Innovate UK CR&D
£25k to £4m+- Best fit
- UK SMEs with a genuine ML or AI technical unknown
- What it covers
- Feasibility, industrial research and collaborative R&D across any sector.
Artificial Intelligence grants
AI grants UK wide for companies with genuine technical risk
AI grants in the UK have never been more available, or more competitive. Nearly every UK and EU programme now funds AI, which means assessors see hundreds of applications wrapping a commodity model in a new interface. Awards go to teams with a defensible technical unknown, a novel architecture, a hard data problem, or a domain where general models fail. Innovate UK runs regular AI focused competitions including the BridgeAI programme for adoption in high-growth sectors, ARIA backs genuinely frontier approaches, Horizon Europe funds trustworthy and secure AI at European scale, and Dstl funds defence and security applications. The AI Research Resource (AIRR) also gives UK researchers and companies access to national-scale compute through Isambard-AI and Dawn, which can de-risk the technical plan in an application. We help you find and prove the technical unknown that separates a funded project from a wrapper.
£190m+ won by our team
PhD-level specialists who have won these awards themselves.
Funders funding artificial intelligence
'%3e%3cpath%20id='s'%20d='M0,162.5%2022.041947,230.338137%20-35.664619,188.411863H35.664619L-22.041947,230.338137z'/%3e%3cuse%20xlink:href='%23s'%20y='-400'/%3e%3cg%20id='s5'%3e%3cuse%20xlink:href='%23s'%20transform='rotate(30)%20rotate(-30,0,200)'/%3e%3cuse%20xlink:href='%23s'%20transform='rotate(60)%20rotate(-60,0,200)'/%3e%3cuse%20xlink:href='%23s'%20transform='rotate(90)%20rotate(-90,0,200)'/%3e%3cuse%20xlink:href='%23s'%20transform='rotate(120)%20rotate(-120,0,200)'/%3e%3cuse%20xlink:href='%23s'%20transform='rotate(150)%20rotate(-150,0,200)'/%3e%3c/g%3e%3cuse%20xlink:href='%23s5'%20transform='scale(-1,1)'/%3e%3c/g%3e%3c/svg%3e)
- Dstl
Programmes that fit artificial intelligence companies
01
Innovate UK CR&D competitions
Mission-led calls; strongest fit for foundational or applied ML innovation from UK SMEs.
02
Innovate UK BridgeAI
Targeted support and funding for adopting AI in high-growth sectors such as agriculture, construction, creative industries and transport, where a proven model is applied to a genuinely new domain.
03
AI Research Resource (AIRR)
National-scale compute on Isambard-AI and Dawn for UK AI research and development; a credible AIRR allocation strengthens the technical plan and feasibility case in a grant application.
04
Horizon Europe SecureAI
Digital, Industry and Space cluster call for trustworthy and secure AI research.
05
ARIA programmes
Programmatic funding for genuinely frontier, high-risk approaches.
06
Dstl and defence innovation
Defence and security AI awards where dual use technology and UK sovereign capability score well.
Compare AI grants UK: which programme fits your project
Most AI founders apply to the wrong call. This table matches the main UK AI funding routes to the stage and risk profile they actually fund.
Innovate UK BridgeAI
Varies by call- Best fit
- Adopting AI in high-growth sectors such as agriculture, construction and transport
- What it covers
- Sector-specific adoption support, feasibility and innovation vouchers.
ARIA
Programmatic awards- Best fit
- Genuinely frontier, high-risk AI research
- What it covers
- Long-term, high-tolerance research where failure is an acceptable outcome.
Horizon Europe SecureAI
€2m to €4m typically- Best fit
- Trustworthy, secure or responsible AI with EU partners
- What it covers
- Collaborative EU research on privacy, robustness and AI assurance.
Dstl / UK Defence Innovation
Varies by call- Best fit
- Dual-use or sovereign defence and security AI
- What it covers
- Operational AI, autonomy, secure systems and national capability.
AI Research Resource (AIRR)
Compute access- Best fit
- Compute-intensive AI R&D that needs national-scale resources
- What it covers
- Access to Isambard-AI and Dawn supercomputing to de-risk the technical plan.
What assessors actually score
- Why an off-the-shelf model does not already solve this, argued against named alternatives rather than a straw man
- Data access, provenance and evaluation methodology, including how training and test data are separated
- Benchmarks and baselines a technical assessor will respect, with quantitative success criteria
- A credible position on model risk: hallucination, bias, drift and what you will measure in the feasibility stage
- Evidence the team has shipped machine learning systems before, not just used APIs
- A commercial route that survives the question of what happens when the frontier models catch up
What counts as a technical advance in an AI grant application
Every UK funder now asks the same underlying question: what does your project know at the end that the field did not know at the start? Applying a commercial model to a new dataset is integration work and scores poorly. These are the advances that read as genuine innovation to an AI assessor.
A method that does not exist off the shelf
A new architecture, training regime, optimisation approach or evaluation method, described precisely enough that a specialist could argue with it. Name the closest published alternative and state why it falls short.
A data problem nobody has solved for your domain
Scarce labels, regulated or sensitive records, sensor noise, distribution shift in the field. Set out how you obtain, govern and evaluate the data, and what breaks if the assumption fails.
A quantified performance target against a real baseline
Accuracy, latency, cost per inference, energy per token, false negative rate. Assessors want a measured starting point and a target, not a claim of state of the art.
