According to Bain & Company, the AI market will need to generate nearly $6 trillion in annual revenue by 2031 to fund current growth and the expansion of data centers . Consumer and business applications combined will generate no more than $1.8 trillion. The rest will have to come from new markets, though it is unclear which ones those will be.
This is what is claimed in Bain & Company’s Technology Report 2026, which was released this week. Annual spending on AI infrastructure could reach $1.5 trillion by 2031, including upgrades to GPUs, memory, and networking equipment. If capital expenditures (capex) account for about 25 percent of revenue, the AI market will need to generate nearly $6 trillion annually.
The hyperscalers, Microsoft, Google, Amazon, Meta, and Oracle, alone could collectively spend $780 billion this year. That’s nearly five times as much as three years ago. We’ve already seen this reflected in this year’s quarterly results: AWS spent $44.2 billion on capex in the first quarter, 76 percent more than a year earlier. Alphabet is targeting $180 to $190 billion this year. The data centers themselves are also getting bigger: from 1 gigawatt today to 9 GW campuses by the end of the decade.
A $4.2 trillion shortfall
To put it mildly, there is a gap in the projected budget. According to Bain, consumer products will generate $200 to 400 billion in 2031 through subscriptions and advertising; this assumes that these subscription models will continue to exist. Business adoption will bring providers an additional $1 to 1.4 trillion. This will mostly involve API pricing, which has so far been higher than the cost of the same access via a subscription. Combined, that amounts to no more than 1.8 trillion, leaving a shortfall of about 4.2 trillion.
According to Bain, that gap must be filled by four new sources. Ads in chatbots and AI search account for $100 to 200 billion. Autonomous vehicles, drones, and industrial automation represent another $400 billion. The greatest gains can be made with physical AI. This includes simulations, digital twins, and robotics, such as humanoid robots. This category yields 900 billion, assuming a 10 percent reduction in R&D and production costs. New applications such as AI-driven drug development and materials research are not assigned a specific value. The picture is still unclear and does not provide a solid foundation for trillions in revenue within five years, although the growth figures for AI revenue since 2023 have been impressive.
Not the only calculation
Other parties have reached conclusions similar to those of Bain & Company. Earlier this month, Schroders calculated that every $1,000 billion in annual data center investments should generate about $1,200 billion in annual value for customers. This is despite the fact that investments are often made in hardware that depreciates rapidly, especially when a new generation of chips is released. Reuters outlines the risks, ranging from cost overruns to power shortages and “stranded capital”, which could all hinder AI expansion.
According to Bain, the infrastructure is being built before the demand exists. To finance this sustainably, an additional percentage point of global GDP growth per year is needed. Whether the applications will be ready in time to foot the bill remains to be seen, according to the consulting firm.