This is my third quarterly update on our progress at Sovereign AI. As before, this represents my own views on our work.

Sovereign AI was set up with a simple thesis: adopting and developing frontier AI is a strategic national capability, and so the UK must be home to more companies building critical parts of the global AI supply chain. Alongside improving the general business environment, we believe the Government can do this by investing directly into British frontier AI companies, and by providing strategic support only the Government can offer so that the best teams can build and scale here to compete globally.

This is the first full quarter since our launch in April, and one where we have hit our targets on the investment side and team build out. We have now invested in, or are in final legals with, 12 UK AI companies, with 5 announced so far. We have an enviable 100% win rate and our portfolio companies have raised a combined £4B. More on this below.

It’s also been a strong quarter for the wider British AI ecosystem. UK AI companies raised a record £10 billion in the first half of 2026. That is more than four times what was raised in the same period last year. The UK attracted 39% of all European venture investment and now has multiple teams that have raised over $1B building frontier-models in fields as diverse as material science discovery to autonomous vehicles.

And it’s also been a strong quarter for the impact of these new technologies across the UK. British Chambers of Commerce research found that a vast majority of SMEs now actively use AI, with adopters seeking significant productivity gains. There’s been breakthroughs in adoption in the public sector too. Kent and Canterbury Hospital became among the first in the NHS to use an AI tool that spots infections before they take hold. A survey of 1,000 NHS healthcare professionals published this month found nine in ten are already using AI in their work to cut through the 1.3 hours (!!) of admin they average on top of a full day of patient care, with seven in ten saying it is reducing their bureaucratic burden. In June, Extract, an AI tool built by the government’s AI incubator, was launched to every council in England, and a companion tool is now being tested in three councils to half the time a routine household planning decision takes. This is the everyday impact we want to see, and can all benefit from.

Globally, this quarter has been one of increased capabilities and demand, alongside mind-melting breakthroughs in maths in particular. And while performance of the frontier models improves, many use cases are also getting cheaper, thanks to the continued competitiveness of open-weight models (which are generally, if not always cheaper to run), more efficient approaches to compute and a competitive set of frontier model companies dropping prices over time.

So, by any account, generative AI models and the chips that run them continue to be one of the most important areas of investment to improve our health and economic and national security in the world today.

However, the UK still faces significant challenges in its effort to benefit from this technology, and the economic and security risks remain stark. In the last three months alone we have seen numerous examples of these national vulnerabilities play out. Export-restrictions have been placed on frontier models. The cyber capabilities of these models have increased significantly, improving the ability to defend against cyber-attacks for those who know how to use them, but increasing the risks of security breaches for those who don’t - with a growing number of serious incidents reported. While capable open-weight models provide welcome alternatives to closed systems, almost all of these originate from China-based labs creating another risk of over reliance on a single nation for our AI capabilities. And even efforts to access frontier models from other countries like Japan have been thwarted by basic issues like outdated GDPR requirements.

While the UK has seen continued investment and has strong fundamentals, companies here remain behind peers in the US & China in terms of the capital available to them and broad adoption to date. The capital being invested into this space globally, in particular the amounts being spent on data, chips and datacentres, is set to be the single largest capital expenditure project in modern history. We still have too few strategically valuable companies in critical areas like frontier models, chips and compute.

So, despite great progress, it remains urgent for our national security and economy that the UK continues to support founders to build the models, chips and supporting infrastructure required for the safe adoption of this technology. In short, while we are optimists at Sovereign AI about Britain’s position and potential, we are sober about the importance and urgency of the work we are doing.

Quarterly Report #3

Investments

Since the last update we have publicly announced three further investments.

OLIX is building a new approach to compute that integrates SRAM architecture with photonics, tackling the memory and energy constraints that today’s GPU architectures are running into.

Cusp AI is building AI models to develop new materials, using molecular simulation to design advanced materials starting with energy and sustainability use cases.

Isomorphic Labs is the most advanced AI drug discovery company in the world, led by Sir Demis Hassabis and developing novel therapeutics to tackle disease.

We have done this with 3 fulltime, and very hard working, investment professionals.

