UKRI Transformative Sustainable AI Technologies Grant 2026–2027: Up to £602,500

UKRI Transformative Sustainable AI Technologies Grant 2026–2027: Up to £602,500

The UK Research and Innovation (UKRI) Transformative Sustainable AI Technologies funding opportunity is an open research grant for ambitious ideas that could materially change how artificial intelligence is designed, built, deployed, or maintained. It is aimed at speculative, high-risk, high-reward work rather than incremental improvements to an existing model or a routine sustainability assessment. Projects must be technology-led, start on 1 February 2027, and last for no more than 24 months.

The opportunity is particularly relevant to UK research groups working across algorithms, AI compute, data, software, hardware, systems, and research infrastructure. UKRI also specifically welcomes applications led by eligible research technical professionals, including research software engineers involved in digital research infrastructure. The call is broad enough for interdisciplinary teams, but the central research contribution must advance the underlying technology of sustainable AI. A proposal that is mainly social, legal, environmental, policy, or ethical research without a substantial technical core is outside the stated scope.

Key details

The Funding Finder identifies the opportunity as open and reports a £12,050,000 total fund. In the aim section, UKRI describes an investment of up to £10 million over two years. The page does not explain the difference between those figures, so applicants should use the live Funding Service record and their research office for current financial guidance rather than estimating award numbers from either headline amount.

What the grant is designed to support

UKRI describes sustainable AI as more than reducing electricity use. The scope includes environmental effects such as carbon emissions, water use, and critical-material consumption; social and ethical effects that affect quality of life and communities; and economic sustainability. The funded project must nevertheless focus on novel AI technologies, computational approaches, methods, tools, architectures, or techniques that could significantly improve the sustainability of future AI systems.

Examples in the official call include energy-efficient algorithms, new architectures and systems, smaller task-specific models, AI for resource-constrained environments, hardware and software co-design, efficient processors, and hardware that is easier to repair, disassemble, recover, or update. These examples are prompts, not a closed list. A proposal can take a different route if it explains the technical mechanism by which the work could change the sustainability of AI.

The strongest concepts are likely to connect a concrete bottleneck to an unusual technical intervention. For example, a team might address a limitation in AI compute, data movement, model maintenance, or infrastructure resilience. The project should make clear what is genuinely new, why existing approaches are insufficient, and how the proposed research could establish a new direction. It does not need to promise a finished commercial product during the grant period. UKRI explicitly says projects are not expected to deliver tangible outcomes such as a new model or spinout within two years, and failure of an especially risky research goal will not automatically make the work valueless if it produces useful knowledge.

Who can apply and how teams are structured

The project lead must be based at a UK research organisation eligible for UKRI funding and must satisfy UKRI’s individual eligibility rules. The opportunity is open to research groups and individuals. Digital research technical professionals, including research software engineers, are expressly invited to apply as project lead or project co-lead when they meet the eligibility requirements. UKRI encourages researcher co-leads when they have made a substantial intellectual contribution and will participate in the resulting research.

Teams should cover the expertise needed for the proposed work rather than adding names for appearance. The Funding Service supports roles including project lead, UK project co-lead, international project co-lead where an approved agreement applies, researcher co-lead, specialist, professional enabling staff, research and innovation associate, technician, and visiting researcher. The application should explain what each core member contributes and how the team will work together.

There is a limit on participation: an individual may be involved in no more than two applications to this opportunity, and only one of those applications may name that person as project lead. Joint leadership is allowed, but the application must identify one administrative project lead responsible for submission. Proposed job-sharing arrangements should be justified in the team-capability section.

International researchers cannot apply as ordinary project leads under this call. They may participate as project co-leads through the UKRI-RCN Money Follows Cooperation agreement or the UKRI-IIASA agreement. Other international collaborators, and UK partners not based at approved organisations, should be included as project partners. Confirm the relevant route with the host research office before building the budget around an overseas collaborator.

What UKRI will and will not fund

The full economic cost of the project can be up to £602,500. UKRI will fund 80% of that FEC. The maximum is a ceiling, not a target: smaller proposals are welcome, and every cost must be justified and proportionate to the proposed ambition. The published costs that can be supported include contributions to project-lead and co-lead salaries, research, technical, and project-management staff, research consumables including compute, travel and subsistence, data preservation and sharing, dissemination, estates, and indirect costs.

