What 2025 taught us about AI in Britain and what comes next

A year ago, every pitch deck I saw led with ‘AI-first.’ By autumn 2025, the same founders were describing themselves as ‘AI-enabled.’ That subtle shift in language tells you almost everything you need to know about what happened to artificial intelligence across Britain in 2025.

The year began with breathless predictions about AI transforming every industry overnight. It ended with clarity about what actually works, who is winning, and where the real opportunities lie.

As someone who spends my days advising startups, investing in early-stage companies, and sitting on boards, here is what I learned and what I think it means for 2026.

Three things 2025 taught us

1. From fearing AI to using it

Over the past year, we’ve stopped fearing AI would take our jobs and started learning how we could use it to do them better.

This is not because concerns about AI were unfounded, but because reality proved more nuanced than the headlines suggested. The ONS reported that around 23% of UK businesses now use some form of AI, up from just 9% in 2023. Yet only about 4% have cut headcount as a direct result.

That does not mean the labour market is untouched. A KCL study found that firms heavily exposed to AI reduced junior roles by several percentage points, even as average pay rose. There may be fewer entry-level positions but there is also a greater demand for people who know how to work alongside AI. The ladder is not disappearing, but it is being redrawn.

2. Britain’s healthcare system begins a quiet digital revolution

I spent six years helping build Pharmacy2U into the UK’s largest online pharmacy. Back then, digital prescriptions were a novelty. In 2025, they became the norm. The number of people able to see their GP records online has jumped from around 1.3 million to over 30 million in just a few years.

This year saw growing momentum around digital platforms, with AI-assisted systems now supporting clinicians with administration, triage, and decision-making. Across the UK and Europe, these AI-driven diagnostic platforms have become prime targets for partnerships and acquisitions, signalling the model is here to stay.

3. AI investment spread beyond London

London still dominates British tech in terms of capital, headquarters, and media attention. But in 2025, the map began to look more interesting.

Funding data from the first half of the year indicates this brilliantly. The UK raised more venture capital than any other European country, with AI startups attracted a record share. Crucially, around a third of those deals were completed outside London, with founders raising rounds in Cambridge, Manchester, and across the devolved nations.

Having worked across Moscow, Paris, Amsterdam, New York, San Francisco, and London, I have seen how geography shapes innovation. The diversification we are seeing reflects how there is a growing drive to tap into different talent pools.

Three predictions for 2026

1. 2026: the year of the “tiny giant”

We will see more billion-pound businesses, like Lovable, that are built by teams of fewer than ten people.

Founders today are going further on £500,000 than earlier generations did on £5 million. When one person with good tools can write code, generate marketing assets, and analyse data, the traditional link between headcount and output starts to weaken.

This will continue to open the door for more diverse and geographically dispersed founders to build companies that can compete with anyone.

2. The NHS will become Britain’s biggest AI user

In 2025, AI in the NHS was still mostly discussed in terms of pilots and experiments. I believe this will change, as AI companies like OpenAI hire figures such as George Osborne to develop ties with government and further AI adoption.

See Also

The startup survival guide: 7 reasons new businesses fail

The combination of an ageing population, chronic workforce shortages, and relentless budget constraints means there is simply too much pressure on the system to ignore tools that can safely increase productivity.

As AI-assisted triage, administrative automation, and clinical decision support moves from pilots to wider deployment, the NHS will be on its way to becoming one of the most significant users of AI in the British public sector.

3. The AI bubble will deflate but it will not burst

Every technology boom must eventually answer the question of if it’s providing real, durable development or just creating another bubble?

AI companies accounted for a record share of UK venture funding in 2025. The bubble question will only get louder as investors start looking towards returns, not just growth. But I do not think we are heading for a crash. A correction, yes. A deflation, probably. A dot-com-style wipeout? No.

Unlike the dot-com bubble, enterprises are not merely experimenting with AI, they are embedding it into mission-critical workflows. When you are using AI to run payroll, diagnose patients, or manage your supply chain, you do not cancel the subscription because markets wobble.

Some valuations have run ahead of reality, and there will be casualties. But the companies that survive will be those solving real problems, using AI as an ingredient, not as the entire recipe.