Mobilising the Nation

Taskforce members contributed their expertise, evidence and challenge to the report; membership of the taskforce does not indicate endorsement of the report's analysis or recommendations:

Chair

• Steve Rigby, CEO, Rigby Group

Members

• Abdul Qureshi, Managing Director, Business Banking, Barclays
• Amanda Grant, Chief of Strategic Ventures – Data & AI, OneAdvanced
• Anthony Hills, Regional Director, UK & Ireland, NVIDIA
• Colin Bannon, Chief Technology Officer, BT Business, BT
• Darren Hardman, Corporate Vice President & CEO, Microsoft UK, Microsoft
• Dom Hallas, Executive Director, Startup Coalition
• Hugh Milward, Vice President, Corporate Affairs, Microsoft
• Imran Shafi OBE, UK Managing Director, Nscale
• Johanna Hutchinson, Chief Data Officer, BAE Systems
• Karl Havard, Chief Commercial Officer, Era4
• Max Beverton-Palmer, Head of Public Policy, UK, NVIDIA
• Sarah Walker, Chief Executive, UK & Ireland, Cisco
• Simon Walsh, CEO, OneAdvanced
• Stephen Brown, Head of Barclays Eagle Labs, Barclays
• Polly Mackenzie, Co-founder & CEO, Zinc Innovation Partners

Secretariat

Zinc Innovation Partners provided the secretariat for Project SAIL.

The world is moving at an unprecedented pace. It is more than two years since the AI Opportunities Action Plan was commissioned and more than 21 months since it was published. Much of the plan remains relevant, but the technology, the investment landscape and the geopolitical context have moved rapidly. The UK’s priorities must move with them.

Through our SCC business, Europe’s largest independent IT services provider, we see customer demand, investment priorities and emerging technology trends at close quarters. We have used that network to bring together the most senior leaders from across the UK technology industry, including individuals operating at the frontier of AI, infrastructure, data and digital services. Some have contributed publicly, others privately.

This work was commissioned by Rigby Group, with Zinc Innovation Partners acting as secretariat, and its conclusions draw on the experience of businesses investing in, building and deploying AI.

Our starting point is the UK’s economic strength and resilience.

The UK will not lead every part of the global AI economy, nor does it need to. But it can secure a globally significant position in the deployment, adoption, application and design of AI. That requires deliberate choices about the capabilities the UK needs.

Compute is fundamental. The UK needs enough AI infrastructure on its shores to support economic growth, public services and national resilience. This is not an argument for buying British. Much of that infrastructure will be built and run by international partners, and that is a strength. It is an argument for making sure the UK has the capacity, the continuity of access and control that its most critical services need.

That requires a clearer, practical definition of sovereignty. Ownership is only one consideration. Location, operational control, legal jurisdiction, supply chains, connectivity and energy all matter, and different uses need different levels of assurance. Government should give buyers a framework for judging which workloads need the highest level of control, and which can be met by the best services available, wherever their providers are based.

Skills are equally important. AI adoption across our services-led economy is a major productivity opportunity, worth around £75 billion a year by 2035. Capturing that opportunity will require a much faster approach to workforce adoption, technical skills and organisational transformation.

The state also holds two extraordinary assets: its data and its purchasing power. Used intelligently, alongside the capabilities of UKRI, including Innovate UK, they can stimulate adoption, improve public services and strengthen the UK’s technology ecosystem.

None of this can be delivered by government alone. It is clear we need to shift from mobilising teams in government to mobilising the whole nation. But that requires a shift in mindset as well as skillset. A government that collaborates first, not last. Businesses ready to adapt and share. A national commitment to ensuring that our citizens benefit from the AI transformation, and feel confident that they will do so.

So this report makes an offer: a Prosperity Compact for the Nation. We make 10 recommendations, each of which a lead part of government could deliver within 12 to 24 months, and against each we set out the role business will play.

The opportunity is substantial, but it will not be realised through ambition alone. It requires clear priorities, coordinated policy, investment and execution.

