The UK just spent £7 billion on AI. Here's the bit for SMEs.

London Tech Week wrapped on 12 June with a sequence of announcements that, taken together, set the shape of UK AI policy for the rest of the decade. The numbers landed in clusters: £1.1 billion for the AI Hardware Plan, £200 million for adoption and skills, £150 million for a hardware fund led by former Intel CEO Pat Gelsinger, £50 million for deep tech, and around £6 billion of private commitments from AMD, Nebius, Amazon, Ark, and others. Add the supporting open-source AI moves and the regulatory letters to Ofcom on growth, and the package is coherent enough to read as a strategy rather than a list.

Most of it goes to large firms and academic institutions. A genuinely useful slice is reachable by smaller UK businesses, and worth knowing about before the application windows close.

The headline numbers

Three government announcements bracket the week.

£1.1 billion AI Hardware Plan, announced on 8 June. The largest single piece, £750 million, goes to a national AI supercomputer due for deployment in 2030, plus £400 million for next-generation chips. Of that, £150 million is for inference chips from British startups, to be bought this summer; £250 million is for specialised chips as the technology matures. A further £120 million funds an AI Hardware Innovation Programme for chip design. £45 million goes to skills, on top of an existing £35 million. The British Business Bank puts £150 million into a new fund run by Pat Gelsinger's Playground Global, which is opening its first non-US office in the UK. Named UK hardware firms already in the pipeline include Arm, Fractile, Olix, Oriole Networks, and Callosum.

The strategy is explicit. Technology Secretary Liz Kendall described AI as "the defining currency of economic and hard power" and said control of the hardware "will hold the keys to the future". The target is a 5% slice of the projected $1 trillion global AI chip market by the early 2030s, which would imply roughly £50 billion in annual UK revenue and the high-paid jobs to match.

£200 million for AI adoption and skills, announced the same day. This is the SME-facing money. The headline scheme is Bridge AI, at £100 million, designed to match British companies with AI tools, plus a skills and assurance layer. Tech Town expansion adds £53 million for regional initiatives, building on the Barnsley pilot. The Spärck AI Scholarships expand by £4 million, funding 50 industry placements for university scholars. AI Growth Zones get £5 million each for local upskilling, with the first AI Advisory Growth Lab focused on legal services. The AI Skills Boost programme, which is already running, has passed 1.7 million completed courses. The partners named on the announcement include Cisco, IBM, Deloitte, Google, Anthropic, Microsoft, OpenAI, BT, Rolls Royce, Accenture, and EDF.

Around £6 billion of private investment, summarised at the week's close on 12 June. AMD committed £2 billion over five years. Nebius announced £1.7 billion of UK infrastructure tied to NVIDIA compute. Amazon committed more than £1 billion plus up to 4,000 jobs in Northamptonshire. Ark put £807 million into the Longcross Park campus. Oxford Quantum Circuits raised £260 million, the largest UK quantum funding round to date. Eros Innovation contributed £265 million for sovereign AI capability, with another 3,000+ jobs attached. The geographic spread is London, Northamptonshire, the Midlands, and Edinburgh, which is part of the political point as much as the economic one.

The numbers are real and the strategy is recognisable: subsidise the capability, attract the capital, and lean on procurement to spread adoption. Britain has run versions of this for cloud, telecoms, and life sciences before. The novel piece this time is the depth of commitment to the chip layer, where the UK has previously been on the design side and absent from the manufacturing side.

What's in the package for SMEs

Strip the press-release language away and there are five concrete things a smaller UK business can actually use.

Bridge AI (£100 million). A scheme that, in essence, brokers an AI tool to the firms that want one. The matchmaking is the headline. The assurance layer is what makes it serious: it is meant to address the "is this tool safe and does it actually do what it says" problem that stops most SME adoption. The detail of how Bridge AI is administered is still being published; if you are interested, the route in is via DSIT's gov.uk channels.

Tech Town (£53 million). Regional money for local AI initiatives, building on the Barnsley pilot. The model is a designated area with concentrated funding for skills, infrastructure, and adoption. If your business is in or near one of the named towns, the local council and chamber of commerce are likely to know more than the gov.uk page.

