TL;DR

Britain’s output rose 0.4% in July, and one corner of the services sector did a disproportionate share of the work. Computer programming and consultancy expanded 3.5% on the month, adding 0.12 percentage points to GDP, and the ONS says AI and cloud businesses feature heavily among its top earners. It is the third strong reading in a row for that industry, though the statistics office is explicit that it cannot put a number on AI’s contribution.

The headline numbers

July followed 0.3% growth in June and a flat May. Measured over three months, output was also 0.4% higher, making eight rolling periods of expansion in succession. Services did the lifting, while production and construction each shrank 0.5% across the quarter.

Information and communication grew 2.4% in July alone. On the ONS’s rounded figures, its programming component accounts for roughly 30% of the month’s entire GDP growth. Across three months that activity was up 4.4%, and scientific research and development climbed 7%. Manufacturers of computers, electronic and optical goods had a strong month too, rising 5.2%.

What the ONS will and will not say

The bulletin chooses its words carefully. It notes that businesses involved in AI and cloud work are prominent among those reporting the biggest turnover in programming and information services, then cautions that its survey design makes their precise effect difficult to isolate. Businesses also credited hot weather and the World Cup for some of July’s strength.

That caution is worth respecting. Last month’s second-quarter figures showed programming output up 3.7%, following 3.8% in the first quarter. A third strong print makes coincidence harder to argue, but turnover at AI-linked firms is not the same thing as productivity gains among the businesses buying their services.

Looking forward

These are early estimates. Survey response rates for July ranged from 80% to about 87% depending on sector, and the entire series opens for Blue Book revision on 15 October.

The ONS’s own business survey also shows the pressure beneath the headline: in late August, 59% of firms expressed some degree of worry about energy prices. An AI build-out that leans on power-hungry data centres will not escape that constraint.