Project SAIL is intended to support that next phase of thinking. Industry stands ready to invest, contribute and work alongside government to ensure the UK remains at the frontier of this technological and economic transformation.

Steve Rigby
Chair, Project SAIL
CEO, Rigby Group

The development of AI presents a great opportunity for business and the nation. The government has focused significant policy attention on unlocking it. The strategic framework put in place by the 2025 AI Opportunities Action Plan is already producing dividends. But it is time for a reset in the nation’s approach.

Government machinery has been reorganised since the Plan was published, creating new opportunities and new demands on both systems and teams, but the direction of travel has not yet been clarified.

The reward for getting this right is significant: most of the roughly £75 billion a year the OBR’s central productivity forecasts suggest that AI could add to the economy by 2035 depends on business adoption, not solely invention, and a coordinated national effort is required to turn that potential into growth that reaches every region - with the state rebuilding the efficacy, legitimacy and trust that delivery depends on.

But the risks from getting it wrong are severe, and growing. The withdrawal of two frontier AI models in June 2026 showed that UK access to models built abroad can change with little notice. And public opposition to data-centre infrastructure is rising in Britain, following a pattern already visible in the US.

This report, developed with input from private sector technology leaders, is built on a comprehensive review of the implementation record from the government's AI Opportunities Action Plan, published in January 2025. The review shows a consistent pattern: progress has been rapid where government controls the machinery directly, and far slower wherever delivery depends on coordination with other parties. That gap is this report's central diagnosis, and closing it must be the founding task of the Prime Minister’s AI Taskforce. Section Two sets out the evidence in full.

The way through this is not a faster version of the same process, but a wider one, which we describe as a Prosperity Compact for the Nation.

Government cannot secure adoption, build infrastructure or earn public consent by itself. It needs a reciprocal compact with the parties who can help deliver each: business, the places hosting AI's infrastructure, and the working people whose jobs it will change. Section Four sets out that compact, including what each side must contribute.

Our report turns the compact into 10 recommendations. Each names an accountable part of government, sets a delivery window of 12 to 24 months, and is paired with a statement on the role of business. Section Five sets out the recommendations in full.

To reap the rewards of AI in, and for the citizens of, the United Kingdom, we must mobilise the nation into action.

Summary of 10 recommendations

The 2025 AI Opportunities Action Plan was written in different times. The government now faces increasingly pressing and complex strategic challenges. Each of these raises the stakes and changes what government and business need from each other.

First, the machinery of government has been reset. Where the Department for Science, Innovation and Technology once held broad leadership on AI, digital functions are now split across three departments, each with its own budget and priorities. The Minister for Artificial Intelligence, the government's AI Taskforce under Lord Vallance, and the wider Cabinet Office now provide clear leadership from the centre, but reorganisations of this kind inevitably turn attention inwards and consume delivery bandwidth that would otherwise go to implementation. The Action Plan is no longer the mechanism by which progress on AI is tracked, and nothing has yet replaced it. Business and government alike need a revised framework for direction and accountability.

Second, the AI sovereignty debate has matured. In June 2026, the United States government applied export controls that required an American AI company to suspend foreign nationals' access to two newly released frontier models. The company disabled both models for all customers to comply, and restored access 19 days later, after the controls were lifted. One business represented on the Project SAIL taskforce had to spend the following three weeks visiting all their regulated clients to reassure them that no kill switch for their product existed. While the models had only been available for three days, the episode made continuity, fallbacks, and what ‘sovereign’ actually means live questions that the government must help resolve. The European Union has already proposed binding sovereignty rules for cloud and AI. The UK has yet to set out its own position. Resilience and security are not simply a technical question of domestic versus international supply and hosting, but a strategic priority for the UK.

