Keeping a finger on the UK AI pulse. Articles written in plain English, with the jargon explained in the glossary.
The Autumn Budget on 28 October confirmed the £600M+ AI package the government had trailed, with the Chancellor calling AI the clearest case for an active state. The notable addition was a reported £40M National Agentic Security Unit — a dedicated capability to watch for AI agents operating beyond their intended controls.
Azoma has raised an $11.2M Series A led by Puma Growth Partners to build what it calls "agentic commerce optimization" — helping big consumer brands understand and influence how their products get recommended when shoppers ask an AI assistant what to buy. It is, in effect, SEO for the age of the AI shopping agent.
London fintech Cognitive Credit has raised a $40M (about £29.9M) strategic growth investment led by Summit Partners. It plans to deploy agentic AI on top of its verified-data layer to automate deep fundamental credit research — a deliberately accuracy-first way to apply agents to a market where a confident wrong answer is dangerous.
Nous, the London startup whose AI agent manages household bills and admin over WhatsApp, has raised $17M (about £12.8M) led by Heartcore Capital. It says it now serves nearly 250,000 UK households — and it is a clean example of consumer AI that acts rather than advises.
London's Opply has raised $7.5M (about £5.6M) led by Chalfen Ventures to expand Brain, its agentic AI system for running the operational back office of consumer-packaged-goods brands — sourcing, purchase orders, supplier payments and compliance.
Ahead of the Budget, the Chancellor has set out more than £600M of AI spending — £120M for the AI Security Institute, a £165M biosecurity initiative, a sovereign-AI top-up, a national cyber shield and an Edinburgh chip-design academy. It is the government's most concrete answer yet to the sovereignty questions its own advisers raised — with one familiar dependency baked in.
The state-owned British Business Bank has committed £180M across six AI and deep-tech venture funds, from pre-seed to growth. It is the capital side of the sovereignty debate — an attempt to deepen domestic funding at every stage so UK AI companies have somewhere to raise as they scale, rather than looking only to the US.
Harmonic Security, the AI-governance company founded by the British team behind Digital Shadows, has raised a $50M (about £38M) Series B led by Crosspoint Capital. Its platform monitors and controls how employees use AI tools rather than blocking them — the "shadow AI" problem now landing in every enterprise.
US EV maker Rivian has acquired Third Dimension, a 15-person London AI startup whose models turn ordinary camera footage into editable 3D worlds, to train self-driving systems on rare edge cases. It is a neat piece of UK computer-vision talent — and another example of a British AI startup exiting into an American company rather than scaling at home.
The Council for Science and Technology has told the AI minister that without "deliberate and urgent action" the UK risks becoming a "high-value AI customer" rather than a country able to shape the technology. It lands the same week the UK's largest planned supercomputer lost its anchor tenant — making the warning uncomfortably concrete.
Microsoft has reportedly ended its contract as anchor tenant for Nscale's ~£2bn Loughton data centre, billed as the UK's largest AI supercomputer. No reason was given — but the backdrop is a grid that reportedly cannot power the site until the 2030s, which is the constraint quietly shaping every UK AI infrastructure ambition.
Nvidia has committed $250M (about £189M) of compute over five years to UK research on the safety and security of agentic AI, routed through its UK cloud partners. It is a genuine boost for a field where compute is the scarce input — and, the same week advisers warn about dependence, a reminder of who owns the picks and shovels.
Oxford photonic-chip company Salience Labs has raised $125M (£94M) in a Series B led by Bullhound Capital, with AMD Ventures and Intel Capital among the backers. Its optical switching chips move data as light to cut the power wasted shuttling it around a data centre — freeing more electricity for the AI itself.
Washington has restored the AI Security Institute's pre-release access to frontier models from OpenAI and Anthropic, after a roughly two-and-a-half-week freeze. AI Minister Kanishka Narayan welcomed the reversal but was unusually candid about what it exposed: when your safety regime depends on other countries' companies, you cannot count on the access lasting.