It is hard to judge a portfolio of work after just a few months, but between our five announced companies, and the seven in final legals, the portfolio is starting to take shape in line with our original goals. A few observations on it:

First, on our priority sectors. The portfolio maps closely to the focus areas we published at launch: several companies in Compute and Infrastructure, several building new Frontier Model architectures, two in AI for Health and Life Sciences, two in Scientific Discovery, and companies whose work directly supports trust and security in AI deployment.

The stages range from pre-seed through to Series C, which is exactly the early-stage, high-conviction profile we set out to build. The later stage companies have been announced first, this is the nature of building a new pipeline of investments, but going forward we will have far more seed and Series A stage companies, in line with the risk profile we wish to take. Each one, if successful (and we appreciate it is always a big if ) could have a significant impact on the AI economy on a global scale, and we believe can contribute significantly to the UK’s position in it. It is great to see patriotic founders and teams willing to take a bet on a new initiative like SovAI, and we’re grateful to everyone who has applied for funding, even if we can only support a few.

Of the twelve companies, all have their headquarters in the UK, and eleven will have a UK legal top-co ( we made a notable exception for Isomorphic, who are based here and are investing heavily into UK life sciences and AI). Amongst the portfolio we are supporting two companies to ‘flip’ into UK legal jurisdiction after establishing elsewhere. As we have always made clear, we require a heavy UK presence and compliance with UK law, including copyright law which is reviewed as part of our due diligence process and an issue we will seek additional legal guidance on where relevant. As should be obvious, this will often not be a relevant question - asking a chip developer to make specific legal representations about copyright law makes as much sense as asking a fighter-jet developer to make specific representations about fruit & veg law.

I mention the latter points in particular as there have been a couple of inaccurate articles on these topics. We get corrections to such articles where we can but we are a small team so can’t always respond in time. In fact the Economist ran a good piece this week on the asymmetric nature of this issue. It’s been eye-opening to see how often the same small number of people misuse FOI requests, costing taxpayers’ money and civil servants’ time. I hope transparent updates like this reduce the opportunity to abuse a well-intentioned and important tenet of Government transparency.

Our offer to founders

Providing capital is only a small part of what we aspire to do, and the next quarter will be about delivering on our wider offer to founders.

Compute. Demand for our compute programme, run by the AI Research Resource, has been extraordinary: hundreds of applications requesting many multiples of the entire year’s allocation. Alongside time on Isambard-AI, the UK’s most powerful supercomputer, we now provide cloud credits so that companies can access more usable capacity flexibly, in the form that fits their workloads where Isambard hasn’t been a fit.

Rapid Procurement Challenges. In the coming weeks we will launch our first procurement challenges, where we have worked with other Government departments to set out problems they need solved and will invite UK AI companies to compete to solve them, with SovAI as the paying customer. Procurement and advance market commitments like this remain one of the most powerful levers any government has to spur innovation, and we have spent this quarter working on what we think will be the first ever series of procurement challenges focused on British start-ups. It seems common sense that the Government should be able to procure innovative solutions from start-ups, not just established companies, but the work that has had to go into making this possible is extraordinary, and while it will be a drop in the ocean of Government spend initially, I hope it forms the foundation of a much bigger change in Government’s approach to buying frontier-technologies. Follow the SovAI accounts and website. We will share more at launch in the coming weeks.

Nationwide launch events. We followed up on our successful launch in King’s Cross in April with further AI community events in Newcastle, Glasgow, Cardiff and Manchester in the last three months - meeting with hundreds of ambitious founders along the way. We are planning more events like this throughout the rest of the year and will be heading to Belfast in October.

The SovAI job board. We have launched a jobs page on our website surfacing live roles across the portfolio. If you want to work at the frontier of British AI, whether as a laser architect in London ( yes it’s a thing!) or director of quality in Bristol, start there. While there are only c.60 live roles in the UK at the moment, this number and the geographic spread will grow significantly in the coming months as we announce more investments.

Team. We have adapted to the organisational changes in Government in the last few weeks. We have a new Deputy Director fully onboarded alongside our Managing Partner, and we have several new hires joining in the coming weeks across the investment and platform functions, with more announcements to follow. The calibre of people wanting to go into public service and leave exceptional private sector jobs is inspiring. Open roles are always listed on the website.

