Let us start with the news that broke yesterday morning.
Ineffable Intelligence, a London-based frontier AI lab co-founded by David Silver, has closed ~£870 million in seed financing led by Sequoia Capital and Lightspeed Venture Partners at a ~£4.03 billion post-money valuation. This is UK's largest seed financing ever recorded.
Silver left Google DeepMind to chase what he calls endlessly learning superintelligence, using reinforcement learning to build AI agents that improve through experience rather than static datasets. The company has no product. It has no revenue. It has been in existence for five months. It was incorporated last November, and Silver was appointed director in January 2026.
Now, to the state of British AI today.
The money
UK startups raised ~£6.16 billion in venture capital in Q1 2026, a 60% increase on the same period last year and the strongest first quarter since 2022, according to HSBC Innovation Banking UK and Dealroom. AI was the defining theme of the quarter, with UK AI startups raising a record ~£4.58 billion, nearly three quarters of all venture capital raised. This represents a fivefold increase in AI's share of UK VC since 2022.
Five years ago, AI accounted for roughly 13% of UK venture capital. In Q1 2026 it accounted for 74%. That is not incremental growth. That is a structural reorientation of British investment capital around a single technology.
More than 100 AI funding rounds were completed in the quarter. Twelve rounds of £75 million or more contributed ~£4.03 billion, accounting for 65% of total capital raised. The average round size increased by 41% year-on-year.
The largest individual rounds tell the same story in more concrete terms. Nscale raised ~£1.2 billion. Wayve raised ~£950 million. ElevenLabs raised ~£420 million. Synthesia raised approximately ~£160 million. And now Ineffable Intelligence has added ~£870 million on top of all of that.
This is what a funding boom looks like from the inside.
What our directory data shows
The AI Founders UK directory tracks 175 verified founders across 110 companies, all building in Britain. Together they have raised £14.5 billion. These are not 175 randomly selected people. They are the ones our research has determined to be among the most significant names in British AI right now.
Here is what the data shows about who is building, and where.
The geographic story is London and almost nowhere else. 70% of founders in the directory are based in London. Cambridge is second at 12.9%. Oxford is third at 6.5%. Every other city in the UK shares the remaining 10.6%. Edinburgh has four founders in the directory. Manchester has two. Birmingham, the second largest city in the country, has three.
London is currently King’s Cross-maxxing its AI scene. Now home to Google DeepMind, Meta AI, Anthropic, OpenAI, Wayve, Synthesia, and Isomorphic Labs within roughly a square mile of each other, with Project Prometheus also in talks to secure a space in the same neighbourhood. That density is real, and it is self-reinforcing. Talent clusters, capital follows talent, and companies follow both.
But the case against it is worth making too. The companies in Edinburgh, Manchester, and Bristol are, on the whole, closer to university research, operating with lower salary costs, and building for different customer bases. Dexory in Hastings has raised over £190 million and deploys autonomous robots in warehouses across Europe. Lichen AI in Cambridge is quietly transforming how organisations prepare for AI adoption.
The innovation is not only happening in London. The money mostly is.
12.6% of founders in the live directory are female. That is 22 women from 175 founders. The broader UK tech ecosystem average sits around 11%, so British AI is marginally above that bar. But marginally above a low bar is still a low bar. The representation exists at the directory level. The capital is not following it.
This is not a pipeline problem. There are exceptional women building AI companies in the UK right now. Tabitha Goldstaub, Dr Catherine Breslin, and Diana Robinson founded Lichen AI in Cambridge in 2025. Meryem Arik co-founded Doubleword, one of the seven companies selected by the Sovereign AI Unit in its first cohort. The problem is visibility, funding access, and the structural biases in who gets access to which investors. This newsletter will name them, and make sure they are not invisible.
Enterprise AI dominates by a wide margin. 36% of companies in the directory are building for enterprise clients. Biotech AI and AI Infrastructure are both at 12%. The rest is split across Legal AI, Healthcare AI, Voice AI, Robotics, and a thin slice of Consumer AI.
Britain is building AI for businesses. Whether that is a strategic advantage or a missed opportunity depends on what you think the next decade looks like. The consumer AI products that most people actually interact with daily like Claude, ChatGPT, Gemini are American. The enterprise AI tools that most British companies will run on in five years have a reasonable chance of being British. That is a different kind of ambition, but it is a real one.
The average company is 8.1 years old. The founding wave that produced the current generation of British AI companies peaked between 2013 and 2017. Those companies are now approaching Series C and beyond. The next wave is just beginning. Three companies in the directory were founded this year including ineffable Intelligence.