AI Growth Zones (£5 million per zone). The smaller cousin of Tech Town. The Growth Zones are clustered around specific sectors. The first AI Advisory Growth Lab is on legal services, which is interesting timing: it follows the filing we wrote on the Silent Ransom Group attacks on law firms, and tells you the government has noticed that the legal sector is both a target and a candidate for AI productivity gains.

British Business Bank £50 million deep tech fund. Routed through Longwall Venture Partners' Longwall 4 fund. Ticket sizes £500,000 to £2 million per company. About 15 companies. Sectors: advanced manufacturing, clean energy, defence, life sciences. Stage: early. The British Business Bank has now backed 54 funds for roughly £2.7 billion of finance in total; this is one of them. If you are at seed or Series A in any of the four named sectors, Longwall is worth a conversation.

Open-Source AI Builder Fund. £500,000+ of compute resources, 160,000 GPU-hours from the UK's AI Research Resource. Designed to turn open-source AI prototypes into operational tools for public services, libraries, the NHS, and the like. The mentoring scheme attached gives winners of the Hack for Impact hackathon access to the government's in-house AI team. Ten UK developers under 30 sit on a Dev Board that helps shape policy through the year. NVIDIA is a partner. The pitch from AI Minister Kanishka Narayan was direct: "The best AI tools in the world won't be built behind closed doors by a handful of companies. They'll be built by people who ship code, share it, and let others make it better."

What's actually a marketing announcement

Two of the items in the package are easier to be sceptical about.

The £1.1 billion AI Hardware Plan is for a supercomputer that lands in 2030 and a chip industry that will mostly benefit large firms. The chip startup money is real and meaningful, but only if you are already in the chip-design game. For an SME building products on top of AI rather than under it, the hardware plan is background noise.

The AI Skills Boost is a free training course on gov.uk. 1.7 million completed sessions is a good number; it is not by itself evidence that any of those sessions has changed how a UK business uses AI. The skills layer is where most of these schemes succeed or fail. If the Bridge AI matchmaking links a firm to a tool and the staff cannot use it, the firm goes back to ChatGPT on a personal credit card, which we covered in the recent filing on your staff are using AI, you're paying twice.

The regulatory shape underneath

Two letters that landed quietly on 10 June are worth reading alongside the spending announcements. Liz Kendall wrote to the new Ofcom Chair, Sir Ian Cheshire, and to Ofcom itself, on "growth goals". The letters were not headline-grabbing and the published page does not include the body text. The framing matters anyway: the Secretary of State is publicly asking the telecoms and online-safety regulator to prioritise growth. The same letter style has been used with the Competition and Markets Authority earlier in the year. The pattern is the broader one of asking regulators to be more pro-investment, which we covered in the filing on CMA reform and the new regulatory tone and others.

The other piece of the institutional architecture is the UK AI Security Institute. On 26 May the UK and Australia agreed to set up a joint research and information-sharing arrangement, with Australia's new AI Safety Institute as the counterpart. That is the international half of the AISI's work. The AISI's UK function is to test the models that the large labs release; the Australia link extends the same testing capability across the Five Eyes partnership.

What an SME should actually do

Three things will turn this package from a press release into a useful input.

Apply to Bridge AI when the next round opens. The matchmaking is the part of the £200 million scheme that is most directly useful to a smaller firm. The application route will be via DSIT or via one of the named partners. Get on the mailing list now.

If you are seed or Series A in advanced manufacturing, clean energy, defence, or life sciences, talk to Longwall. £500,000 to £2 million is a useful cheque size for that stage, and the British Business Bank backing means the fund's runway is reliable.

If you ship code, look at the Open-Source AI Builder Fund. 160,000 GPU-hours is a real resource. The path in is via the Hack for Impact hackathon and the i.AI mentoring scheme. UK open-source contributors grew 7% year-on-year in Q1 2026 to over 38,000 active people; the new schemes are one of the routes the government is using to keep that growth visible.

The broader lesson, for a board reading this, is that the UK is making a multi-year bet on being an AI-adopting country and an AI-building one. The procurement and assurance machinery is being built to match. Firms that engage early get the benefit of being known to the schemes; firms that wait will be selecting from what is left when the schemes have already placed the bets they care about.

How Steelwise can help

Reading the UK AI funding landscape to identify which schemes are actually relevant to your business, drafting the applications, and connecting the AI policy you write internally to the assurance language the schemes ask for is the kind of practical work we do with clients. Get in touch.