I expect in the next quarter we will see an even faster increase in model performance, with falling token costs driven by local models. In particular we are monitoring progress in recursive self-improvement, embodied AI (world models & robotics), cyber defence and more efficient compute. We are continuing to refresh these areas of focus to match our goal and we are grateful for the engagement, the challenge and the support on this we have had.

As ever, this update reflects my own views, and if you’d like to get involved, the website is the front door.

Firstly, I want to say a huge thank you for the incredible support we’ve received from the tech industry, researchers, investors, and the Government since starting this role. The energy in the British AI ecosystem, and the progress being made, is incredible.

The last month in particular has been one of significant delivery. We held our formal launch at Wayve’s London HQ. We executed on our first investments. We published the criteria the fund will focus on initially. And we’ve hired critical team members – including our Managing Partner, who we’ll announce in the coming weeks. Behind the scenes, we’ve developed new financial instruments, developed our business case with HMT, set-up our internal investment processes, and put MoUs in place with the most important public finance bodies in Government to ensure deep alignment.

It’s been many long hours and unglamourous work, and I’m incredibly proud of what the team delivered.

I also feel we’ve added to an important part of the national conversation around AI sovereignty. As the Secretary of State shared clearly at RUSI last week:

“If we retreat from progress we retreat from the world, leaving this powerful technology to be exploited by other nations to their advantage and our disadvantage. The choice isn’t between a world that has AI and one that does not. It is a choice between a world where we shape our AI future, based on our own interests and values, or where we are left at its mercy and whim.”

AI Sovereignty, in critical parts of the AI stack, is not a choice - it is something we must achieve. AI will progress with or without us. The only question is whether Britain is at the table. One of the principal ways we will be at that table is by ensuring the UK is home to some of the world’s most important AI companies, which is our goal at SovAI. We are more determined than ever to deliver on that. To be insular in our outlook and modest in our ambitions would rob future generations of opportunity and betray our rich scientific and commercial heritage.

Quarterly Report #2

Progress against Q1 goals

In the last update I set out four goals for the quarter: to set-up an independent legal and operating structure; hire the leadership of the investing function; refine our offer to founders; and launch the fund itself. We delivered on all four.

The fund formally launched, on schedule, on 16 April at Wayve. The Secretary of State opened the event and we had over 150 attendees and national press coverage. Reaching that milestone on the timeline we set ourselves - quicker than many newly formed commercial venture funds achieve - reflects an enormous effort by the team and our partners across Government.

The legal and operating structure is now agreed, with our Investment Committee able to operate with a level of independence and pace I believe is unique within Government, and we have completed our first investments, more on that below.

On the team, we’ve hired our Managing Partner for the investment function, who will be announced in the coming weeks. We’ve also brought in further senior expertise from venture, deep tech and the Civil Service. We will be announcing the lead on that part of our work in a similar time frame.

On the offer to founders, the response from the entrepreneurial community has been far warmer and more engaged than I expected. So far, we have received over 130 applications for compute access – totalling more than 60 million GPU hours of demand – and almost 100 direct applications for equity funding. That signal alone tells me the offer is landing. We are now working to expand our resources, and deliver on the promises. Our KPI to respond within a week is just about holding, we’re still waiting on a critical Outlook integration ( yes....) to automate this process more. You can apply now via our website, but please do check out our focus criteria first!

First investments

We’ve publicly announced our first two equity investments:

Callosum, founded by Cambridge PhDs Danyal Akarca and Jascha Achterberg, builds systems-level software that orchestrates AI workloads across heterogeneous chip architectures – co-optimising models, workloads and silicon in real time. As inference becomes the dominant compute constraint and the hardware landscape fragments beyond traditional GPUs, this orchestration layer is set to become one of the defining platforms of the next generation of AI. Headquartered in London, and accelerated by ARIA, Callosum was Sovereign AI’s first equity investment and a strong validation of both the ARIA and SovAI model.

Ineffable Intelligence, founded by David Silver – Professor at UCL and formerly head of reinforcement learning at Google DeepMind – is building a new class of self-learning systems designed to discover knowledge from their own experience rather than from human-generated data. We co-invested alongside the British Business Bank in a $1.1bn seed round led by Sequoia Capital and Lightspeed Venture Partners, with participation from Index Ventures, Google, Nvidia and others.