Third, public opposition to critical data infrastructure is growing. In the United States, organised local opposition has stopped fully permitted data centre projects, and moratoriums have been announced even in states that are otherwise pro-growth (Thompson and Sun, 2026). Reporting from those places shows that AI itself is only a surface-level concern; distrust of corporate and government promises, fear of stranded assets, and electricity costs run deeper. The same arguments have begun to appear here. The UK context is specific, higher energy bills and successive dry summers have put data centres’ resource use under particular public scrutiny. Britain has a window - we think of roughly 12 months - to build a compact that earns public consent before that pattern takes hold.

The reward for getting this right is real. The government's stated ambition is for Britain to be the fastest adopter of AI in the G7, with benefits reaching every region. Microsoft's diffusion measure currently places the UK eighth in the world, with just over two in five adults using a generative AI product - but the rate of growth is slowing while several Asian economies accelerate (Microsoft, 2026). The Office for Budget Responsibility’s central scenario is that AI could raise UK productivity by around 2.3% by 2035 (OBR, 2025) - equivalent to roughly £75 billion a year in today’s prices¹, before the costs of investment are taken into account. Most of that depends on adoption - people and organisations using the technology well, in practice - rather than invention. Access to skills is one of the principal barriers to AI adoption: around a fifth of people in work feel confident using AI in the workplace, and only 11% of employers have had staff undertake AI training in the last 12 months (DCMS, 2026).

It is time for a new approach.

1 Our calculation, applying the OBR’s central productivity estimate to projected UK GDP in 2035

The assessment

The AI Opportunities Action Plan was published in January 2025 and the government accepted all 50 of its recommendations. In its One Year On update, published in January 2026, the government stated that it had met 38 of the 50, with 12 in progress.

We conducted our review 19 months after the plan launched. We examined each recommendation independently and compared the government's account of its own progress against evidence from independent and third-party sources. The purpose was to separate announcements, funding commitments and appointments from implementation and effect. We used four assessment categories:

Across the 50, we recorded 14 as substantial progress, 27 as early activity, six as weak or short of the ask, and three as lacking the information to judge.

Underneath those four patterns is a consistent shape. Government is good at standing up a tool with a capable team behind it. The rapid prototyping unit built systems that work, the gated procurement process ran end to end and produced a contract, the AI Security Institute recruited the people it said it would, the citizen-facing chat service is running with users, and independent survey evidence shows a strong rise in basic AI adoption among small firms, which is the clearest positive outcome in the whole review.

The government is weaker on anything structural. The roles, standing services, mandates and reforms intended to make adoption systematic are where delivery thins out. There is no verifiable AI lead for each government mission. There is no single central service for scaling successful pilots, only several separate programmes doing parts of the job. There is no interoperability mandate, only guidance that departments should follow. The National Data Library has produced a curated set of links to already-published datasets, a guidance manual, repairs to the old data.gov.uk directory and five pilot projects, which satisfies the recommendation as drafted and leaves the underlying problem where it was.

We observe that across the board, the ambition of a recommendation is not the determining factor of whether a recommendation could be achieved, but rather the number of parties who had to act on it. We call this a coordination ceiling. Below the ceiling, where one capable team could build a thing and put it into service, delivery is strong. Above it, where a recommendation has required shared infrastructure, several departments moving together, or a change in the law, delivery thins into guidance, strategy documents and pilots. The ceiling, and not the quality of the ideas, is what the Action Plan ran into.

The 50 recommendations

As we have seen, good initiatives on AI can stall when they hit the ceiling beyond which coordination at scale will be required. The importance of coordination has only grown since the Plan was published.

Responsibility for the wider AI agenda, including skills, now runs across at least five departments, with the boundaries between them still being drawn. Skills England sits in the work and pensions department; the curriculum and the apprenticeship levy sit with education; digital skills and technology regulation sit with culture; cross-government adoption sits at the centre; the research funding that would commission new data sits with business and science.

Beyond simply civil service coordination sits the nation as a whole. Adoption at scale is a change in the behaviour of firms that the government does not control - and with devolution accelerating that complexity is increasing. Regional benefit depends on councils, combined authorities, grid operators and developers agreeing terms. Skills provision is devolved, and in Wales and Scotland it is devolved again. Sovereignty, in any of the senses now in circulation, is a property of contracts between buyers and suppliers, not of a strategy document. The state can set the direction for all of these. It cannot deliver any of them alone, and the evidence of the last 19 months is that it has not.

