Just how big is $6 trillion? AI v world's largest economies

Artificial intelligence will need to bring in $6 trillion a year by 2031 to pay for the data centres and chips being built to run it. That's more than the entire economy of Germany, Japan or the UK today.

In Bain and Co.'s annual Technology Report, the consultancy estimates that global spending on AI infrastructure could reach $1.5tn a year by 2031. That covers new data centres and computing capacity, plus constant upgrades to chips, memory and networking. If capital spending runs at about a quarter of revenue, as it does for cloud providers, the industry needs $6tn in annual sales to fund it.

Existing AI services won't come close. Consumer subscriptions, advertising and enterprise tools such as coding assistants could together bring in $1.2tn –$1.8tn. That leaves a gap of about $4.2tn that Bain says must come from products that barely exist yet: autonomous machines, factory robotics, and new breakthroughs in drug discovery and materials science.

David Crawford, the report's lead author, said the industry needs “a wave of innovation that will dwarf what mobile and cloud unlocked”.

Put against the world's largest economies, the scale of that target is extraordinary. On forecasts from the International Monetary Fund (IMF) for 2026, only China and the US are bigger.

The comparison has limits. Revenue is not the same as GDP, which measures value added. And by 2031, the International Monetary Fund (IMF) expects India and Germany to have exceeded $6tn, which would put AI's target fifth rather than third.

Still, the $6 trillion figure today highlights how the infrastructure being built is way behind the demand and just how large it sits among the biggest economies in the world.