These companies are perfect examples of the SovAI fund’s purpose in action: Work with exceptional founders, invest alongside world-leading funds, take risks that if they pay off, could be transformative to Britain’s economic and national well-being. Thank you David, Danyal and Jascha for taking the risk of working with a little known Government entity!

We have six more companies at the final stage of our Investment Committee as of this week, and expect to announce further investments in the coming month.

Investment Criteria

At our launch we laid out our investment criteria, agreed by the Sovereign AI board. This includes our sectors of focus, which will always be public on our website, they are:

01 Compute & Infrastructure

02 Foundation Models

03 AI In Health & Life Sciences

04 Scientific Discovery

05 Trust, Safety & Assurance

Alongside these criteria, as we state online as well, are requirements that a company is anchored in the UK, and complies with British law. These two points are obvious, and we have been explicit about them many times.

I knew that an important effort like SovAI would attract interest, and much of it is welcome when offered in good faith. However, a small minority of very online people have spread misinformation about this, and sent in comments to newspapers (we have had a correction made already, but as they say a lie travels faster than the truth). So to share our principles again:

As it states on our website, in our diligence processes, and as we have shared many times: we only invest in companies that comply with UK law, including copyright law.

Second, we are focused solely on companies that contribute to the UK’s AI sovereignty. Every company we invest in must have a heavy UK nexus: for instance founders and engineers based here. In a vast majority of cases this also means a UK legal ‘top-co’. But there will be a few exceptions. For instance, many brilliant UK founders go through Y-Combinator or Entrepreneur First US and start with a Delaware topco even when they then build their company in Britain. (For what it’s worth, I’d advise against doing that. We lived through this at Balderton with GoCardless, and I think it’s better to start with a UK topco). As the Secretary of State explicitly said in her launch speech, part of our mission is to bring companies like that, which were founded elsewhere back to the UK – to make this a place that you can build and scale here while winning everywhere.

So of course we will occasionally work with companies that have some operations or a legal entity abroad. Britain isn’t an insular nation, we want our best companies to have offices across the globe, SovAI is here to help capture more of the value they create here, by making the UK the most compelling place to build in.

If Sir Demis Hassabis himself were to start a company that happened to have a US topco, but was entirely based in the UK, should we support it? Of course we should. The only people who oppose these nuances are those who oppose the UK having any AI companies at all.

Hiring update

We’re continuing to build out the team. Alongside our Managing Partner announcement in the coming weeks, we are actively recruiting two Investment Partners to lead deals end-to-end, build pipeline, and help shape the funds’ investment strategy. Full details of the role – responsibilities, compensation and how to apply – are on our website at sovereignai.gov.uk/about.

If you, or someone you know, would be a strong fit, I’d encourage you to apply or reach out directly. Other roles within SovAI are listed on the site as well.

Focus for Q3

Four priorities will shape the next quarter.

Continued investment in our priority areas. As mentioned, we have a full pipeline of early stage AI companies that meet our key criteria to execute on. As more investment professionals join in the coming months, we expect to ramp up our investment activity.

Strategic research collaborations. In addition to the equity investments and compute grants we’ve made publicly, there is significant work going on in a number of critical research programmes. Our specific focus at the moment is automated labs and data for life science. I can’t wait to share some of the breakthroughs the team are working on in the next update.

Procurement. We have launched our latest procurement programme and have been fortunate to receive hundreds of survey responses and a large amount of data feeding back on the best way to deliver this. Advance market commitments – where SovAI directly procures solutions from UK AI companies that meet a bar set by other parts of Government – remain one of the most powerful levers we have. Getting this right is a priority for us and we will be running more processess on this, in particular around novel compute, in the coming months.

UK tour. As announced on our various social channels, we are beginning a tour of the UK to connect with more founders and investors. While we have a specific sector focus, AI, rather than a geographic one, the talent, ideas and ambition we want to back are everywhere, and Sovereign AI must be too. I will definitely be at the Manchester event and hopefully others.

