UK AI Companies Up 79% to 10,488 as Revenue Hits £33.6bn
The UK had 10,488 AI companies in 2025, up 79% in a year, and AI-related revenue rose 41% to £33.6 billion, according to the government's latest AI sector study. For marketers, that means more AI suppliers to choose from, a more crowded market for AI firms to sell into, and more vendor risk.
01Telemetry
Key findings at a glance
- 10,488AI companies identified in the UK in 2025, up 79% from 5,862 in the 2024 study
- £33.6bnAI-related revenue in 2025, up 41% on 2024, with dedicated AI firms almost doubling to £9.2bn
- £18.6bnAI-related gross value added in 2025, up 58% on the previous year
- 130,400people employed in AI-related roles, a 51% increase on 2024
- 1,752new AI companies incorporated in 2025, dedicated and diversified firms combined
- 68%of UK AI companies are dedicated AI firms (AI is their main revenue source), up from 56% in 2024
- 84%of dedicated AI companies work on natural language applications, the most common capability
- 53%of dedicated AI companies are tagged with agentic AI capability, ahead of AI development and training at 45%
- 12%of the 3,292 dedicated AI firms in the 2024 study had changed status a year later, most of them dissolved
- 38%of AI businesses that import AI products or services say they are entirely dependent on them
- 91%of AI businesses using compute expect the cost of compute to rise over the coming year
- 25%of AI businesses cite unclear regulation as a significant barrier, up from 16% in 2024
02Scope
What the report covers
The Artificial Intelligence sector study 2025 was published on 29 October 2026 by the Department for Science, Innovation and Technology (DSIT) and the Department for Business, Innovation, Science and Trade (BIST). Perspective Economics led the research with Beauhurst and Ipsos. It is the fourth edition of an annual series that sizes the UK's private-sector AI suppliers: companies that develop and sell AI products, services, platforms or hardware, not businesses that simply use AI. It identified 10,488 active AI companies, up 79% from 5,862 in the 2024 study, and puts AI-related revenue at £33.6 billion (up 41%), gross value added at £18.6 billion (up 58%) and AI-related employment at around 130,400 people (up 51%). Our report on the 2023 edition covered 3,713 firms and £14.2 billion of revenue, so the sector has almost tripled by company count in two years.
The study is careful about what the jump means. Some of the 79% increase reflects better coverage and the growing habit of companies describing their products as AI, but the authors say it is underpinned by genuine market expansion. About 1,750 AI companies were incorporated in 2025 alone, and the dedicated share of the sector rose from 56% to 68%, which the study reads as a wave of small, AI-native entrants. Revenue is still concentrated: large firms make up around 4% of AI companies but generate 82% of AI-related revenue, although the SME share rose from 15% to 18%. Fundraising by dedicated AI companies more than doubled to £6.3 billion across 917 deals in 2025, with rounds of £100 million or more taking around 60% of the value.
Geography and supply chains complete the picture. London, the South East and the East of England hold 72% of registered offices, and 1,372 internationally headquartered AI companies operate in the UK, 54% of them from the US. The survey adds a dependency layer: 47% of AI businesses import AI products or services, most commonly off-the-shelf large language models (73% of importers), and 38% of importers say they are entirely dependent on those imports. Earlier editions are in the AI sector study collection on gov.uk.
03Briefing
What this means for UK marketers
There are two readings here, depending on which side of the table you sit. If you buy AI tools for marketing, the supplier market has never been bigger. If you sell an AI product, you are now one of 10,488 companies, and most of them describe themselves the same way.
For buyers, the churn figure matters more than the growth figure. Of the 3,292 dedicated AI companies in last year's study, 385 (12%) had changed status a year later, and 74% of those were dissolved. 93% of the firms that changed status were micro businesses. Before you connect a small vendor's tool to your CRM, product feed or content workflow, ask where your data is processed, how you would export it, and what happens to your account if the company closes. The ICO's guidance on AI and data protection is the starting point on the data side, although the ICO says it is under review after the Data (Use and Access) Act. Budget for price rises too: 91% of AI firms using compute expect its cost to go up, and interviewees described paying usage-based costs to suppliers while quoting fixed prices to clients. Expect some of that squeeze to show up at renewal.
For sellers, "AI-powered" no longer sets anyone apart. 84% of dedicated AI companies work on natural language applications and 53% are tagged with agentic capability. Interviewees said UK firms compete best through sector focus and domain expertise, because the large model developers are moving up the stack into applications. Your positioning, landing pages and search content should lead with the problem you solve for a named sector, the evidence you have and the deployment experience behind it. That is the brief we work to in digital marketing for UK B2B brands, and it starts with a site that explains the product clearly, as covered in our SaaS and tech start-up web design guide.
Finally, supply is running ahead of demand. ONS data put business AI use at 30% in September 2026 (see our report on AI adoption and jobs), and the study's interviewees described UK buyers as cautious, with long procurement cycles and pilots that stall before deployment. Content that explains implementation, cost and proof will win more of those buyers than another capability claim.
04Method
How they got the data
Check the method before you quote the number.
AI has no SIC code, so the study builds its population from a business-focused taxonomy developed with BIST and an expert advisory panel. Firms were longlisted from Companies House, Beauhurst, Lightcast, UKRI's Gateway to Research, FT fDi Markets, HMRC trade data, Tussell procurement records and web data, then verified between March and June 2026 with automated and manual checks, including language-model tagging of company websites. Only UK-registered, active, commercial firms that supply AI products or services are in scope; companies that only adopt AI, charities and universities are excluded. Revenue, employment and GVA are modelled from filed accounts (Bureau van Dijk FAME and Beauhurst) where available, with conservative per-employee estimates for unknown values, and diversified firms count only their estimated AI-related share. The authors stress these are experimental estimates, not official statistics, and that year-on-year changes mix real market growth with methodological improvement.
The survey findings come from a multi-modal Ipsos survey of AI firms in May 2026 with 376 responses, 87% of them micro businesses with fewer than 10 employees, so they describe the experience of small AI firms rather than the large companies that earn most of the revenue. The research team also completed 50 in-depth interviews, mostly in April and May 2026: 30 with businesses, 10 with academic and policy stakeholders and 10 with investors. Interview findings are presented as context, not as representative measures, and are written up in the annex on qualitative findings.