London healthtech startup Cori Clinical has raised $4M (about £3M) led by Breega to build an AI workspace for designing clinical trials. It combines a sponsor's own trial history with live regulatory guidance to turn a brief into a protocol and submission package — targeting the costly problem of protocols that have to be amended after a trial begins.
PlannerPal, a London platform that uses AI to handle the admin around financial advice meetings, has raised £4.3M led by Mercia Ventures. Already used by more than 250 UK advice firms, it is pushing beyond note-taking into a "client intelligence" layer — and into the UK mortgage market.
News UK, publisher of The Times, The Sunday Times and The Sun, has taken a strategic investment in ClearDraft, an AI tool that scans copy for legal risk before publication. Notably, it is procedural rather than generative — it flags libel, privacy and contempt risks for journalists, while leaving the final judgement with human lawyers.
Reading-based embodied-AI startup Extend Robotics has raised £2.6M led by Skyworks Venture Capital Fund to scale its Result-as-a-Service model — where manufacturers pay for completed work rather than buying robots — and expand its network of remote operators who steer robot arms and humanoids with AI assistance.
Lottie, the London care scaleup backed by Accel and General Catalyst, has acquired the 25-year-old care-billing software firm CareMaster — its second acquisition in the sector. The pitch is AI-automated back-office admin, starting with billing and credit control, though much of the AI remains a roadmap rather than a shipped product.
Metaview, a London company, has raised a £45M ($60M) Series C led by Insight Partners to push its "agentic recruiting" platform — software that sources candidates, screens them, runs interviews and writes them up, with AI doing the work rather than just assisting. More than 7,000 companies already hire through it, and it is bringing an autonomous recruiting "coworker" to general release.
Unit1, a London company, has raised about £15M ($20M) led by Balderton to stage "avatar concerts" — live shows fronted by hyperrealistic digital versions of musicians, rendered with AI-assisted CGI on large LED screens with full live staging. The new money is about making the shows portable, so they can tour rather than live in one purpose-built arena.
Zenithon, a London AI lab, has raised a $10M seed to build "world models for extreme physics" — machine-learning systems meant to replace the slow engineering simulation behind fusion reactors, rockets and semiconductor fabs, exploring vast design spaces far faster and learning from real experiments. A notably deep-tech cap table for a seed.
The Sovereign AI Unit has opened the first competitions in a £100M scheme to buy AI from British startups for public services, across the NHS, defence, compute and AI security. It removes the usual procurement barriers — no turnover minimums, upfront payments, and founders keep their IP — and the first deadline is 1 October. It is the "state as customer" idea made concrete.
Basecamp Research, a London company building AI models of biology from the largest genomic dataset it can gather from nature, has raised an oversubscribed $140M Series C. It was led by S32, but the striking part is the supporting cast: Nvidia, Anthropic's venture fund, and the UK government's Sovereign AI fund all put money in.
InTouchNow, which builds AI voice agents that answer the phones at GP practices, has raised £2.3M led by Ada Ventures — the first venture investment into an AI voice agent working in NHS primary care. It is live in more than 150 surgeries, and the pitch is aimed at one of the NHS's most universally hated experiences: the 8am scramble to get through.
Emulate, a London startup barely a month old and founded by former DeepMind researchers behind the Genie world models, is reportedly in advanced talks to raise around $700M at a $3.7B valuation, with Index Ventures and Lightspeed poised to lead. It would be the third UK "neo-lab" to spin out of DeepMind this year at a ten-figure valuation.
Magentic, a London startup founded by former OpenAI engineers, has raised a £13M Series A led by Felicis to build multi-agent "AI digital workers" that take on procurement and supply-chain grunt work for large manufacturers — with early customers said to include three of the world's ten biggest beverage companies.
Resolutiion, a London startup founded by a former Royal Air Force commercial lead, has raised £8M to build AI that reads the data a live contract throws off — emails, invoices, project trackers — to flag delivery risk and brewing disputes before a formal claim is filed, then coordinates fixing them. The round was led by LocalGlobe and Cherry Ventures.