Sprints as a way through

The government's answer to the coordination ceiling is speed. The Prime Minister's AI Taskforce is an empowered unit with a single clear objective, a ring-fenced budget, a high tolerance of risk and a direct line to the Prime Minister (Cabinet Office, 2026). It is understood to be planning its work in sprints of three to six months. Its leadership has been drawn from the parts of the AI system that delivered best under the Action Plan.

The Taskforce is said to be inspired by the Vaccines Taskforce of 2020 and commentators have urged it to similarly adopt one mission and pursue it with the powers of direction the vaccine programme enjoyed (Coleman and Lawrence, 2026). This is a powerful model for well-defined and well-boundaried problems - and it is a mode government has proven it can work well in and deliver results through.

Where sprints need support

Sprints suit problems that can be bounded, and many start-ups and the businesses that work with them would not want government to lose the habit. But the vaccine programme was of two parts. Buying the vaccine was a sprint; getting it into arms was a mobilisation, delivered through institutions and relationships that already existed.

AI adoption is an execution problem of the second kind: changes in the behaviour of two million firms, several hundred local authorities, dozens of regulators and the public. We also need an operational capability that can:

• Build standing services. A standing service, a mandate or a reform requires several parties to accept a permanent change in what they do. A temporary unit with a temporary remit has neither the authority nor the lifespan to secure that. A taskforce sprint can write the service agreement. It cannot hold the other side to it in three years' time.
• Secure consent. The next phase of AI in Britain runs through planning committees, grid connection decisions, NHS Trust procurement panels, and the boardrooms of mid-sized firms. Each of these decisions will be taken by people weighing the costs of AI today against benefits they cannot yet see. Consent is built by parties the community already trusts. Speed at the centre, applied to a decision being made elsewhere, mostly produces friction.
• Mend fragmentation. A unit at the centre of government that depends on other departments to deliver is only as fast as the slowest of them. The AI Taskforce's advantage is its line to the Prime Minister, but this attention cannot be assumed over the medium term (Key, 2026). Capability needs to offer benefits beyond leadership attention to hold together a broad coalition around a significant transformation.

Mobilisation as the way through

The Taskforce’s sprints and a wider mobilisation can work together: the Taskforce can set direction and move quickly where it controls the levers, while the compact brings in the businesses, places and institutions needed to deliver at scale.

The question then is not how to speed up, but how to form a new mode of government operation. This is a question about incentives and trust as much as levers and mechanisms, and it is one that business is well placed to help answer. It is a project of shared mobilisation. That is why Project SAIL recommends a new approach built on collaboration. We call it the Prosperity Compact for the Nation, and we set out its logic, and what it should contain, in the next section.

In the next phase of the UK’s AI journey, the government needs to change how systems beyond its control behave. But mobilisation cannot be commanded. Businesses will not redesign how they work, workers will not trust new systems, and communities will not welcome new infrastructure because the government has told them to. They will do it because they have a reason of their own to want it to succeed - willingness, incentives, and beyond either, active enthusiasm and commitment.

It is no great surprise that we see scepticism about the benefits of AI adoption. Britain has a long record of world-class research that is scaled and commercialised elsewhere. DeepMind and Arm, the best-known examples from the AI era, are both now foreign-owned. If AI more generally follows that pattern, there will be little here to share: no domestic surplus for firms, regions or workers to hold a stake in, and no basis for the bargain the compact is asking business and government to make. Only a country that captures a share of the AI value chain can afford to fund the returns to society that will establish and maintain consent.

Previous periods of rapid technological growth did not settle distributional questions by winning the argument. They settled them through bargains. The postwar European settlement rested on a broadly corporatist arrangement in which labour accepted wage discipline, capital accepted taxation and codetermination in various national flavours, and the state built the welfare systems that made the whole thing tolerable. The lesson is not that Britain should return to 1950s corporatism. It is that the fast diffusion of a general-purpose technology has historically required an explicit bargain about who gets what from the growth it produces.

