Innovate UK Funding Opportunity: Efficient, Structured and Controllable AI Systems

Innovate UK has announced the Efficient, Structured and Controllable AI Systems competition, offering up to £1.66 million in grant funding to support UK SMEs developing the next generation of artificial intelligence and machine learning technologies.

The competition aims to advance novel, reusable AI mechanisms that can overcome the limitations of existing AI architectures and learning systems. It focuses on technologies capable of improving computational efficiency, adaptability, structured reasoning and controllability, creating foundations for future AI-powered products, platforms and services.

This opportunity is particularly relevant to research-intensive AI start-ups, deep-tech SMEs, university spinouts and pre-commercial companies developing proprietary AI technologies.

Key Dates and Funding

Competition Opens: 12 October 2026

Competition Closes: 18 November 2026, 11:00 am UK time

Total Funding Available: £1.66 million

Eligible Project Costs: £50,000 to £150,000

Project Duration: 1 to 3 months

Project Start: By 1 April 2027

Project End: By 30 June 2027

Funding Rates:

Micro and small enterprises: Up to 70% of eligible project costs
Medium-sized enterprises: Up to 60% of eligible project costs

Funding will be awarded as a grant for eligible feasibility studies and industrial research projects. Applicants must provide the remaining project costs through their own contributions or other eligible funding sources.

What Does the Competition Support?

The competition seeks to support the development and experimental validation of novel AI and ML mechanisms that could underpin future generations of intelligent systems.

Projects must demonstrate a meaningful technical advance in the underlying AI architecture, learning process, representation or control mechanism, rather than simply applying existing AI tools to new applications.

Supported developments may include:

New learning mechanisms that improve how AI systems acquire, retain and update information.
Adaptive architectures capable of dynamically adjusting computational processes.
Causal and counterfactual models that improve prediction and decision-making.
Modular AI systems that reuse learned computational components across different tasks.
Geometric and physics-informed learning methods that improve generalisation and robustness.

Projects are not required to develop a complete commercial product. Instead, they must establish experimental evidence demonstrating the technical feasibility of the proposed AI mechanism.

The resulting technology should have the potential to support multiple future applications, products or markets.

Priority Themes

Applicants must align their projects with at least one of the four competition themes and select at least one specific priority within each chosen theme.

Theme 1: Resource and Data Adaptive Learning

This theme focuses on AI mechanisms that reduce dependence on large datasets, repeated model training and unnecessary computation.

Priority areas include:

Active and probabilistic learning, including uncertainty-aware selection of informative data, experiments or interactions.
Continual and selective learning that enables AI systems to acquire new knowledge without repeated full model retraining.
Data-efficient learning under limited or expensive observations.

Theme 2: Causal and Intervention Grounded Models

This theme supports technologies that improve AI systems' ability to understand causal relationships, predict outcomes and make decisions under changing conditions.

Priority areas include:

Causal representation learning.
Interventional and counterfactual learning.
Causal discovery under partial observability.

Theme 3: Learned Modular Computation and Routing

This theme targets AI architectures capable of selecting, combining and reusing computational components to improve efficiency and adaptability.

Priority areas include:

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

This theme supports AI systems that incorporate geometric relationships, physical principles and structural constraints into their learning processes.

Priority areas include:

Equivariant and invariant learning architectures.
Geometric and relational learning on structured or non-Euclidean domains.
Physics-structured learning incorporating physical symmetries, constraints, conservation structures or differentiable dynamics.

Who Can Apply?

The competition is open to UK-registered micro, small and medium-sized enterprises (SMEs).

Eligible applicants must meet the following requirements:

Be a UK-registered SME.
Carry out the funded project activities in the UK.
Intend to exploit the project results within the UK.
Spend most of the funding within the UK.
Submit no more than one application as the lead organisation.
Demonstrate a clear route to protecting and commercially exploiting the resulting technology.

Subcontractors are not permitted under this competition.

Applicants must also meet Innovate UK's financial viability and subsidy control requirements.

