Subversive Ventures
← Uncommon Sense
Field Notes

The Prize and the Pitfalls: What AI Could Really Deliver for UK SMEs by 2030

24 Sept 2025•David Ashenden
The Prize and the Pitfalls: What AI Could Really Deliver for UK SMEs by 2030 — Photo by Possessed Photography on Unsplash

Photo by Possessed Photography on Unsplash

I said previously ambition is cheap and execution is the real test. The same applies here. Ministers talk about "AI superpower" status. Consultants draw charts that slope endlessly upward. But the question for small firms is much more basic.

Can they afford it, can they power it, and can they find the people to make it work.

The adoption rate is still low. Nine per cent of firms were using AI in 2023, with forecasts of twenty-two per cent by 2024. By 2025, surveys show adoption has picked up - around a third of SMEs now report using AI, according to YouGov, while the British Chambers of Commerce puts the figure closer to 35 per cent. A separate survey by SumUp suggests more than half of SMEs are either already using AI or plan to within the next year**.** A chatbot or a bit of predictive text is not real adoption, though. Embedding systems into processes is the hard part, and that is where most firms still stumble.

The prize is often framed in billions. Government papers talk about an extra four hundred billion pounds of growth by 2030. McKinsey sketches scenarios where GDP rises by a fifth. PwC is more cautious, suggesting three to four per cent gains over the mid-2030s. Even the low estimates are worth chasing. But the outcome rests on questions that rarely make the headlines.

Ambition is the headline. Execution is the compounding.

Feeding the beast: power and data

AI eats electricity. Data centres are multiplying across the UK, pulling on already stretched grids. The government has proposed AI growth zones to prioritise planning and power connections. Without this, SMEs risk being priced out or simply cut off when cloud costs surge. The irony of firms trying to digitise only to be throttled by the national grid is not lost on anyone.

National Grid has warned that electricity demand could increase by up to 50 per cent by the mid-2030s, partly due to data centre growth. Much of that demand will cluster around London, Manchester, and the Thames Valley where hyperscale facilities are already piling up. SMEs, who rely on the same energy mix, will feel the pinch through higher costs long before they see the benefits.

AI is only as strong as the current in the socket.

Data is the other side of the coin. Without reliable, accessible, and affordable data infrastructure, small firms will never get beyond pilots. The government promises a National Data Library and sector data trusts, but many SMEs are still drowning in spreadsheets, PDFs and paper invoices. If the pipeline into AI systems is dirty or incomplete, the results will be useless.

Garbage in, garbage out is not a cliché. It is day one of SME AI.

Building skills where they count

Britain is rich in elite AI researchers but poor in middle-tier skills. SMEs do not need another academic paper. They need staff who can manage data, tweak models, and integrate tools into daily work. Colleges, apprenticeships, and practical bootcamps will decide whether adoption spreads or stalls.

A bakery that can forecast demand better, a logistics firm that can optimise routes, a solicitor’s clerk who can safely supervise drafting tools — these are the people who will determine whether AI makes a difference. The elite may win the research prizes, but the middle tier will decide if adoption ever reaches scale.

AI literacy should become as normal as spreadsheet literacy.

Making adoption affordable

Software licences are not the real cost. Integration, change management, and the hours lost while systems bed in are the killers. For SMEs running on thin margins, a poorly executed pilot can feel like a tax. That is why modest vouchers or tax breaks matter more than grand prizes. They cover the painful middle ground where enthusiasm usually dies.

Large corporates can write off failed pilots. A small business cannot. Without financial cushioning, adoption will stay patchy. Singapore’s experience is telling: SMEs there adopted digital tools rapidly because government covered up to half the costs. Britain does not need to copy the subsidy, but it must reduce the risk that the journey from pilot to production becomes a cliff.

Liability remains murky. If an AI system makes a mistake in a clinical or financial setting, who carries the can. Cyber risk is rising as firms open more data pipelines. Public trust is brittle, and a few high-profile failures could sour attitudes quickly. SMEs do not want 50-page frameworks. They want one-page rules for their sector, written in plain English.

The real cost of failure is not a bad model. It is a lost year of confidence.

Looking outward

The international comparisons remain useful. The United States has the tools everywhere but adoption remains patchy. Germany relies on regional trainers who show up in person, building trust slowly. Singapore keeps it brutally simple with pre-approved solutions and co-funding. Britain will need to borrow the bits that fit its size and complexity, and accept that it cannot copy them wholesale.

By 2030 a realistic outcome is not a revolution.

It is a normalisation of SMEs using AI tooling daily. Power connections that can actually handle the load. Training programmes that build a middle class of AI-literate workers. Affordable pathways that let firms scale without betting the business. That will not grab headlines, but it will raise productivity and take the edge off the UK’s flatlining economy.

The measure of success is not how many supercomputers we build or how many reports we publish. It is whether a small logistics firm in Birmingham runs more efficiently, or a regional law practice serves clients faster. Do that across hundreds of thousands of firms and the economy moves. Fail to do it, and we will be left with slogans and stalled pilots.

The test is not whether Britain wins the AI race. It is whether the average SME in 2030 feels more capable, more productive and less stressed than it does today.

References:

Chart on Aggregated Electricity Demand - https://www.neso.energy/publications/future-energy-scenarios-fes

UK SME AI Adoption Trends and Scenarios to 2030 (illustrative) Sources:

← More Uncommon SenseGet new pieces by email