That is why we propose a Prosperity Compact: a reciprocal agreement between government, business, places and citizens, structured to start a virtuous circle of capturing value for the nation. What follows sets out that circle, and the compact that sets it moving.

The virtuous circle

1. Build well. We build resilient and secure AI infrastructure and drive adoption in ways that create value in the short and long term, mobilising collaboration between public and private sectors.
2. Capture. Britain keeps a meaningful share of the value AI creates. Companies scale here, compute revenue lands in the regions hosting infrastructure, and enterprises retain the intellectual property and learning generated on top of third-party models.
3. Stake. Captured value pays for visible benefits: higher wages, skills investment, retained business rates, and lower energy bills near data centres.
4. Consent. Visible benefits change who argues for a project and who argues against it. Residents who see lower bills, councils that retain business rates, citizens that see environmental responsibility, and workers who see wage growth have a reason to back a data centre or grid connection at the planning meeting - enabling us to continue to invest in building well.

Where to begin

A compact of this kind is not built in a single report or a single year. It is a standing practice of dialogue and consent-building that has to run for as long as the AI economy keeps changing what it asks of people, places and firms. What follows are proposals for the first 12 months of that practice: 10 recommendations where government can act to mobilise and incentivise wider change.

The compact has three interfaces: between the state and business, between the AI economy and places, and between the AI economy and people. At each interface there is a reciprocal relationship between what business brings and what government can deliver.

It starts with building well, stage one in the virtuous circle we aim to set in motion. Before there is anything to share, there has to be something built well enough to be worth sharing. That is why the recommendations that follow lean so heavily toward building: definitions, procurement, data, compute, resilience and adoption support.

Even if we get this right, capture remains the critical challenge, where the circle is most fragile. Below, we seed the earliest, most tractable form of capture - planning gain for communities hosting data infrastructure - as well as recommending building new collaborative capacity to understand skills and the labour market to minimise negative effects of AI on employment.

Nevertheless, securing Britain's share of a global value chain and ensuring that citizens benefit from it will be an ongoing challenge for decades. How AI-generated activity is owned and taxed is one of the central questions governments must answer in the years to come. That dialogue must happen in partnership with business, and with one critical thought in mind: trust is the scarce resource; a visible stake in prosperity for firms, regions and workers is how trust is built; and without it there will be no gain for business or citizens.

The following 10 recommendations set out a roadmap for the AI Taskforce and the Minister for Artificial Intelligence to mobilise AI for the nation within the next 12 months. The recommendations initiate the compact by building well, and starting the processes that will enable us to capture value from AI for people, places and the nation as a whole.

Measuring success

Our review found that progress on the Action Plan was mostly reported as announcements and funding, with little evidence of impact. The next phase needs a clear measure of success. Government’s estimates of AI’s value to the UK economy vary widely, from the OBR’s cautious central forecast to DSIT’s £55-140 billion by 2030 (DSIT, 2026d). None of these estimates net off the costs of investment. The AI Economics Institute, working with the OBR and ONS, should publish a single methodology for measuring AI’s net contribution to the economy, including how widely the gains are shared, and report against it annually.

We welcome the progress the Government has made on AI. We are backing the recommendations in this report, and in particular two that resonate with what we are hearing from our customers; a clear definition of sovereign AI, and a second stage of support for firms that have started adopting AI and stalled.

A shared and simple definition will educate buyers and unblock procurement, which today can be held up by uncertainty over what ‘sovereign’ means. Secondly, too many organisations are stuck in trials and pilots. How can we enable adoption at scale, at pace and safely for our customers?

We urge the government to act on the recommendations in this report so that adoption accelerates at the pace the opportunity demands to get the UK AI Economy moving at pace.

