The UK Office for National Statistics put real gross domestic product 0.4% higher in July from June, after 0.3% growth in June and zero growth in May. Services rose 0.4%, production 0.2% and construction 0.1%, making the services line the main accounting basis under the monthly print. The same bulletin put GDP 1.6% above July 2025 and three-month output through July 0.4% higher. [E1]
Measured against the forecast ledger, the surprise is not just a change in mood. Reuters reported that its poll had expected no monthly growth, 1.2% year-on-year growth and 0.3% growth over the latest three months, below the reported 0.4%, 1.6% and 0.4% readings. [E4] ONS listed the monthly GDP estimate for release at 07:00 UK time, so the print was new hard data rather than a recycled forecast. [E2]
A useful technology detail sits inside services, where computer programming, consultancy and related activities expanded 3.5% in July and contributed 0.12 percentage points to real GDP growth. ONS said many of the businesses reporting the largest turnover in that category were involved in artificial intelligence and cloud computing. [E1] That wording supports an AI-linked classification, not a causal calculation that AI itself added 0.12 percentage points to national output.
That distinction matters because the release gives a contribution from an industry category, while the causal claim would require firm-level detail that the supplied record does not establish. The government release mirror points to the same July monthly GDP estimate, which is useful as a public posting but does not add a separate decomposition of the AI or cloud component. [E3] A cleaner causal reading would need ONS respondent or industry detail showing how much of the turnover gain came from AI work rather than from the broader programming and consultancy category.
The macro read is therefore stronger real activity at the start of the third quarter, with a services-heavy composition and a technology-heavy sub-ledger. Reuters supplied the outside forecast comparison: economists had expected flat monthly GDP and slower annual and three-month growth than ONS reported. [E4] That makes a simple weak-growth story harder to sustain, but it does not by itself settle any later policy question.
One revision fact would break this read: if the October 15 reopening of the full time series after Blue Book and quarterly national accounts revisions materially lowers July services or monthly GDP, the apparent momentum would be smaller. One ONS-detail fact would also break the AI line: if the largest turnover reporters in the category were mostly cloud or consultancy businesses with no meaningful AI activity, the classification would be weaker. Until either fact appears in the record, the supported conclusion is narrower and sturdier: July beat the consensus, services carried the month, and the official AI-cloud wording is a signal inside the services ledger rather than a measured causal contribution. [E1]