Founders and investors are cautiously optimistic of the UK Government’s AI ambitions, they tell Tech Funding News on the launch of the £500m Sovereign AI Fund.
For founders looking to call – and keep – Britain their home, there has always been a question as to whether its ambition matches reality.
The UK is known as a great place to start a company, but not necessarily to scale. For example, the jewel in the UK’s AI crown, DeepMind, was quickly gobbled up by Google. The British government wants to make sure that doesn’t happen again, and has set its sights on establishing the country as “an AI maker, not just an AI taker.”
The Sovereign AI Fund was officially launched Thursday evening at the London headquarters of autonomous car company Wayve.
Alex Kendall, the CEO of Wayve, said a vehicle like this would have made a “meaningful impact” on his company in its early days.
“We started with a bold, contrarian idea and a small team in Cambridge, and while the UK offered world-class talent and a strong testing environment, scaling a frontier AI company requires long-term backing and conviction. Initiatives like this help give ambitious British companies the confidence to start, stay, and scale here in the UK,” he tells TFN.
The launch reflected that ambition, he says. “There was a clear sense that this is a moment of intent for the UK,” he adds, and “a shared understanding that the UK cannot afford to sit on the sidelines.”
This pragmatism was matched with “a huge amount of optimism and excitement in the room, and ultimately the belief that the UK has what it takes to lead in AI,” he says.
For some of the companies in the room, the energy was more than symbolic. James White, chief revenue officer at Cosine which is one of the seven companies in the fund’s first cohort, announced on LinkedIn as he headed into the event that the company had been granted half of a million GPU hours on Isambard-AI, infrastructure support that is worth “millions of pounds.”
“Our ambition is simple: to be the DeepMind that stays British,” White said.
Ideas versus execution
For Justin Lane, co-founder of CulturePulse AI, a dual-use AI startup currently registered in the US, the UK’s ambitions are making him rethink his long-term strategy. “We’ve always had a root in the UK, so to speak,” Lane, a born-and-bred American who came to the UK to study, tells TFN.
Now, he’s considering making it official and redomiciling. It marks a step-change in the status quo, where startups normally head to the US to both register and scale.
Lane says that the UK has “always punched above its weight” in deep tech innovation relative to its population size, serving as a useful middle ground between the research and development strength of the EU and the capital markets of the US.
But he is also candid about where the friction lies. For early-stage companies in deep tech and defence sectors defined by long R&D cycles and notoriously slow sales pipelines, the rising cost of living and doing business in the UK adds pressure at precisely the moment when runway matters most.
“It’s a great idea,” he says of the new Sovereign AI Fund. “But great ideas are only as good as their execution, and in the AI space today, speed is everything.”
It is a view shared across the ecosystem, though many are holding tight to see how. Founders and investors who have been watching this week’s developments remain cautiously optimistic as they wait to see what the fund is deployed and what non-monetary support looks like in practice.
What’s on offer
Since 2020, the UK has had more AI companies founded per capita than anywhere else in Europe. It has world-class research universities, a deepening pool of technical talent, and now, a £500 million government-backed Sovereign AI Fund with its first cohort already named.
The fund, backed by the Department for Science, Innovation and Technology, is the UK’s latest attempt to keep its best AI companies from scaling elsewhere. It was announced as part of a wider £2.5 billion AI and quantum programme confirmed by Chancellor Rachel Reeves earlier this year.
Beyond capital, companies can gain access to the UK’s national supercomputing infrastructure, specifically the Isambard-AI supercomputer in Bristol and Dawn in Cambridge, as well as a procurement pathway into government contracts.
This includes an Advance Market Commitment programme worth up to £100 million, under which the government commits to becoming an early customer for eligible UK AI hardware companies.
Companies that bag a cheque from the fund also gain fast-tracked visa decisions – within one working day – and access to an initial 10 cost-free visas for top R&D talent.
The fund, which also provides hands-on support to its companies, has now released its first batch. As TFN reported yesterday, seven startups have already been named across biological foundation models, physical AI, sovereign inference infrastructure and agentic AI.
Of the seven, one – Callosum – received an undisclosed equity investment, while the remaining six were allocated access to the AI Research Resource supercomputer network.
Overcoming a key AI bottleneck
For some portfolio companies, the compute allocation is already making a tangible difference. Martell Hardenberg, a partner at Antler whose portfolio includes one of the seven companies named in the first cohort, Prima Mente, says the GPU hours allocated through the AIRR network remove “a genuine bottleneck at exactly the right moment.”
Hardenberg says Prima Mente’s biological foundation model targets 70 trillion tokens across five modalities simultaneously. “That kind of training requires serious compute infrastructure,” he tells TFN. “This directly accelerates their research roadmap.”
“The most valuable thing the fund can give deep tech startups is not a cheque but access to resources that are genuinely difficult to acquire privately,” he adds.
For founders like Lane, whose work sits at the intersection of AI and national security, the procurement element stands out as one of the fund’s most positive signals.
“It shows that thought was given to the practical issues of working with government, and that goes beyond just capital,” he explains. “It’s also about the structure of B2G sales itself – it’s a serious green light, and one we also see reflected in the NATO Innovation Fund.”
Lane sees both initiatives as meaningful progress and believes that if the procurement access delivers in practice, it could become a decisive factor for companies like CulturePulse when choosing where to base themselves.
Translating support into deployment
For the fund to fulfil its ambition, compute access alone will not be enough. The real question is whether that access will translate into real deployment wins, and whether those wins will reach founders beyond the capital.
Nicola Sinclair, general partner at Twin Track Ventures and a former Royal Air Force officer who spent nearly two decades in defence, also points to the operational details.
The size of the fund, although previously scrutinized for being modest in comparison to what the US and individual tech companies are spending on AI infrastructure, is not what she’s focused on.
“I’m paying much less attention to the fact that there is a fund, and much more attention to how they look like they’re going to operate as a way that helps to unify this customer base,” Sinclair tells TFN.
Sinclair sees the fund’s structure addressing a problem she has encountered in governments worldwide. There is often a disconnect, she says, between the people who understand what a technology can do for a government customer and those who can assess whether the company behind it is actually viable.
“We don’t often get those two things at the same time,” she explains. A fund that can apply both filters simultaneously is, in her view, one of the most underappreciated aspects of the announcement.
What comes next
The fund’s first cohort has already surfaced another pressing question, and that is who the fund is really designed to reach in the first place. Of the seven companies named, six are based in London, and one is in California, which Jamie Hardesty, head of tech ecosystem development at Sunderland Software City, says is concerning.
“The Sovereign AI Fund is positioned as a national intervention. But the initial cohort being entirely London-based raises a fair question about who can actually access that capital,” Hardesty says. “The issue isn’t a lack of ambition or capability. It’s distance to capital, to networks and to decision-making.”
For Hardenberg, there may be a longer-term opportunity to look for within the NHS. If the fund can help broker responsible data access between AI startups and the NHS, the UK would have a structural advantage in biological AI that few other countries could match.
As Lane put it, however, great ideas are only as good as their execution. “If this fund helps supercharge the AI industry and helps overcome the hurdles that exist around startups, then the moment is going to set a really interesting precedent – not just for the UK, but globally,” he says.
“But if the moment turns out to be more talk and PR than action, it will just go down with a long list of other well-intentioned ideas.”
