Key findings
- The first estimate for Q2 2026 showed UK GDP growth of 0.4%, with information and communications contributing almost half of the expansion according to analysis of ONS data.
- Computer programming, consultancy and related activities rose 3.7% quarter on quarter, while hardware investment also strengthened — consistent with an AI infrastructure build-out, though not proof that AI caused all of the growth.
- The policy challenge is diffusion: technology-sector growth does not automatically translate into productivity gains for ordinary SMEs, regions or workers.
- Sanctuary's preferred lens is capability diffusion — adoption, workforce judgement, local ecosystems and access to capital — rather than headline AI investment alone.
A clearer economic signal
The UK's first GDP estimate for April to June 2026 reported 0.4% quarterly growth. Reuters' analysis of the ONS release noted that information and communications contributed almost half of the expansion and that computer programming, consultancy and related activities rose 3.7% on the quarter after strong growth in the previous quarter (Reuters, 2026) ↗. The ONS release itself is the authoritative statistical record for the quarter (ONS, 2026) ↗.
It is reasonable to describe this as evidence consistent with an AI investment cycle. It would be too strong to treat every pound of sector growth as caused by AI. Programming and consultancy contain many activities, and macroeconomic output has multiple drivers.
Hardware matters because AI is physical too
The same reporting highlighted stronger investment in computer hardware and ICT equipment. That is a useful corrective to the idea that AI is only software. Compute capacity, data infrastructure, energy, connectivity and specialist labour are all part of the productive base.
For decision-makers, this means an AI strategy cannot stop at buying licences to generative tools. It has to consider workflow redesign, data readiness, staff capability, cyber controls, procurement and whether the organisation can turn a technical tool into a repeatable economic process.
Growth can concentrate
Emerging UK research points to the geographic concentration of AI firms and the importance of local skills and economic ecosystems. Ashraf, Coyle and Debnath's 2026 preprint finds strong concentration of UK AI entities in London and argues for place-sensitive interventions; as a preprint, it is useful emerging evidence rather than a settled conclusion (Ashraf et al., 2026) ↗.
The strategic risk is that national AI success coexists with weak diffusion: a cluster of highly productive firms grows while smaller firms elsewhere remain low-adoption users or consumers of imported technology.
The Sanctuary perspective: measure diffusion, not announcements
A more useful scorecard would ask how many SMEs have moved from experimentation to integrated use; whether AI adoption improves cycle time, quality or customer outcomes; whether managers can govern it responsibly; and whether regional firms are building capabilities rather than only purchasing tools. Business support should help firms move from curiosity to validated use cases.
This turns AI policy into a management question. The goal is not maximum adoption. The goal is productive adoption where the benefit exceeds the cost and risk.
Workforce capability is the transmission mechanism
Government-backed Skills for AI research finds that AI use is rising faster than workforce capability and that effective upskilling is practical, contextualised and linked to real tasks rather than generic awareness (DWP and Skills England, 2026) ↗. That provides a plausible mechanism for diffusion: firms that build judgement and workflow capability are better placed to convert technology into productivity.
The same logic supports learning and capability development as an economic-development tool rather than a separate HR activity.
From AI boom to broad-based value
The next stage of the UK AI debate should be less about whether the country is 'winning AI' and more about how value travels. Can a construction SME reduce estimating time without degrading quality? Can a local authority improve service triage while protecting rights? Can a high-street business use forecasting or customer insight without becoming dependent on opaque systems?
Those are smaller questions than national compute investment, but collectively they determine whether an AI boom becomes durable productivity and shared prosperity.
Research foundation
References
- Reuters (2026). AI boom is starting to show in UK economy's performance. Reuters.Source ↗
- Office for National Statistics (2026). GDP first quarterly estimate, UK: April to June 2026. ONS.Source ↗
- Department for Work and Pensions and Skills England (2026). Skills for AI: What works for AI upskilling in the UK. GOV.UK.Source ↗
- Waqar Muhammad Ashraf, Diane Coyle and Ramit Debnath (2026). Code, Capital, and Clusters: Understanding Firm Performance in the UK AI Economy. arXiv preprint.Source ↗
- Department for Science, Innovation & Technology. Hero image: administrator-supplied photograph. CC BY 2.0.Image source ↗
Related Sanctuary capabilities
From analysis to implementation.
Discussion
No approved comments yet.