Digital investment nearly doubles since 2019 yet AI’s growth contributions questioned

Since 2019, investment in digital infrastructure in the UK has nearly doubled, according to official data. However, economists are questioning whether the recent gains in productivity are truly attributed to artificial intelligence. The Office for National Statistics (ONS) revised how digital infrastructure spending is calculated, suggesting a surge in investment from £6.3 billion in 2019 to around £11.6 billion in 2024.

The updated estimate now includes data centres, software, cables, and other digital infrastructure elements, aligning the UK with standards in other major economies.

Building, software, and database investments accounted for about 77.6% of digital infrastructure spending between 2020 and 2025. By 2025, investments in assets like data centres and fibre optic cables were estimated to reach £5.6 billion, marking a 94% increase since the start of the pandemic. Additionally, around 171 data centre construction projects tracked by Barbour ABI could potentially quadruple current data centre investment levels by 2030.

The ONS revised its calculations as part of an effort to better track AI-related investment. Surveys indicated that more than a third of businesses with 10 or more employees adopted AI in 2026, boosting digital infrastructure investment. Yet, leading City economists have warned that AI has not been the driving force behind higher growth despite optimism that the UK's productivity would surpass previous trends.

Analysts from the Resolution Foundation, Morgan Stanley, and the London School of Economics have separately argued for a "productivity boom" in the UK, based on payroll data showing a 1.1% annual growth rate since late 2024, compared to the ONS's estimate of 0.2%.

However, AI's growth contributions to productivity remain a subject of debate. Morgan Stanley researchers indicated that private sector productivity in the second quarter of the year stood at 1.8% over a 12-month period, while Pantheon Macroeconomics warned of "little evidence" that AI was making workers produce more output per hour.

Surveys showed that AI adoption is increasing, but economists Rob Wood and Elliott Jordan-Doak noted that AI has not translated into "material headcount reductions." While most businesses surveyed reported that AI had "no material impact" on headcount over the past three years, the proportion of firms cutting jobs due to AI remained low and stable.

The economists believed that significant changes were occurring beneath the surface for some workers, although recent productivity gains might not signal the start of a sustained trend.