UK Launches £100M Sovereign AI Procurement Competitions for Public Services

The UK government launched the first procurement competitions under its £100 million Sovereign AI R&D Procurement Scheme on August 31, 2026, opening four challenges that invite British AI startups to build systems for the NHS, defence, computing efficiency, and AI agent security. Chancellor John Healey MP announced the launch at the G20 Finance Ministers and Central Bank Governors meeting in North Carolina, according to a joint release from HM Treasury and the Cabinet Office.

The scheme is the latest part of Sovereign AI, a government unit designed to help British AI companies start, scale, and win globally. It targets a specific barrier in public procurement: smaller companies are frequently locked out of contracts with national and local government because they lack the turnover, cash reserves, or track record of larger firms. The scheme removes those barriers, the government said, with upfront payments available where appropriate to give early-stage founders the funding they need to build and prove their technology.

Successful companies will keep the intellectual property they create, allowing them to take innovations beyond the pilot stage and develop commercial products for customers in the UK and abroad.

The Four Challenges

The first competition, run with the Department of Health and Social Care, asks firms to develop AI systems that can automate workflows, coordinate care, and support decision-making across health services. The government said the challenge supports the ambitions of the NHS 10 Year Health Plan while improving productivity and staff experience.

The second, led by the Department for Business, Innovation, Science and Trade and ARIA’s Scaling Inference Lab, will back technologies that improve the efficiency of AI computing infrastructure. The stated goal is to help deliver the government’s ambition to significantly expand the UK’s public AI compute capacity while reducing costs for researchers and businesses.

The third challenge, working with the Ministry of Defence, tasks firms with developing solutions that securely connect data and frontier AI capabilities across defence systems. The fourth, run in partnership with the National Cyber Security Centre, will support technologies that help organisations understand, manage, and mitigate the risks associated with increasingly capable AI agents — work the government framed as helping unlock the safe adoption of AI across the economy.

The competitions focus on demonstrator-stage technology: solutions that have moved beyond early research and development and are ready to be tested in real-world operational environments before wider deployment. By becoming an early customer, the government said, Sovereign AI will help companies cross one of the hardest stages in their growth — turning a promising prototype into technology with a proven customer and a credible route to wider adoption. Successful firms will work with government departments to develop demonstrator-stage technologies with the potential to scale across the public sector and beyond.

Ministers on the Scheme

“Britain is home to some of the most innovative AI companies in the world, and this government is backing them to start, scale and succeed here in the UK,” Healey said in the announcement. He described the competition as first-of-its-kind and said he was determined Britain would play a lead role among G20 countries in harnessing AI for jobs, public services, and growth.

AI Minister Kanishka Narayan pointed to the scale of public procurement as the scheme’s underlying rationale. “Every year the British state spends billions procuring products and solutions tackling some of the most important challenges facing society – from health, to our national defence,” Narayan said, adding that the scheme would open those opportunities to British AI innovators.

Sovereign AI’s Operating Model

Sovereign AI was launched on April 16, 2026, by Technology Secretary Liz Kendall, who described it in a launch speech as combining the speed of venture capital with the strength of the nation state. The unit is chaired by James Wise, with Josephine Kant as head of ventures, and operates with an investment committee that makes its own decisions on which companies to back.

The government describes Sovereign AI as different from any previous government-backed unit, acting like a venture capital fund with the muscle of the state behind it. Beyond capital, the unit provides fully funded access to the UK’s largest supercomputers, fast-tracked visa decisions for research and development talent, and legal-fee support for startups that incorporate as UK limited companies. Kendall said in April that the unit had already made two direct investments, the first in AI infrastructure startup Callosum, and had signed five right-of-first-refusal deals covering companies working on Alzheimer’s and Parkinson’s, defence and security, and continuously learning AI agents.

AI Economics Institute Goes International

Alongside the competition launch, Healey confirmed that the UK plans to open the AI Economics Institute to international cooperation with G7 countries, focused on information and evidence sharing. The institute, established as a joint research organisation of HM Treasury and the Department for Science, Innovation and Technology, is chaired by Nobel Prize-winning economist Simon Johnson and follows the operating model of the AI Security Institute.

According to its official introduction, the institute is not a policy-setting body; its role is to build the data infrastructure, analytical capability, and external relationships needed to assess AI’s effects on productivity, labour markets, and growth, and to ensure that evidence reaches policymakers. It incorporates the Department for Science, Innovation and Technology’s Future of Work Unit and has agreed a joint statement of collaboration with Anthropic, Google, OpenAI, and Microsoft.

Applications for the competitions will be assessed by the Sovereign AI team, participating government departments, and independent technical experts against published competition criteria. The first four competitions are intended to test the delivery model, with further challenges expected in future phases of the scheme as the programme expands.