Technology Readiness Level Requirements

Projects must begin at Technology Readiness Level (TRL) 1, 2 or 3, reflecting the early-stage nature of the competition.

Innovate UK expects projects to achieve approximately 0.5 to 1 TRL of advancement during the funding period.

Applicants must justify their starting TRL and demonstrate how the proposed experimental activities will advance the technology.

The emphasis is on validating fundamental AI mechanisms and resolving technical uncertainties, rather than developing market-ready products.

Expected Project Outcomes

Successful projects must deliver the following:

Prototype development: A prototype algorithm, architecture or critical subsystem demonstrating the proposed AI mechanism.
Experimental validation: A structured testing environment and methodology with predefined performance metrics.
Technical evidence: Quantified experimental results against named baselines or hypotheses, including ablation studies and negative controls.
Innovation defensibility: Evidence of technical novelty and a credible strategy for protecting the resulting intellectual property.
Future development plan: A clear rationale for further technological development, scalability and commercial exploitation.

At the end of the project, applicants must submit a technical white paper outlining the work undertaken, experimental results, technical limitations, remaining risks and innovation defensibility strategy.

Key Considerations for Applicants

Applicants should develop proposals that clearly demonstrate the following:

Technical novelty: A significant advance beyond existing AI methods and architectures.
Experimental rigour: A falsifiable hypothesis, appropriate baselines and measurable validation criteria.
Feasibility: A realistic work plan capable of delivering meaningful experimental evidence within one to three months.
Intellectual property: A clear route to ownership, protection and defensibility of the resulting innovation.
Scalability: Potential for the underlying AI mechanism to support future applications, products or platforms.
Commercial potential: A credible pathway towards future development and market exploitation.

Innovate UK will also apply a portfolio approach, seeking to fund a variety of technologies, markets, technological maturities and research categories.

The competition guidance indicates an estimated 4% chance of success, based on experience from similar competitions. This highlights the importance of preparing a technically rigorous and well-supported application.

Potential for Future Funding

This competition represents Phase 1 of a potential two-phase programme.

Subject to future business case approvals, Innovate UK may introduce a second funding phase to support further development of successful technologies.

However, no funding has currently been allocated for Phase 2, and applicants should not assume that further support will become available.

How Inventya Can Help

Developing a competitive application for this opportunity requires more than demonstrating an innovative AI concept. Applicants must establish a compelling technical hypothesis, measurable experimental objectives, a credible intellectual property strategy and a realistic pathway towards future commercialisation.

Inventya works with innovative SMEs to develop funding applications that communicate technical novelty, commercial potential and alignment with funder priorities.

Inventya's support includes:

Eligibility and scope assessment: Reviewing the proposed AI technology against the competition's eligibility requirements, TRL expectations and four priority themes to identify the most appropriate positioning.
Technical proposal development: Supporting the preparation of a clear, evidence-based application that explains the novel AI mechanism, underlying technical uncertainty, proposed architecture and anticipated technological advances.
Experimental validation planning: Helping applicants structure their experimental methodology, define measurable success criteria, identify suitable performance benchmarks and demonstrate how the project will validate its technical hypothesis.
Intellectual property and defensibility strategy: Supporting the articulation of background IP ownership, expected foreground IP, competitive differentiation and credible barriers to entry.
Commercialisation and scalability planning: Developing the commercial rationale for the technology, identifying potential applications and markets, and outlining a realistic pathway towards further development and commercial exploitation.
Project planning and financial preparation: Supporting the development of realistic work packages, milestones, deliverables, risk assessments and project budgets aligned with Innovate UK's funding requirements.

With experience in innovation funding, technology commercialisation and strategic business development, Inventya helps AI-focused SMEs translate early-stage technical research into compelling, evidence-led funding proposals.

Innovate UK Funding Opportunity: Efficient, Structured and Controllable AI Systems

The latest funding, scale up and business growth tips, straight to your inbox

Commercial Services

Innovate UK Business Growth (previously Innovate UK EDGE)

Contact

Social Links

Other