Thank you all again for your engagement, feedback, the questions and the patience. As ever, this update reflects my own views, and if you’d like to support please do get in touch via the website.

Best,

James

( P.S the launch videos are now up on our excellent YouTube channel )

You can follow regular updates from the Sovereign AI Unit through DSIT, the Minister for AI and the Secretary of State for DSIT on all social media platforms. But in the interests of further transparency, I plan to release a quarterly report like this with updates on our focus and progress. My hope is this will raise awareness of the great work being done by the Unit, and to make sure you keep us accountable and on mission in the years to come.

Feedback on its format and content is always welcome – but please keep in mind that this represents my own views and our work purely within this Unit - on which I’m ruthlessly focussed. While I will be as transparent as I can be, the nature of the work we do means sensitive commercial or security work will be omitted.

Quarterly Report #1

In her speech at Bloomberg’s Offices last week, the Secretary of State for DSIT succinctly laid out the goal and approach of the Sovereign AI team:

“We have announced our new Sovereign AI Unit to accelerate success, by investing in and supporting UK AI companies…today I can announce the Unit will be a standalone team, highly empowered, operating with maximum autonomy. So it can be fast and decisive.

Because that’s what the AI start up sector demands. Government adapting itself to the speed and pace of business, not the other way around.

And so we say loud and clear to UK start up founders, particularly in AI: You are our future and we will do everything we can to back you.”

This is our promise, and the operating principles, at the heart of the Sovereign AI Unit.

What AI sovereignty means continues to be the focus of global debate. Some argue that Sovereign AI means countries should work together to “pause” AI development. Others argue each country must own each part of the AI infrastructure stack themselves, from electron to token.

In my view, these options are technologically untenable, and economic and cultural dead-ends. Instead, the UK must continue to adopt and contribute to the best infrastructure and tools available to us globally, supporting and shaping the development of AI tools for everyone.

But we can’t be blind to the macroeconomic and political circumstances we’re in. As well as being part of AI’s global development, sovereignty demands that we have both resilience in our own AI infrastructure, and sectors of global excellence in AI anchored in the UK. Resilience and excellence.

So, as the Secretary of State put it above, the goal of the Sovereign AI unit is to accelerate the success of these UK-anchored companies working in areas of AI we believe are of national priority, to benefit the wealth, health, security and values of the UK.

My role as Chair is to help establish the structure of the Unit, hire the team and provide on-going support on the strategy and tactics in relentless pursuit of this outcome.

We have unique tools within Government to achieve this. This includes capital to make direct equity investments as little as £1M to over £20M, which we hope to deploy into dozens of companies in the years to come on commercial terms.

Perhaps more importantly, alongside this direct capital, we can grant access to domestic compute, unique datasets and provide advance market commitments that could uniquely transform and accelerate a company’s success.

What we are not is a policy unit. While we recognise that there are many important areas of policy development required to ensure the UK is a great place to anchor a global AI company - there are many other parts of Government that lead on this work. However, we hope our work does help inform some of those decisions.

We are also not a grant making institution. We are here to invest our resources and capital on a commercial basis. There are many well-funded programmes to support research and R&D in the UK, such as Innovate UK, ARIA and others. Every decision we make is with the view that the taxpayer should be rewarded for taking a risk and supporting AI companies. Not only is this fair for the taxpayer, I feel strongly that better decisions are made when you know performance will be measured by a concrete benchmark, in this case financial returns amongst other metrics.

We will share more on the soft and hard metrics the Unit will work towards in coming months. You can judge our success against these goals and operating principles. Below is a summary of our major efforts in the last 3 months.

Q1 2026 Goals, Progress and Requests.

Structure:

Our goal for Q1 is to establish the legal and operating structure of the Unit and be ready to launch under this new structure by April 2026. Given the unique levers we have at our disposal, and the skills required to deliver them, this is no easy task. However, we have a model in place which will allow for as much independence of our Investment Committee as possible, while embedding them with world-leading experts in the Civil Service, so that we can match the speed founders should demand of us as investing partners while providing unique types of support.

Despite the difficult challenge of balancing autonomy with proper governance, we are well on track to formalise this structure on a quicker timeline than many new commercial VCs can achieve.