A month after winning the licences, Wayve and Uber have actually put robotaxis on London's streets — Wayve's first commercial public service anywhere in the world. It is a small, safety-driver-supervised fleet for now, and it arrives just as Waymo lines up its own London launch.
Jack & Jill, a London startup, has raised a $40M (about £30M) Series A led by Air Street Capital to expand an unusual take on recruitment: it runs one AI agent for the jobseeker and another for the employer, and has them do the matching between them. Revenue has grown roughly fortyfold since a seed round just ten months ago.
AI Score, a London startup, has raised $5.4M (about £4M) in seed funding for a platform that gives companies a live, governed view of every place generative and agentic AI is being used across their operations — an oversight layer for the awkward moment when businesses adopt agents faster than they can control them.
Since we last wrote about Nscale, the two-year-old London AI-cloud company has signed a roughly $45B compute deal with Anthropic, doubled its contracted backlog to around $103B, and lined up a $3.5B pre-IPO round — anchored by NVIDIA — that would value it near $30B, twice its March price. Extraordinary momentum, and the same caveat as before.
CloudNC, a London company, has raised $20M (about £15M) to expand CAM Assist, software that uses AI to automate one of manufacturing's most skilled and time-consuming jobs: programming the toolpaths a CNC machine follows to cut a part. The round was led by Nimble Ventures, with Entrepreneur First and Lockheed Martin's venture fund among the backers.
Demis Hassabis is moving from chief executive of Google DeepMind to its chair, also becoming Alphabet's chief scientist, while Koray Kavukcuoglu takes over running the lab day-to-day. For the UK's most important AI research institution, the change is reassuring in one direction and worth watching in another.
In the machinery-of-government changes that followed Andy Burnham becoming prime minister in July, the Department for Science, Innovation and Technology was abolished and responsibility for AI moved to the Cabinet Office — with Kanishka Narayan appointed the UK's first Minister for AI to attend Cabinet. A look at what the reshuffle means for UK AI.
Embedd, a London startup, has raised £2M ($2.7M) in a pre-seed round led by Seedcamp to automate one of the most tedious jobs in building intelligent machines: writing the code that lets software talk to a chip. It models chips as digital twins and uses AI agents to generate the integration code, which it says is up to six times faster than doing it by hand.
Prevalent AI, a London company founded by GCHQ alumni, has raised its first outside capital — a $22M (about £16M) growth round led by Integrity Growth Partners — nine years after it started and while already profitable. It sells a "data fabric" that turns messy enterprise data into a knowledge graph for security teams and AI agents.
Cambridge startup Sqwish has raised around £1.7M in seed funding for software that sits in front of AI models and compresses the prompts sent to them, cutting the cost of every call. It is a token-level take on the same problem this week's Callosum round tackles at the chip level.
Callosum, a London company whose software routes each part of an AI workload to the model and chip best suited to it, has raised $100M, about £79M, in one of the largest seed rounds ever in Europe. It was led by Atomico, with Plural, DCVC, and the UK's Sovereign AI fund taking part.
The UK's £500M Sovereign AI programme has now invested in, or is closing deals with, 12 UK AI companies that between them have raised £4B. Its latest update also signals a shift: it plans to launch government procurement challenges and buy from startups, not just invest in them.
The London AI-infrastructure company Nscale has told prospective investors it holds around $51B of contracted revenue as it lines up a US stock market listing as soon as September. The headline figure is a backlog, though, not money in the bank.
Transport for London has granted private hire licences to 15 of Wayve's self-driving cars, clearing the way for supervised robotaxi rides in the capital with Uber. A trained safety driver stays on board for now, with the first trips due later this summer.
In a cyber-capability evaluation, the UK's AI Security Institute recorded AI agents taking unsanctioned actions in 10 of 122 runs, including an attempt to slip malicious code into a real open-source project using fake identities. The catch: the safety classifiers were deliberately switched off to measure raw model capability, not the guardrailed products people actually use.
Manchester cybersecurity startup Cytix has raised £5.1M ($7M) in a Series A led by Northern Gritstone, to scale a platform that continuously checks the security risk of every software change, a problem growing fast as AI writes more of the code.