A research question, not a product story
The strongest AI applications read as science, not marketing. Combine AI has won more than £1.4m across ten Innovate UK awards, including a £546,004 Investment Accelerator for next generation meeting efficiency, because every application was framed around a different technical unknown, from the characteristics of virtual meetings to dynamic AI-enabled writing support, rather than re-running the same product pitch.
Uncertainty you are honest about
A funded project has a way of failing. Naming the scientific risk, the go or no go point and the measurement that settles it is what makes a feasibility study fundable.
Assurance built into the work, not bolted on
Bias, hallucination, drift monitoring, explainability and alignment with UK AI assurance expectations. On Horizon Europe trustworthy AI calls this carries direct marks. Holistic AI secured £596,670 for its AI risk management platform by anchoring the proposal in the technical uncertainty of measuring and auditing model risk at scale, and mapping the work to emerging AI regulation so assessors could see the standards work was research, not consultancy.
You can read how each of these played out in practice on the Combine AI and Holistic AI case studies, both recorded on UKRI's Gateway to Research.
Grant funding is non-dilutive: you keep the equity and the intellectual property. For AI companies burning capital on compute and specialist hires, that usually buys a year of runway before a priced round. Check whether your project qualifies.
AI funding results from our team
Our PhD-level writers have won competitive AI funding across assurance, machine learning tooling and applied AI, including more than £1.4m across ten Innovate UK awards for Combine AI, £596,670 for Holistic AI and £158,942 for Bronze AI.
“GrantUp provided clear guidance, support and review of our application and we wouldn't hesitate in recommending them.”
Where we work in artificial intelligence
Machine learning
Computer vision
NLP & language models
Autonomy
Edge AI
AI safety & evaluation
Artificial Intelligence grants in our directory
Live programmes from the UK grant directory that suit artificial intelligence projects. Deadlines move between rounds, so each listing links through to the current detail.
Innovate UK
[IUK] Small medium enterprise champions: Frontier AI – phase 1
The Frontier AI SME Champions competition aims to advance state-of-the-art AI and machine learning in the UK by supporting SMEs to validate ambitious innovations that could deliver step-change capabilities and enable new products, services or platforms.
Individual award £150-250KCloses 11 Nov 2026
View programme Innovate UK
[IUK] Efficient, Structured and Controllable AI Systems
Innovate UK industrial research competition for novel reusable AI mechanisms that are efficient, structured and controllable. Projects start at TRL 1–3. Total eligible costs £50,000–£150,000.
Total eligible costs £50,000–£150,000.Closes 18 Nov 2026
View programme EIT Urban Mobility
[EIT Urban Mobility] Strategic Innovation Open Call - Mobility data management
Open to consortia of innovators, including businesses, researchers, educators and cities, working on urban mobility solutions. This is a multi-beneficiary call requiring a minimum of two independent legal entities working together, established in two different European Member States and/or Third countries associated with Horizon Europe.
Up to 2 million EURCloses 1 Dec 2026
View programme Innovate UK
[IUK] BridgeAI Deployment Catalyst: AI for Assurance and Compliance
Expected to open. Innovate UK BridgeAI Phase 1 strand for AI systems that monitor, test, verify, validate or evidence compliance. UK SME AI suppliers with an MVP work with a UK deployer. Phase 1 grant request £100,000–£200,000; later phases are invite-only.
Phase 1 grant funding request £100,000–£200,000.Closes 8 Dec 2026
View programme Innovate UK
[IUK] BridgeAI Deployment Catalyst: AI for Decision Critical Systems
Expected to open. Innovate UK BridgeAI Phase 1 strand for AI systems that help people make high-risk or high-consequence judgements. UK SME AI suppliers with an MVP work with a UK deployer. Phase 1 grant request £100,000–£200,000; later phases are invite-only.
Phase 1 grant funding request £100,000–£200,000.Closes 8 Dec 2026
View programme Innovate UK
[IUK] BridgeAI Deployment Catalyst: AI for Operational Systems
Expected to open. Innovate UK BridgeAI Phase 1 strand for AI systems whose primary purpose is to perform, control or coordinate operational activity. UK SME AI suppliers with an MVP work with a UK deployer to evidence real-world deployment readiness. Phase 1 grant request £100,000–£200,000; later phases are invite-only.
Phase 1 grant funding request £100,000–£200,000.Closes 8 Dec 2026
View programme Relevant client wins
Grants our team has won for companies solving comparable technical problems.
Women in Innovation Award
Award won
Highly competitive national award won
TextMine
AI / software: document intelligence
Read the case study
Innovate UK grant
£158,942
For Creator-Assist tool
Bronze AI
AI / software: interactive machine learning
Read the case study
Innovate UK grant
£331,000
For AI diversity impact tool
HR Data Hub
AI / software: workforce analytics
Read the case study
Innovate UK grant
£596,670
Awarded for AI risk management
Holistic AI
AI / software: AI assurance
Read the case study
Innovate UK Smart Grant
£341,810
Smart Grant secured
Neural Voice
AI / software: conversational AI
Read the case study
Innovate UK (multiple competitions)
£1.4m+
More than £1.4m of Innovate UK funding across ten awards
Combine AI
AI / software: productivity AI
Read the case studyR&D tax relief for AI and software teams
Model development, data pipelines and work to resolve technical uncertainty in AI systems can qualify for R&D tax relief. Our R&D tax credit specialists for AI and software plan both together. Use the R&D tax relief calculator to see what your spend could return, or read accountant vs R&D tax specialist.
Fund your artificial intelligence R&D
Talk to us about your grant funding plans and we'll help you make it happen through effective, results-driven grant writing.