UK reveals the first startups backed by its £500 million sovereign AI fund
Artificial intelligence has become a matter of state in the United Kingdom. And this time, it is not just talk.
The British government just revealed the first batch of companies backed by the UK Sovereign AI Fund, a £500 million fund created to retain talent within the country and strengthen AI infrastructure on British soil. The announcement came from the Department for Science, Innovation and Technology, known as DSIT, and it marks the practical kickoff of a strategy that had been in the works for months.
This move is part of an even larger £2.5 billion program that combines artificial intelligence and quantum computing in a clear government bet that technology is, in fact, a matter of national security. The fund was officially announced by Chancellor Rachel Reeves during her March Mais Lecture and arrived with the promise of transforming the relationship between the British state and the country’s AI startup ecosystem.
The problem the fund is trying to solve is nothing new. British startups have a strong track record in research but stumble when it comes time to scale. The lack of growth-stage capital in Europe pushes many founders across the Atlantic in search of money and infrastructure. The fund arrives as a concrete attempt to change that landscape, offering everything from equity investments to supercomputer access for those building the future of AI.
But who made the cut? And did the support reach the ones who actually need it? 🧵 Let us break it all down.
Why the UK decided to create a sovereign fund for AI
The decision to create a sovereign fund specifically for artificial intelligence did not come out of nowhere. Over the past few years, the UK had been watching a pretty worrying pattern unfold: companies born out of elite British universities like Oxford and Cambridge would develop revolutionary technologies, attract attention from all over the world, and then, when it was time to truly grow, they would end up getting acquired by American tech giants or simply relocate their operations to Silicon Valley. This phenomenon, informally known as brain drain, siphoned off not just talent but also intellectual property that could have generated jobs and tax revenue within the country.
The UK has an incredibly rich history in AI research. It is home to DeepMind, for example, and to companies like Wayve, which develops autonomous vehicles powered by world models for physical AI. The scientific foundation is there. What was missing was the capital and infrastructure for those companies to grow without having to leave the country.
The geopolitical context also weighed heavily in this equation. With the United States and China racing at breakneck speed to dominate global AI infrastructure, it became clear to British policymakers that sitting on the sidelines was not an option. Governments around the world increasingly view artificial intelligence as a core component of national security, and they want sovereign capability within their own borders, including foundation models and data centers. In that sense, the UK Sovereign AI Fund is not merely an economic initiative. It is a political statement that the government understands the structural role AI will play in the coming decades, both in the economy and in national security.
On top of that, the European venture capital ecosystem still faces longstanding difficulties sustaining larger funding rounds. While an American startup can raise a Series B or C without leaving the state of California, a British company often has to accept unfavorable terms or seek out foreign investors who, in many cases, demand relocation as a condition for the investment. The sovereign fund steps in precisely to fill that gap, acting as an anchor of capital that allows startups to grow with autonomy and without having to give up control to outside investors.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
How the fund’s support model works
One of the most interesting things about the UK Sovereign AI Fund is that it does not operate in a single way. The support model combines different mechanisms, which gives it the flexibility to serve companies at various stages of maturity.
Of the seven companies announced in this first batch, only one, Callosum, received a direct equity investment. The exact amount was not disclosed. The other six were granted access to the AI Research Resource supercomputer network, known as AIRR. That might sound modest, but for startups working with large-scale AI models, access to compute is often the single biggest bottleneck when it comes to advancing research and product development.
Martell Hardenberg, a partner at Antler whose portfolio includes Prima Mente, one of the companies that received supercomputer access, described the benefit as transformative. According to him, it directly accelerates the research roadmap for companies whose primary limitation is a lack of computational power.
DSIT also announced that the fund secured a right of first refusal on future investments for some of the selected companies. In other words, when these startups go out to raise new rounds, the government will have the option to participate before any other investor. Beyond that, the department revealed it is in talks with another 30 companies about potential AIRR access, signaling that this first wave is just the beginning of a much broader program.
The selection process for the seven companies followed an open and competitive format, with evaluations based on strategic relevance, technical quality, scalability potential, and concrete computing needs.
The seven startups chosen and what each one does
The first batch of companies backed by the fund covers areas ranging from biological foundation models to AI for national defense, passing through sovereign inference infrastructure, agentic AI, and engineering biology. Here is a closer look at each one.
Callosum
Founded in 2024 and headquartered in London, Callosum was created by two Cambridge neuroscientists, Danyal Akarca and Jascha Achterberg. The company has already raised $10.3 million from investors including Plural Platform, Bare Metal Ventures, Entrepreneurs First, and angel investors.
What Callosum does is pretty ambitious: it is building AI infrastructure that allows different models to work together and run on varied hardware. The ultimate goal is to make artificial intelligence more accessible, efficient, and affordable. And it is the only company in this first batch to receive a direct equity investment from the sovereign fund, which says a lot about how the government views the potential of this technology.
Prima Mente
Founded in 2023 and also based in London, Prima Mente was created by Hannah Madan, Ravi Solanki, and Jonathan Wan. So far, the company has raised $1.4 million from investors including Beyond Capital, Bluebirds Capital, Episode 1 Ventures, Antler, and Octopus Ventures.