The Ineffable moment
It is worth spending more time on Ineffable because it is not just a funding story. It is a signal about what British AI research is now capable of commanding.
Silver holds a doctorate from the University of Alberta, where he studied under Richard Sutton, widely regarded as the father of reinforcement learning. In a 2025 paper co-authored with Richard Sutton, Silver predicted that experience-based learning would ultimately dwarf the scale of human data used in today's systems. Ineffable Intelligence is the practical test of that thesis.
Rather than scaling transformers on more text, Silver bets on world models and agentic reinforcement learning systems that learn through simulated environments. When this works, it could unlock capabilities that pure language model scaling cannot reach, particularly in robotics, science, and complex planning.
The Ineffable deal compresses the standard funding timeline to near zero. Where most startups spend years proving revenue before commanding unicorn valuations. Silver cleared a £4 billion valuation before shipping a usable product.
That fact reshapes the incentives for every senior researcher currently inside DeepMind, Anthropic, and OpenAI's UK teams. The message the market just sent is clear: if you built something foundational inside a major lab, you can raise a billion British Pounds on your name alone. Watch for more departures from the London AI labs in the next 12 months.
The government picture
On April 16, 2026, the UK government launched the £500 million Sovereign AI Unit at Wayve's King's Cross headquarters. The timing was pointed. The launch happened while the consequences of OpenAI pausing its Stargate UK data centre project were still reverberating through Westminster, with OpenAI citing energy costs and regulatory uncertainty.
The Sovereign AI Unit is a direct bet that the UK can retain the companies it produces. The persistent challenge is that while the UK excels at startup creation, it struggles to retain its most promising companies as they scale. DeepMind was acquired by Google. Featurespace was acquired by Visa. Arm went to SoftBank. Autonomy went to HP. The pattern of British research and British exits benefiting foreign balance sheets is long and well-documented.
The grants programme opened April 16, 2026, with applications due May 7 and June 5, 2026. A separate Strategic Assets Grants Programme offers awards between £1 million and £9 million for building high-value AI datasets and autonomous laboratory infrastructure. The government is also opening £80 million in AI procurement talks with tech firms, with potential contracts worth up to £5 million per project launching as early as July 2026.
Whether £500 million changes the retention problem is genuinely unclear. France committed €109 billion in AI investments. The UK's £500 million amounts to roughly 0.6% of that commitment. But the French figure includes private sector commitments. The comparison is real but not quite as stark as the raw numbers suggest.
What is unambiguously true is that the direction of travel is right. Equity investment, sovereign compute, fast-tracked visas, and direct procurement are exactly the tools that keep founders building here. The question is whether the scale is sufficient for the ambition.
This week in UK AI funding
Ineffable Intelligence has raised ~£870 million seed, with Sequoia and Lightspeed leading the round at a ~£4.03 billion post-money valuation. Europe's largest seed financing ever.
Cloudsmith raised £42.4 million in a growth round led by TCV to accelerate its cloud-native artifact management platform, which handles AI-generated code at enterprise scale. DOJO AI raised £4.44 million from Armilar for its multi-agent marketing platform. Lua raised £4.29 million from Norrsken22 for an operating system for human-agent collaboration.
Locai raised £1 million pre-seed from Fuel Ventures to build off-cloud AI inference infrastructure that runs models directly on users' devices, cutting cloud API costs for SaaS companies. Rivan raised £25 million led by IQ Capital to scale domestic synthetic natural gas production, using AI to optimise conversion of renewables and captured CO2 into pipeline-ready fuel. inploi raised £3 million seed from YFM Equity Partners for its AI-driven talent agents platform, automating recruitment sourcing and screening for high-volume sectors including healthcare and hospitality. QMatter raised £900,000 pre-seed from 55 North for quantum compression tools that use AI to accelerate drug discovery and materials simulation on standard hardware.
The pattern across the smaller rounds is consistent with the broader market. Infrastructure and enterprise tooling are attracting capital at every stage. Flashy demos are not. Investors are putting money into companies that make AI reliable, deployable, and audit-ready. The lab-to-production gap is where British companies are winning.
74%.
That is the share of all UK venture capital that went to AI companies in Q1 2026. Not tech broadly. Not software. AI specifically. Every three pounds invested in a UK startup this quarter, two of them went to an AI company.
Five years ago that number was 13%.
The shift is not reversible and it is not slowing down.
Founder spotlight
This section is an open call. We are looking for UK AI founders willing to answer one question honestly: what did you get badly wrong, and what did you learn from it? Reply to this email. The most honest answers become the first spotlights.