The estimate and the method
According to an estimate released on Wednesday by Bloomberg Economics and reported by IT之家, artificial intelligence supported about a quarter of the UK’s economic expansion in 2026. The contribution attributed to AI activities is 0.45 percentage points out of overall growth put at around 1.8 percent.
The figure does not come from a single budget line: the research firm aggregated distinct areas, from the manufacture of computing hardware to the construction of data centers, from the development of models and applications to data storage and cloud services. The number should therefore be read as a composite measure, not as the revenue of a homogeneous sector.
Where the contribution comes from
A central element of the report concerns the nature of this contribution. The boost to growth comes mainly from production and research related to the technology itself, rather than from its widespread adoption in businesses and services. It is a distinction that changes the meaning of the estimate: the weight of AI largely reflects the activity of those who build and develop the tools, not yet the productivity gains of those who use them.
Economist Ana Andrade, cited in the material, says that the technology is injecting dynamism into the UK economy. The same economist expects that, with broader use, the benefits will progressively translate into productivity increases across sectors. This is a stated expectation, not an already measured result.
The risk of a repricing
The report links the estimate to the recent wave of selling in AI-related stocks. According to the authors, the impact on the UK would depend on the shape the correction takes: a scenario of disorderly decline, compared to a financial crisis, or an orderly realignment of valuations. The material does not quantify either hypothesis.
A second reference comes from the Bank of England, which had previously estimated the contribution of AI adoption to this year’s growth at about one fifth. The two figures do not coincide and are not directly comparable in light of the available data: they come from different institutions, and it is not stated whether they use the same definition of AI-related activity.
It therefore remains to be seen how the market correction will evolve and whether the contribution will remain concentrated in the production chain or shift toward widespread productivity gains, which the report identifies as the next expected but not yet observed phase.
