award Details
• Opens: 12/10/2026
• Closes: 18/11/2026 at 11:00am
• Opportunity type: Funding
• Location: East Midlands, East of England, London and the South East, North East (England), North West (England), Northern Ireland, Scotland, South West (England), Wales, West Midlands, Yorkshire and The Humber
• Award: Project total costs must be between £50,000 and £150,000; up to 70% of costs can be covered, depending on business size.
• Organisation: Innovate UK
• Sector: Robotics & AI
• Project duration: between 1 and 3 months
• Project start and end: start by 1 April 2027 and end by 30 June 2027
• Funding location: Any organisation receiving funding must carry out its project work in the UK, intend to exploit the results in the UK, and spend most of the funding within the UK.
Eligibility
• To work alone, your organisation must be a UK registered micro, small or medium sized enterprise (SME).
• Subcontractors are not allowed in this competition.
• An eligible organisation can only lead on one application. Any further applications by the same organisation as lead will be made ineligible.
• Eligible projects must have total costs between £50,000 and £150,000; last between 1 and 3 months; start by 1 April 2027; and end by 30 June 2027.
• Any organisation receiving funding must carry out its project work in the UK, intend to exploit the results in the UK, and spend most of the funding within the UK.
• Up to 70% of costs can be covered, depending on business size.
• Technology Readiness Level: Projects must start at TRL 1, 2 or 3 at the time of application and are expected to make an advance of 0.5 to 1 TRL during project delivery.
Scope and Required Outcomes
• The aim of this competition is to advance the development of novel reusable AI mechanisms that will form the foundations for next-generation AI and machine learning systems.
• Projects must drive development of a new core AI or ML technology by establishing experimental evidence for a new architecture, learning process, representation or control mechanism.
• Developed technology must have the potential to underpin capabilities, products, platforms or services across multiple future applications or markets.
• Applications should show a clear route to defensibility, such as protectable IP, proprietary data advantage, specialist know-how or other credible barriers to entry, with a path to scale.
• Projects are not required to develop a complete commercial product. They must demonstrate the technical feasibility of the core mechanism, architecture or critical subsystem being developed.
• Proposals must demonstrate that AI or ML innovation is the core technical contribution and principal source of any potential competitive advantage likely to be generated if successful.
• Projects must deliver experimental validation of the hypothesis; a clear validation methodology against predefined metrics; evidence of technical novelty and a technical asset; a scaling rationale for the technology and business model; and a technical white paper.
• To be in scope, applications must sufficiently describe technical novelty and hypothesis-driven experimentation; what technical asset the project will create; how it could underpin future products, platforms or services; and the justification of TRL classification.
Specific Themes
• Theme 1. Resource and data adaptive learning: Learning mechanisms that reduce dependence on large datasets, repeated full model training or unnecessary computation by adapting what information is acquired, retained or updated according to uncertainty, task requirements or expected information value. You must focus on one of the following priorities:
- active and probabilistic learning, including uncertainty aware selection of informative data, experiments, simulations or interactions
- continual and selective learning mechanisms that enable new knowledge or capabilities to be acquired without repeated full model retraining
- data efficient learning under limited or expensive observations, using probabilistic representations, structured priors or uncertainty to improve learning efficiency
• Theme 2. Causal and intervention grounded models: Mechanisms that learn, discover or exploit causal and counterfactual structure to improve generalisation, prediction, intervention selection or decision making when underlying conditions or data generating processes change. You must focus on one of the following priorities:
- causal representation learning
- interventional or counterfactual learning
- causal discovery under partial observability
• Theme 3. Learned modular computation and routing: Learning mechanisms that discover, select, compose or reuse internal computational components, enabling models to dynamically route computation and reuse learned functions across tasks, environments or novel combinations. You must focus on one of the following priorities:
- learned modular architectures and computational routing
- reusable learned functional primitives or subroutines
- adaptive composition and reuse of learned computational components
• Theme 4. Geometric, equivariant and physics structured learning: Learning mechanisms that exploit or discover geometric, symmetry, relational or physical structure to improve data efficiency, generalisation, robustness or consistency beyond what can be achieved by learning these regularities from data alone. You must focus on one of the following priorities:
- equivariant or invariant learning architectures
- geometric and relational learning on structured or non-Euclidean domains
- physics structured learning, including methods incorporating physical symmetries, constraints, conservation structure or differentiable dynamics
• Theme selection requirement: Proposals must fall within one or more themes and must focus on at least one specific priority within each theme selected. If an application does not align with the theme and specific priority selected, it will not be sent for assessment.
Exclusions
• Innovate UK is not funding projects that: do not sufficiently provide clear background IP ownership or rights; do not sufficiently provide a specific defensibility route; do not align with the competition theme and specific priority areas; do not sufficiently provide a named baseline with metrics and numeric targets for validation; do not provide a falsifiable hypothesis driven experimentation plan; or do not sufficiently address the scope of the competition.
• Additionally excluded are projects that: are primarily literature review studies, requirement gathering, without substantive experimental research and development; propose routine integration, deployment, orchestration or productisation of existing AI tools or third party models without novel AI or ML development; are not delivering measurable and specific objectives; are primarily routine integration or deployment of existing AI tools without substantive technical innovation; don’t have clear technical novelty and feasibility challenge; do not result in defensible foreground IP; focus primarily on non-AI or non-ML research and development, including cross-cutting projects; or focus on prompt engineering, generic agent orchestration, routine Graph Neural Network (GNN) application, or passive only agentic layers.
Briefing and Support
• Online briefing: Innovate UK will hold an online briefing at 11am on Friday 16 October 2026. Click here for the joining link (Zoom required). A recording and slides will be available afterwards.
• Contact: Contact Innovate UK’s Robotics & AI team for queries.
In summary, this funding opportunity supports UK-registered SMEs to validate novel, reusable AI mechanisms across defined technical themes, with clear eligibility boundaries, structured project requirements and dedicated briefing support. Prospective applicants should review the theme priorities, exclusion criteria and validation requirements carefully before preparing their proposals.