Government · UK · Department for Science, Innovation and Technology (DSIT)
DSIT: 3,713 UK AI firms hit £14.2bn revenue in 2023
DSIT's Artificial Intelligence Sector Study 2023 puts the UK AI industry at 3,713 active companies, £14.2 billion in AI-related revenues and 64,539 AI-related jobs. The structure of the sector matters as much as the headline numbers: this is a concentrated market with a small number of large players and a deep base of dedicated SMEs.
Source published Summarised
Read the full report from Department for Science, Innovation and Technology (DSIT)
Key findings at a glance
- 3,713 active UK AI companies identified in 2023, up 17% on 2022 (+543 firms). Net growth is driven almost entirely by micro enterprises entering the sector.
- £14.2bn in AI-related revenues across the UK sector in 2023, up 34% on 2022. Diversified firms drive most of the growth, +£4.3bn versus 2022.
- 64,539 full-time equivalent staff employed in AI-related roles. That is 14,499 more than 2022, a 29% year-on-year increase.
- 80% of all UK AI revenue (£11.4 billion) is generated by large firms that represent just 4% of identified companies. Concentration is structural.
- 73% of AI company office addresses sit in London, the South East or the East of England. Outside that triangle, AI activity is vertical-specific (automotive, energy, manufacturing, agri-food).
- £1.2bn in equity investment into UK AI companies in 2023, down 53% from £2.4bn in 2022. Deal volume only fell 1%; the average cheque shrank from £4.6m to £2.3m.
- 70% of 2023 UK AI equity investment (£822 million) landed in London-headquartered companies. Regional capital flow is more skewed than the headcount distribution.
- £1.5bn raised between January and July 2024 across 150 deals, already exceeding the full 2023 total. The cooling of 2023 looks transitional, not structural.
- 74% of the 297 surveyed AI businesses cite developing AI products as their primary future growth driver. Service-led models are a secondary play.
- 53% of survey respondents flagged access to equity as a major barrier to growth, the most-cited constraint in the study. Capital is the binding factor.
- 59% of UK AI companies are dedicated (AI is the primary revenue source); 41% are diversified. Dedicated firms are smaller on average: 14 staff, £2m revenue.
- 81% of dedicated-segment Gross Value Added comes from the top 10 firms alone. The dedicated tier is far more concentrated than the diversified one.
What the report covers
DSIT's Artificial Intelligence Sector Study is the UK government's annual baseline measurement of the AI industry, delivered by Perspective Economics with Ipsos, Beauhurst and Glass.ai. The 2023 edition was published in October 2024, and the next iteration is in fieldwork now (late April to end of May 2026). The study sits inside DSIT's wider evidence base for AI policy and is the closest thing the UK has to an official sector scoreboard.
The methodology blends desk research, automated web-data classification, official company-registry data (Companies House, Bureau van Dijk FAME), private investment data (Beauhurst), and primary qualitative research. The 2023 study includes 297 survey responses from AI-focused businesses and 45 in-depth stakeholder interviews. A taxonomy of 14 capabilities (Model Development, Computer Vision, Machine Learning across financial services, healthcare, retail and energy, NLP, Autonomous Systems, AI Assurance, and others) is used to classify each company.
The 2023 dataset identifies 3,713 active UK AI companies, of which 59% are dedicated AI businesses and 41% are diversified (AI activity sits inside a broader commercial offer). Aggregate AI-related revenue reached £14.2 billion, employment hit 64,539 full-time equivalents, and equity investment was £1.2 billion. The headline story is growth, but the more useful story sits underneath: structural concentration at the top, micro-enterprise churn at the bottom, and a regional skew that has barely moved.
What this means for UK marketers
Three operational takeaways for UK B2B marketers and ecommerce operators selling into, or alongside, the AI sector.
The buyer is large and concentrated; the addressable mid-market is shrinking. 4% of UK AI firms generate 80% of the revenue. The top 10 dedicated AI companies hold 81% of the segment's GVA. If a UK B2B marketing programme is targeting "the UK AI sector" as a homogenous prospect base, it is wasting impressions on the long tail. Account-based marketing into a defined enterprise list (the top 100 dedicated AI firms by revenue, plus the top 100 diversified AI employers) is structurally more efficient than broad-reach campaigns. The small and medium tier between micro startups and large enterprises is also shrinking as a share of employment, meaning the SMB-marketing playbook that works in most UK B2B verticals does not generalise here.
Geography matters more than it looks. 73% of registered offices sit in London, the South East or the East of England, and 70% of 2023 equity investment went to London-headquartered firms (£822 million of the £1.2 billion total). For UK paid-search and Meta-ads campaigns targeting AI-sector buyers, the most efficient geo-strategy is to weight bids heavily toward the M25 plus Reading, Cambridge and Oxford, and to treat the rest of the UK as a long-tail layer. The exceptions are vertical-specific: automotive AI in the Midlands, energy AI in Scotland and the North East, manufacturing AI in the North West. Vertical creative for those regions outperforms generic AI messaging.
Equity access is the headline pain point, not technology. 53% of AI businesses cite access to equity as a major barrier, more than skills, regulation or compute. Marketing teams at fintechs, advisory firms, legal services, accountancy practices and recruitment specialists serving the AI sector should lead with capital-cycle messaging. SEIS/EIS-ready propositions, growth-loan brokerage, due-diligence services, fundraise PR and US-investor introductions are higher-intent buys than generic "AI growth support" copy. Aligning to that pain inside SEO content programmes and Google Ads campaigns will outperform technology-led narratives in the next 12 to 24 months.
How they got the data
The 2023 study combines four data layers: (1) AI-firm identification using Glass.ai web crawling, Bureau van Dijk FAME, Beauhurst, Lightcast, UKRI Gateway to Research and Financial Times FDI Markets; (2) revenue, employment and GVA estimation using Bureau van Dijk filings and Glass.ai web data; (3) investment data from Beauhurst's high-growth dataset (announced and unannounced rounds); and (4) primary research, an online survey with 297 respondents plus 45 in-depth stakeholder interviews. The taxonomy was updated for 2023 to include Model Development as a standalone capability and to split the previous Strategy/Consulting/Training capability into two distinct categories.
Because there is no Standard Industrial Classification code for AI, the sector population is constructed using experimental methods. DSIT notes that headline 2022-to-2023 estimates remain comparable, but lower-level capability analyses are not directly comparable to the 2022 study because the taxonomy was updated. The full findings are available on gov.uk as the Artificial Intelligence sector study 2023, a 54-page PDF with a parallel HTML summary.
Government · UK
Read the full report
From Department for Science, Innovation and Technology (DSIT) · published 01/05/2026