Team:

Our goal for Q1 is to establish the leadership of the investing function for the new Unit, and bring in expertise from both the commercial world, as well as experienced civil servants who can help us execute at speed on our offer.

We have already started on this effort with 6 new joiners in under 3 months. This includes Josephine Kant, who brings experience in AI and venture through roles with Google and the tech-incubator Y-Combinator, and Konstantin Sietzy, who was critical in the successful set-up of AISI.

We are now actively hiring for three partner roles to form our investment committee, starting with a Managing Partner who will sit on the investment committee and lead on executive decisions across the Unit as well. We have had over 70 applications, and we have interviewed over a dozen candidates so far.

We would love for interested people to apply directly by emailing a CV or sharing a LinkedIn profile, along with a short introduction as to why you would be relevant for the role to Eliza.Cudmore@dsit.gov.uk. She will share a more details and the process back to relevant parties. Below is a fuller description of the role.

Managing Partner Job Description:

The Managing Partner role in the Sovereign AI Unit is a pivotal leadership position, critical to the success of the Unit’s mission to strengthen the UK’s sovereign capability in frontier AI. The Managing Partner will sit at the centre of the Unit. They will help define the Unit’s investment strategy across the full range of the Unit’s priority areas, and oversee the sourcing, evaluation and execution of the deals that will define the Unit. It requires a rare blend of deep investment expertise, understanding of AI and strategic leadership, with strong credibility and influence across the AI ecosystem.

Key responsibilities:

- Develop and lead a high-conviction investment strategy for the Sovereign AI Unit, shaping strategic priorities and designing a due diligence process in line with market best practice.

- Lead the sourcing, evaluation and execution of deals across the full range of the Unit’s portfolio.

- Act as a central convener between industry, government and investors, representing

- Attract and retain world-class talent to enhance the UK’s position in AI

Who are we looking for?

- Currently working as an investor either in a VC firm, Corporate venture team or equivalent.

- Have a minimum of ~5 years’ experience as an investor. Has lead investment rounds.

- Track record of investments across frontier tech, showing adaptability of investment thesis and a capability to show conviction and win investment deals..

- Exposure to AI industry.

- Built, or managed, teams before.

- Has a strong network and credentials within the UK AI & VC ecosystem and beyond

Offer to founders:

While we set-up the structure and the team to lead this new fund, we are also working hard on refining the offer to founders and fellow investors to make the SovAI proposition the most compelling we can. We want to be a first-choice partner when you think about scaling an AI company in the UK.

As a starting point, SovAI will invest as a leader or follower on commercial terms directly into AI companies within our strategy. We will match the commercial and legal terms of the world’s leading venture firms, working alongside them, and commit to operating with the same pace, passion, risk-appetite and ambition as them as well.

But capital is just one of the many constraints facing founders in critical parts of the AI economy today, and we hope to provide much more than that.

We will also be providing compute directly to companies, via grants of 5k - 500k GPU hours that can come with a cash investment. We have already started making small commitments like this and intend to ramp up the offer significantly over the course of the year, to meet the AI Opportunities plan of 20x’ing the amount of domestic compute available in the UK in the years to come.

And we will be making even more important offers to AI companies in the UK through our advanced market commitments – where the Unit will directly procure solutions that meet our bar for other parts of Government, giving the best AI teams a guarantee they will have a customer if they can build a world-leading product Government needs.

To test these offers with the market we have kicked off a series of interviews with dozens of early-stage AI companies in the UK. If you would like to provide feedback on this model, please don’t hesitate to get in touch with me or Josephine on Josephine.Kant@dsit.gov.uk.

Launch

We are working hard to be fully operational by April 2026, while continuing our work in other areas. We will announce more details about that launch, and ways you can learn more or engage with the Unit in the coming months. So, stay tuned!

Thank you all for your enthusiasm and patience. I hope you found this update useful.

Best,

James

The next 24 months will shape both AI, and Britain’s place in the world, like no period in our lifetime.

The compute used to train frontier AI models has grown four to five times every year for over a decade. The scale of growth has held with remarkable consistency: more compute, more data, better models, more capabilities. And at the same time, the cost of using AI has been falling by an order of magnitude each year.