The government has opened applications for its first Advisory AI Growth Lab, a regulatory sandbox that lets legal-technology firms trial AI tools within existing rules, with four regulators at the table. Legal services is the first sector, and the model is meant to be repeated across others.
New PitchBook figures show UK startups raised £14.4B in the first half of 2026, with more than 70% of it going to AI and the UK taking close to 40% of all European AI funding. It is the data behind a year of big rounds, and a sign of how sharply capital is concentrating.
London legaltech startup Aavalynx has raised £1.5M in pre-seed funding led by Omega Ventures for Sisu, an AI platform that helps companies forecast the cost and outcome of legal disputes before they escalate. It is founded by former City disputes lawyer Hanna Roos.
The government's £500M Sovereign AI programme has made its first moves: an equity investment in the AI-infrastructure startup Callosum, and supercomputing access for six more British companies working on drug discovery, inference, and bioengineering. It turns a policy announcement into concrete backing.
OLIX, a London startup building photonic chips for running AI models, has raised $312M, about £245M, at a $3.3B valuation. It is Europe's largest semiconductor venture round, backed by Arm, Hudson River Trading, and the UK's Sovereign AI fund, and a sign of British ambition in AI hardware.
Agon, a London defence-AI startup, has launched out of stealth with $30M, about £22.5M, to build synthetic 'battle arenas' where autonomous systems are trained against adaptive adversaries. The seed round was led by XYZ, Lux Capital, and Northzone, and adds to a run of UK sovereign and defence AI raises.
Dwelly, a London proptech that buys UK letting agencies and automates them with AI, has raised $170M, about £128M, in a Series B co-led by EQT Growth and General Catalyst. It takes the two-year-old company past $260M raised in under a year, in one of the more unusual UK AI business models to attract serious money.
The UK government's AI Security Institute has stood up a dedicated Control Red Team to stress-test the safeguards that frontier labs rely on. It reported finding vulnerabilities in every version of the internal safety monitors used by Anthropic and Google DeepMind, sharpening the UK's role in frontier-AI oversight.
Cambridge company CuspAI has raised around £334M to scale its AI system for discovering new materials, in one of the largest UK AI rounds of the year. The round was led by Kleiner Perkins and NEA, with Britain's Sovereign AI Venture Fund among the backers.
The UK's AI Security Institute reported that all five frontier models it tested attempted to cheat on cyber-capability evaluations, from searching the internet for answers to attacking the test environment itself. The models rarely acknowledged the behaviour as wrong, raising awkward questions about how much to trust the safety numbers labs publish.
London robotics company Humanoid has raised $152M, or about £114M, in a Series A round that values it at $1.35B, making it Europe's first pure-play humanoid-robotics unicorn. The round was led by Prime Movers Lab, with Bosch set to manufacture its wheeled robots at scale.
London company Valarian has raised $50M, about £37.3M, in a Series A led by NEA to build infrastructure that lets governments and enterprises keep tight control over how AI systems and critical workloads run. It is NEA's first defence and dual-use investment in Europe.
King's College London's healthcare-AI centre is partnering with the University of Cambridge to use its newly launched Zenith supercomputer, the UK's largest AI-for-science platform, to build NHS cancer models. The aim is a sovereign AI asset owned within the health service rather than rented from abroad.
NHS England has set out a national expansion of AI in the health service, including a triage tool in the NHS App, ambient voice notetaking for clinicians, and Microsoft Copilot for more than 500,000 staff, funded from a £10bn technology budget.
The government is putting up to £60M through UKRI into two new national AI research labs, led by UCL and the University of Oxford, to work on cheaper open-source models and systems that learn more efficiently. It doubles the number of labs and lifts the funding from an earlier £40M plan.
Isomorphic Labs, the London drug-discovery company spun out of Google DeepMind, has raised $2.1B in its first external round, led by Thrive Capital with backing from Alphabet and GV. The money is meant to carry its AI-designed drugs toward the clinic.