There is no separate matched-funding requirement from the host institutions beyond the standard 20% FEC. Third-party partners may contribute direct or in-kind support, but reviewers must not treat the level of institutional matched funding as an assessment advantage. A host organisation still has to provide an appropriate research environment and infrastructure for the award’s activities, such as facilities and staff development.

PhD studentship costs are not eligible under this opportunity. Equipment costing more than £25,000 per item, including VAT, is also excluded. Smaller equipment items belong under the directly incurred other-costs heading. If a project depends on expensive equipment, redesign the work around existing infrastructure or confirm whether a partner can provide access; do not hide an ineligible purchase in a general equipment line.

How to apply through the Funding Service

Applications must be made through the new UKRI Funding Service. Je-S is not the submission route. The project lead begins by selecting “Start application” on the Funding Finder page, confirming that they will lead the project, and signing in or creating a Funding Service account. The account setup requires selecting the organisation, verifying an email address, and creating a password.

The application is completed through questions in the Funding Service. Applicants can save answers, work offline, and paste text back into the service. Where the system requests an upload, follow its specific file instructions. The lead should leave enough time to inspect the application in read-only view, send it to the research office for checking, respond to any returned edits, and allow the research office to submit the final version. Only the lead research organisation can submit to UKRI.

The formal deadline is 10 November 2026 at 4:00pm UK time, and the service will not accept applications after that point. UKRI warns applicants to allow for internal institutional deadlines. Since submitted applications cannot be changed, a technically complete form that has not passed institutional checks is not a safe submission plan. Ask the research office for its internal date before writing the final timetable.

Required responses and supporting material

The application includes a 550-word plain-English summary. UKRI says this summary is normally made public and should explain the context, challenge, aims and objectives, and potential applications and benefits. Keep confidential information out of it. Write it so that an assessor can quickly understand both the technical idea and why it matters for sustainable AI.

The Vision and Approach response must be supplied as a single PDF of no more than three sides of A4, single-spaced, in 11-point Arial or an equivalent sans-serif font, with margins of at least 2cm. An additional page may be used for a diagrammatic workplan. The filename must use the unique Funding Service number followed by “Vision and Approach,” and the file must be no larger than 8MB. The response should explain the project’s vision, novelty, relevance, timing, objectives, method, milestones, risks, research environment, and route to impact.

The applicant-and-team-capability section has a 1,500-word limit. UKRI requires the Résumé for Research and Innovation (R4RI) format, with 1,000 words for the R4RI modules including references and, if needed, up to 500 words of additions. The modules cover contributions to new ideas, tools, methods, or knowledge; developing others and maintaining relationships; the wider research and innovation community; and broader users and societal benefit. This is not a conventional publication list. Select past contributions that directly support the proposed work and explain the team’s ability to deliver it.

Other application questions address ethics and responsible research and innovation, resources and cost justification, risk and reward, trusted research and innovation, and related requirements. Read the live Funding Service guidance for the complete question set and any upload instructions. UKRI says applications should be self-contained: hyperlinks may point to reference information, but assessors are not required to follow links to fill gaps in the proposal.

How applications will be assessed

The main assessment areas are vision and approach, applicant and team capability to deliver, ethics and responsible research and innovation, resources and cost justification, and risk and reward. For the vision, assessors look for excellent and important work, new knowledge or discovery, timely motivation, and potential benefit to research, society, the economy, or the environment. The call’s high-risk character means applicants should not dilute a genuinely ambitious idea into a safe but incremental project.

The approach must still be credible. Explain the technical framework, why the chosen methods fit the objective, how milestones will be measured, what could fail, and how the team will learn or adapt. The call specifically says that projects should not focus only on benchmarking existing tools or reproducing what industry already does. It also asks applicants to explain why the work cannot be supported through another route, such as responsive-mode funding.

This opportunity uses distributed peer review (DPR), rather than a project-lead response or panel step. Each application receives up to eight reviews, and applicant teams are expected to review proposals submitted by other teams. UKRI says teams will receive around eight to ten proposals to assess. Reviewers must read, score, and provide feedback under the DPR rules; failure to complete assigned reviews by the deadline can lead to automatic rejection of the team’s own proposal, except in limited extenuating circumstances. Applying therefore means accepting a real review obligation, not simply consenting to be evaluated.