That combination of capabilities going up and costs coming down so quickly is rare in history. When it happens - in steam, in electricity, in silicon - it does not reward the countries that wait.

I have been spending time this year with researchers at our best labs, and with the experts at our AI Security Institute (AISI). They all agree on this: we are in the steep part of the innovation S-curve. The models being trained now will not be marginally better, they are going to get very effective, very fast.

And as these AI models become more capable, they are expanding from just making the software we use smarter, to making our physical industries smarter as well. Advances in AI are already unlocking new technologies in transportation, robotics, automated labs and advanced materials.

Critically for Britain, this next phase of embodied AI isn’t something that can be built in an office in Kings Cross. It requires real industrial skill, complex machinery and links into intricate supply chains. This phase of AI should see far more of the investment, and ultimately benefits, spread across our country and with a much stronger connection to place, renewing the cities and regions that have led in previous technological waves, if we give them the powers to seize it.

So the question for Britain is not whether AI will change our economy, society and place in the world. It will. The question is whether Britain shapes that change in our interests, giving us the growth, the tools, the national pay rise we need. Or it is done to us without a say.

Be proud of where we stand

Britain is starting from strength. We have seen £13 billion of tech and AI investment in the UK so far this year alone. We have world-leading companies that the rest of the world envies: Wayve making access to transport cheaper and safer, Isomorphic designing more effective medicines, Fractile and Olix making chips that will make inference more affordable. We have the third-largest AI talent pool. We have the third-largest number of domestic datacentres, and in AISI a security institute that the world relies on.

We should be prouder of our place in this race. But pride cannot mean complacency. We have been here before. And we must be honest about what happened last time.

Three hard lessons

My time as Secretary of state has brought into focus three very hard lessons that we need to learn from as a nation.

First, Britain has lived through great technology waves in living memory, such as the internet and mobile. But Brits did not benefit enough. We had a great starting position: world-class scientists, early companies, and real momentum. But we did not do enough to help the companies we had to scale in a way that benefitted us all. They sold early, moved abroad, or were out-muscled by rivals whose domestic markets and governments backed them more. The jobs, the returns, the tax revenues went elsewhere. Discovery happened here, but the returns were banked by others.

Second, we need globally leading companies, producing AI products the world needs, to have a seat at the table. Right now, my department is working hard to make social media safer for young people. Our job is made vastly more difficult by the simple fact that none of the major social media companies were built and scaled here. We are negotiating with firms whose incentives, headquarters and lawmakers are an ocean away. If you want AI to work for British families and to respect our values, our privacy, our children, we need companies at the table who answer to Britain. We cannot let AI be a repeat of the social media story.

Third, it cannot be stressed enough how sovereign AI capabilities matter for our national security. I said in my speech at RUSI that access to advanced AI is as strategically significant as access to energy or ammunition. The assessments coming out of AISI have only strengthened my view. Frontier models are already meaningful in cyber offence and defence, and the gap between those who have them and those who do not is widening. A nation that cannot run, evaluate and secure advanced AI on its own terms has outsourced part of its defence and national security.

These are just some of the costs of not having frontier technology companies and infrastructure in our borders. We are still paying for previous mistakes. We cannot afford make the same mistakes again.

Make it count

So what do we do? Five things are vital, and we have started on all of them.

We invest in compute and hardware. AI runs on infrastructure. My recent announcements on our hardware programme are the start: backing British chip design, securing our supply chains, and building the training and inference capacity our companies and public services need. And if we want British data to be secure here in Britain — NHS records, defence data, the information of ordinary citizens, then we need more UK-based data centres, built faster, connected to the grid sooner, and located in the regions where investment will matter most, outside London and the South East.

We build a trillion-pound company. Not as a vanity project, but because a company that produces such a valuable part of the global AI economy is the ticket to a seat at the global table. That is why we set up the UK’s Sovereign AI Fund, SovAI, to accelerate our successes. It is hard. Of course it is. SovAI is a team taking real risks, and taking risks is something the government has not been good at. Some of its bets will fail. When they do, there will be criticism, and we will continue to back the team - because now is the time to be bold, not dither. Look at what that small team has done in just a few months, supporting some of the most exciting British AI companies like Callosum and Ineffable. Imagine what they will do in two years with more support behind them.