Time moves fast in this sector, and the original AI Opportunities Action Plan, with the AI Growth Zones initiatives, now feels outdated. This initiative, from leaders within the UK AI Industry has resulted in 10 strong objectives, that will stand the test of time. They will stimulate the economy, enabling UK Sovereign AI to achieve both regional and national growth across the UK. I hope our politicians take notice.

The opportunity is substantial, but it will not be realised through ambition alone. It requires clear priorities, coordinated policy, investment and execution.

Project SAIL has brought together leaders from across the UK technology industry to take an honest look at what the AI Opportunities Action Plan has delivered, and what the next phase requires. The review shows a clear pattern. Government delivers well where one capable team controls the work, and struggles wherever progress depends on many parties acting together. That coordination ceiling, and not a shortage of ideas, is what now holds the UK back.

There is a narrow window to act. Public consent for new infrastructure is not yet settled, the machinery of government is still being redrawn, and the UK’s access to critical AI capability depends on decisions made elsewhere. Other countries are moving quickly. If the next 12 months are spent on announcements and pilots, the UK risks building the technology here while the value, and the public’s support, goes elsewhere.

The conditions for action are favourable. The Prime Minister’s AI Taskforce gives AI a direct line to the centre of government. The AI Economics Institute provides a new basis for measuring what works. Mayors and combined authorities are gaining the planning and spatial powers to host and shape new infrastructure. The Sovereign AI Fund is investing in British companies. And business is ready to invest, share evidence and work alongside government.

The 10 recommendations in this report set out where to start. Each has one lead part of government, a 12-month timeline and a clear role for business. Together, they begin the Prosperity Compact for the Nation: a reciprocal agreement in which government, business, places and people each contribute, and each see a return.

A compact of this kind cannot be agreed in a single report. It needs a standing relationship between government and the businesses building and deploying AI in the UK. Rigby Group is prepared to convene the Project SAIL taskforce as an external advisory council to the AI Taskforce and the Minister for Artificial Intelligence. It would offer independent advice from industry, test policy as it develops, and report publicly on progress against these recommendations in 12 months’ time.

Mobilising the nation starts now, and it starts with government and business acting together.

Our findings rest on an independent review of progress against all 50 AI Opportunities Action Plan recommendations between June and August 2026. This was built from documentary analysis of government announcements and publications, parliamentary material, and independent and third-party sources. Each recommendation was assessed against what it asked for, not against related activity.

We brought together over 30 leaders from industry partners - startups, civil society and policy specialists - to test and extend that evidence through a cross-sector sandpit on 23 July 2026. We also conducted extensive interviews and discussions with private-sector actors to sharpen our analysis and recommendations.

The steering committee set our priority areas, and several members contributed directly to shaping the recommendations.

Benaich, N. (2026). Europe cannot rent its way to AI sovereignty. June 2026.

Cabinet Office (2026). AI to power change at the heart of government as Lord Vallance appointed chair of new PM AI Taskforce. Press release, 24 July 2026.

CBI and Oliver Wyman (2026). The Adoption Decade: Closing the execution divide and making AI work for Britain. Report of the Accelerating AI Adoption Taskforce, August 2026.

Centre for British Progress (2026). Speeding up grid connections to lower bills and deliver growth.

Coleman, C. and Lawrence, E. (2026). The new AI taskforce should have one mission: leverage. Centre for British Progress, August 2026.

Department for Digital, Culture, Media and Sport (DCMS) (2026). AI Skills for Life and Work: summary report. Ipsos and Warwick Institute for Employment Research, 28 January 2026.

Department for Science, Innovation and Technology (DSIT) (2026a). Report on Copyright and Artificial Intelligence. Statutory report under section 136 of the Data (Use and Access) Act 2025, April 2026.

Department for Science, Innovation and Technology (DSIT) and Government Digital Service (GDS) (2026b). Guidelines and best practices for making government datasets ready for AI. January 2026.

Department for Science, Innovation and Technology (DSIT) (2026c). UK AI Hardware Plan. June 2026.