A practical preparation strategy

Start with a one-page technical theory of change. State the current sustainability constraint, the proposed intervention, the part of the AI stack it changes, and the evidence that would indicate progress. Then separate what must be demonstrated during the 24-month project from what may only become possible in later research. This helps the proposal stay ambitious without promising a finished system that the call does not require.

Next, map the team to the workplan. If the project combines algorithms, systems, hardware, and infrastructure, identify who owns each technical risk and who can test the result. Include the research software, computing, data-management, engineering, or experimental skills actually needed. If an external partner supplies a facility or dataset, describe the contribution and confirm that the arrangement is real before submission.

Build the budget from the milestones. Compute costs, staff time, consumables, travel, data preservation, and dissemination should each have a reason connected to an activity. Check the 20% FEC structure with the finance office and remove PhD studentship costs and over-£25,000 equipment. If the project could work at a smaller scale, explain why the selected scale is necessary rather than filling the ceiling.

Finally, prepare for the DPR commitment. Nominate at least two members of the team to participate in peer review, including the project lead, as required by the call. Make sure they have the time and subject knowledge to review other proposals impartially. Also register for UKRI’s applicant webinar on 16 September 2026 if the timing is workable; the official page says the session will provide more information and an opportunity to ask questions.

Common mistakes to avoid

The most common fit problem is presenting an ordinary efficiency project as transformative. A modest reduction in inference cost, a routine benchmark, or an assessment of environmental impact may be valuable but does not by itself meet this call’s emphasis on novel technology and step-change potential. Explain the new mechanism and why it could disrupt an existing research direction.

Another mistake is making the proposal mostly about policy, governance, or responsible use while leaving the technical advance vague. Those issues belong in the responsible-research discussion, but UKRI says the funded work must be technology-led and will not support a majority non-technical project. A social or environmental benefit is important; it cannot substitute for a technical research contribution.

Do not treat £602,500 as an automatic award or as cash available to spend without institutional rules. It is the maximum FEC, with UKRI funding 80%. Do not wait until the deadline day for research-office approval, forget the fixed 1 February 2027 start, or submit a Vision and Approach PDF that breaks the page, font, margin, filename, or file-size rules. Most importantly, do not apply if the team cannot complete the required peer reviews after submission.

Frequently asked questions

Is this a scholarship or personal award?

No. It is a research grant awarded to eligible research organisations for a defined project. It can support eligible staff, compute, consumables, travel, data, estates, and indirect costs, but it is not an unrestricted personal payment.

Can a research software engineer lead the application?

Yes, if the person meets UKRI’s individual eligibility rules and is based at an eligible UK research organisation. UKRI specifically welcomes applications led by research technical professionals involved in digital research infrastructure.

Can international researchers apply?

They cannot apply as ordinary project leads under this call. They may be project co-leads through the UKRI-RCN Money Follows Cooperation agreement or the UKRI-IIASA agreement. Other international collaborators should normally be project partners.

Does the project need to produce a finished AI system within two years?

No. UKRI says tangible outcomes such as new models or spinouts are not expected within two years. The proposal should still set out testable research objectives, milestones, learning, and a credible route for useful outputs to continue after the grant.

Is institutional match funding required?

There is no additional matched-funding requirement beyond the standard 20% FEC. Host support for the research environment and infrastructure is still necessary, and third-party contributions may be useful.

Who should answer questions about eligibility or costings?

For costings and proposal-writing support, UKRI advises contacting the research office first. Questions about the specific opportunity can be sent to [email protected]. Funding Service technical questions go to [email protected] or the helpline listed on the official page.

Official links and next steps

Read the UKRI Transformative sustainable AI technologies funding opportunity in full, including the current scope, cost rules, application questions, and distributed peer-review requirements. Begin through the “Start application” link on that page, which opens the UKRI Funding Service application. Confirm your organisation’s eligibility and internal deadline with its research office, then develop the technical concept, team plan, budget, and Vision and Approach PDF around the published criteria.

As of 26 August 2026, the official page lists the opportunity as open and gives 10 November 2026 at 4:00pm UK time as the final submission deadline. The project must start on 1 February 2027. Recheck the official record immediately before submission for any change to the deadline, application questions, review timetable, or funding-service instructions.