We back British companies. Government procurement should do more for our local champions. We know this model can work, because it built the most valuable technology company Britain has. Back in the 70s, when Britain’s old industrial base was crumbling, the Callaghan government invested in a high risk semiconductor start-up, Inmos, headquartered in Bristol and whose products were manufactured in Newport. That bold move sparked not just a company but an industry, leading to Graphcore and Isambard in Bristol, and to Wales’s amazing semiconductor cluster. And one team of Inmos alumni went to help build what is now Britain’s most valuable company: Arm. Out of that publicly backed bet came the chip designs whose architecture now sits in 99% of the world’s smartphones. One brave procurement decision, four decades ago, helped launch Britain’s most valuable company today. We should have made fifty such decisions since. Now, we have started making them again.

We make AI a priority across the whole of government. Every department will need a plan for how AI makes its services better, faster and cheaper - not glossy pilots, but measurable improvement in the services people actually use. But this can’t be done top down. Look at what is already happening when we let local trusts and clinicians adopt these tools. In East Kent, 85% of NHS clinicians are using an AI tool that spots infections early - catching sepsis sooner and saving lives at a fraction of the usual time and cost. Our clinicians need these superpowers. There are dozens of other examples, and they are multiplying by the day.

We enforce stronger AI sovereignty. SovAI addresses the early stage, and the British Business Bank helps with scale up capital. But we will also use every lever the state already holds, including government procurement that buys British where British is best, so our firms get the revenue and validation they need to scale. This is especially important in defence. And we can strengthen the CMA and other institutions with the tools needed to ensure the AI market stays open, contestable and fair, so the next great British company is grown here, not swallowed early.

Compute, private champions, public adoption, national sovereignty. We need to double our efforts on all of these fronts in the months ahead.

Who is AI for?

But do not mistake my urgency for fear. I am urgent because I have seen what these technologies do when they reach ordinary people. Because the story of the last thirty years is also the story of technology making things easier and cheaper.

Thirty years ago, if you were a child from a low-income family with no connections to investors, starting a business that sold to the world was a dream. Nobody could afford stock, or a store, or a shipping operation. Then came the internet and the smartphone. Suddenly a teenager anywhere in the UK could get their first taste of independence and commerce selling on platforms like Depop or eBay. Millions did.

Thirty years ago, if you were a young actor, musician or broadcaster without connections, the door to TV and radio was closed. Today, YouTube and Spotify have let millions of young Britons create, find an audience, and get a first foot on the ladder - no gatekeeper’s permission required.

And just a few years ago, if you dreamt of building a business, you needed tens of thousands of pounds for software developers, marketers and operations before you could begin. Only the already-wealthy could afford to try. Now, thanks to AI, anyone has the equivalent of a brilliant developer and a seasoned business adviser in their pocket for £20 a month.

These are powers that can transform lives. We can hand them to the next generation - or we can let others decide who gets them.

This is just as true in the public sector. At Citizens Advice, an AI assistant is helping advisers answer people’s questions in half the time, which means twice as many families getting help with debt, housing and benefits from the same overstretched service. Cheaper government thanks to AI is not a pipedream. It is a person getting help today instead of next month.

And it goes beyond the public sector. If we want to reindustrialise across this country - to regain our capability to build - we can only do it by embracing the technologies of the future. Leadership in AI is what lets us compete at the very frontier of new industries in robotics, fusion, quantum, life sciences, and technologies we cannot yet name. But they cannot be dictated to or planned top-down by us in Whitehall. Look at the CHIPS Act in the USA. The federal government set the targets, but it was the individual states who could compete to be the location for these new semiconductor fabs, creating highly skilled jobs in the process. These industries need energy, space, and skills, more independent regions can help make these abundant across the UK. That is how new life comes back to the regions that powered the last industrial age. AI is the enabler for growth everywhere.

No time to lose

Sovereign AI is urgent — because the prize is growth and hope in every part of the country — the two things we have been short of for many years. But the window in which we can still choose our place is measured in months, not decades.

We have the companies. We have the talent. We have the institutions. We have the tools.

Now we need the ambition, dedication and will to make the next 24 months count.