Britain's economy grew 0.4% in July against a City of London forecast of zero, and the largest single contribution came from computer programming. The Office for National Statistics went further than that in its commentary: many of the companies with the highest revenue in July 2026 were tied to AI and cloud computing. This is a national accounts release naming the AI build-out as the thing keeping a G7 economy growing through an energy shock, and it lands six weeks before a chancellor has to write a budget around it.
The month followed 0.3% growth in June. Services carried it, expanding 0.4%, with the sharpest gains in administrative services, programming and computer consultancy. Industrial production rose 0.2%, where higher manufacturing output offset declines in mining and in electricity and gas supply. Over the three months to July, which the ONS treats as the more reliable window, GDP also grew 0.4% — the same pace as the three months to June.
Martin Beck, chief economist at WPI Strategy, said that against weak activity across many traditional sectors, this kind of productivity-raising spending is what the British economy needs.
The backdrop is a war with Iran that pushed energy prices up and left interest rates higher than anyone expected at the start of the year. The economy has held up anyway. Through the first half, with the conflict already under way, the UK was the fastest-growing economy in the G7.
For Chancellor John Healey, preparing his first budget for 28 October, the resilience is useful and probably not sufficient. Economists warn the longer view is worse: oil well above $100 a barrel could push inflation up globally and drag borrowing costs with it, and the shock the Iran war delivered to bond markets may leave Healey choosing between higher taxes and lower spending.
Neither the oil price nor the growth beat is expected to move the Bank of England at next week's meeting. Suren Thiru, chief economist at the accountancy body ICAEW, said the data could strengthen the hawkish camp among rate-setters, while noting that a September rise remains unlikely — most of them expect a slowing economy to bring inflation down eventually, even with US–Iran tensions rising.
Here is what I would not take from this print. Beck's framing describes the spending, not its effect. What the ONS measured is firms buying AI and cloud services, which registers as output in the sector selling them; whether any of it raises productivity in the sectors doing the buying is a question the July data cannot answer and does not claim to. Calling it productivity-raising is a forecast wearing the clothes of a statistic. And the narrowness is its own signal — when one line of the services sector is the largest contributor and traditional industry is flat, a strong month is a concentrated month.
The July figures also contain a reminder to discount them. Liz McKeown, the ONS director of economic statistics, said that as in June, some firms reported effects from warm weather and the FIFA World Cup, helping some sectors and hindering others. Nobody has quantified that, and nobody has separated the durable part of the AI spending from a one-off procurement wave. Those are the two numbers that would tell you whether 0.4% is a trend, and neither exists.
Which leaves Healey building a budget on the strongest quarter Britain has had in a hostile year, powered in large part by companies buying compute. Capital expenditure cycles are not the same thing as demand, and they end on a schedule the Treasury does not control.