The company sits at the intersection of AI and neuroscience. Its platform uses genomic mapping and biological modeling to detect early signs of neurological diseases. The goal is to help medical researchers and healthcare administrators better understand and treat brain conditions.
Doubleword
Also in London, Doubleword was founded in 2021 by Meryem Arik, Jamie Dborin, and Fergus Finn. It is one of the most well-capitalized in the group, with $26.9 million raised from names like Dawn Capital, K5 Global, Atomico, and Octopus Ventures.
The company built a platform to make AI inference, which is basically the process of putting artificial intelligence to work in practice, cheaper. It enables automated model deployment and performance monitoring across different hardware stacks, without being locked into a single cloud provider.
Cosine
Founded in 2022 in London, Cosine is the brainchild of Yang Li, Alistair Pullen, and Sam Stenner. With $6 million raised from Gaingels, Lakestar, and Multimodal Ventures, the Y Combinator-backed company created an AI agent called Genie designed to replace software developers at startups looking to keep their teams lean. Genie fixes bugs, builds features, and refactors code, among other capabilities.
Cursive
The newest company in the group, Cursive was founded in 2025 in London by Olivier Henaff and Talfan Evans. The company is still operating in stealth mode and has backing from SuperSeed. What is known so far is that it is developing generative technology designed to build foundation models and generative infrastructure. Details about how much funding it has raised have not been disclosed.
Odyssey
Here is a curious detail: Odyssey was founded in 2023 by two Brits, Oliver Cameron and Jeff Hawke, but it is headquartered in Menlo Park, California. It is the only company in the group based outside of London. It has already raised $27 million from heavyweight investors including AWS Startups, DCVC, EQT Ventures, Samsung Next, and NVentures, Nvidia’s venture capital arm.
Odyssey is an AI research company developing world models with applications spanning robotics, gaming, education, and defense. Its team includes hires from DeepMind, OpenAI, ByteDance, Tesla, Waymo, Meta, Wayve, and Luma. The stated ambition is to create a ChatGPT-scale moment in its corner of AI.
Twig Bio
Rounding out the list, Twig Bio was founded in 2022 in London by James Allen, Satnam Surae, and Russell Tucker. With $3.79 million raised from investors including Creative Destruction Lab, FoodHack, Zero Carbon Capital, Future Planet Capital, Gaingels, and Seedcamp, the company uses automation and AI to produce sustainable bioproducts like palmitic acid, isoprene, and acetone. Its system combines three components: a biological tool that creates the base ingredients, robots that handle and analyze bacteria, and AI that evaluates results and suggests improvements.
The London concentration and the debate over geographic diversity
One point that has already sparked discussion is the geographic concentration of the selected companies. Of the seven startups announced, six are based in London and one is in California. None represent other regions of the United Kingdom.
Jamie Hardesty, who leads technology ecosystem development at Sunderland Software City in the Northeast of England, pointed out that she knows close to 100 AI-native companies in her region. In her view, there is a shared responsibility to recognize that sovereign capability does not come solely from concentration in the capital.
Another relevant data point: according to PitchBook data, only two of the seven companies have women co-founders. These are numbers that raise legitimate questions about who is actually being reached by this kind of program. If the fund’s goal is truly to build technological sovereignty, making sure that support extends beyond London and that diversity of backgrounds is represented could be just as strategic as choosing the right technologies.
What this move means for the future of AI in the UK
The launch of the UK Sovereign AI Fund and the unveiling of the first group of beneficiaries are, above all, a signal. A signal that the British government is willing to use the weight of the state to compete in a game that, until now, seemed reserved for a handful of countries with much deeper pockets. The £2.5 billion program encompassing both artificial intelligence and quantum computing positions the UK as one of the few European countries with a truly integrated national technology strategy, where different layers of innovation are treated as parts of the same ecosystem.
For startups that did not make this first wave, the message is also important. The fund made it clear that there is a selection process with defined criteria and that new funding rounds are planned. This creates a planning horizon that the British ecosystem badly needed. Knowing that a source of sovereign capital exists for companies working on AI infrastructure changes the calculus for many founders when deciding where to register their company, where to hire, and where to scale. It is a real competitive advantage over other European countries that still lack an equivalent initiative in scale and clarity.
From a global perspective, this kind of initiative can also serve as a reference point. Other European countries facing similar challenges with talent retention and a lack of growth-stage capital may look to the British model as a starting point for their own strategies. The combination of direct investment, access to computational infrastructure, and rights of first refusal on future rounds creates a pretty attractive package for founders weighing their options.
Of course, challenges remain. Sovereign funds have a mixed track record when it comes to agility and the ability to keep up with the fast pace of the technology sector. Bureaucratic processes, overly rigid eligibility criteria, or a lack of governance transparency can turn a good idea into yet another government program that promises more than it delivers. The industry will be watching closely to see whether the UK Sovereign AI Fund can maintain the speed that the artificial intelligence market demands, especially at a time when new models, new architectures, and new applications emerge practically every week. The potential is there. Now it is all about execution. 🚀