Department for Science, Innovation and Technology (DSIT) (2026d). AI Skills Boost: explainer. 28 January 2026.

Department for Science, Innovation and Technology (DSIT) and Warwick Institute for Employment Research (2026). AI Skills for Life and Work: labour market and skills projections. January 2026.

European Commission (2026). Proposal for a Regulation on cloud and AI development (Cloud and AI Development Act). COM(2026) 502, 3 June 2026.

Herbert Smith Freehills Kramer (2026). UK grid connections reform: impact on data centres and demand connections.

HM Treasury (2026a). Financial Services AI Adoption Plan. July 2026.

HM Treasury (2026b). £100 million competition to back British AI companies to fix public services. Press release, 31 August 2026.

Innovate UK (2026). About BridgeAI. Programme page.

Institute for Public Policy Research (IPPR) (2024). Transformed by AI: How generative artificial intelligence could affect work in the UK - and how to manage it. Jung, C. and Srinivasa Desikan, B., March 2024.

Key, A. (2026). A short history of DSIT (2023 to 2026). UK20, July 2026.

Lacroix, T. (2026). Interview on Mistral's European compute plan. VentureBeat, August 2026.

London AI and Jobs Taskforce (2026). Shaping how AI changes work in London. July 2026.

Microsoft (2026). AI Diffusion Report, Q1 2026. May 2026.

Office for Budget Responsibility (OBR) (2025). Briefing paper No. 9: Forecasting productivity.

Ofgem (2026). Curate: consultation on managing demand connections for data centres. 29 July 2026.

Phillips-Robins, A., Tawil, T. and Winter-Levy, S. (2026). The Compute Coalition: how to build the future of AI in the free world. Carnegie Endowment for International Peace.

Piggott, J. (2026). UK Public Sector AI Procurement Tracker. Tussell, 11 August 2026.

Rogers, D. and Westgarth, T. (2026). Six Ideas for Andy Burnham's Sovereign AI Policy. Dylan Rogers (Substack), 19 July 2026.

Roodman, D. and Massenkoff, M. (2026). Worker retraining: what the randomised evidence says. Anthropic Economic Index, August 2026.

Science, Innovation and Technology Committee (2026). Science diplomacy: sovereignty, strategy, and the global race. Third Report of Session, House of Commons, 7 July 2026.

techUK (2026). Europe's Technology Sovereignty Package: what do the Cloud and AI Development Act and Chips Act II mean for UK tech?

The Business Desk (2026). Partnership uses Urmston data centre's heat to warm Trafford Leisure Centre pool. May 2026.

Thompson, D. and Sun, J. (2026). The Great American AI Rebellion. August 2026.

UK Government (2025). AI for Science Strategy. November 2025.

About Project SAIL

Project Sovereign AI Leadership (SAIL) taskforce brought together leaders from across the UK technology sector to review progress on the government’s AI Opportunities Action Plan and set out what should come next. It was convened by Rigby Group in June 2026.

About Rigby Group

Founded in 1975 by Sir Peter Rigby, Rigby Group is one of Europe’s largest privately owned technology companies and investors. It is also one of the UK’s top ten wholly family-owned businesses, and in the top 500 globally. The Group is recognised as both an investor and business operator and a trusted partner to government, defence, global corporations, major charities and thousands of medium-sized organisations.

Rigby Group has three main divisions: SCC, the principal trading business, which is Europe’s largest independent IT services provider; Rigby Technology Investments, which is a leading private capital investor in transformative technology businesses; and the Group’s Real Estate portfolio covering hospitality, commercial and residential real estate assets. Through a £300m technology investment programme, the Group is accelerating development in AI, cloud and cybersecurity.

Together, its businesses employ around 8,000 people, operate in nine countries and generate over £4bn in turnover.

About Zinc Innovation Partners

This report was produced by Zinc Innovation Partners (ZIP), and authored by Luke Charbonnier-Bevan and Polly Mackenzie, with support from the wider ZIP team.

All costs associated with its production were met by Rigby Group.