The artificial intelligence industry will need to generate about $6 trillion in annual revenue by 2031 to justify the enormous capital being poured into data centers, according to a new report from the consultancy Bain & Company published Tuesday.
The calculation is simple and unforgiving. Bain forecasts that annual spending on AI infrastructure — new facilities, expanded capacity, and upgrades to the installed base of GPUs, memory and networking equipment — could hit $1.5 trillion by 2031. If capital expenditures consume roughly a quarter of industry revenue, which Bain calls "an ambitious but reasonable percentage based on trends among cloud providers," the AI market would need to approach $6 trillion a year to sustain the buildout. For more context on this story, see our ongoing AI trends.
"The unprecedented speed and scale of the AI buildout, with billions flowing into chips, data centres, networks and power systems, have focused attention on the challenge of building capacity," the report said. "But the more important question may be whether enough economic value can be created to justify it."
Where $6 Trillion a Year Would Come From
Bain's breakdown of that hypothetical revenue pool is a map of where the industry still needs to invent new business. The single largest piece — an estimated $4.2 trillion — would have to come from revenue on products that largely do not exist yet: new offerings in search, advertising, autonomy and physical AI.
Enterprise productivity, the segment where most AI monetization happens today, would need to grow to between $1 trillion and $1.4 trillion in annual revenue to support gains in software development, sales, marketing, customer service and IT operations.
Consumer-focused services — subscriptions and advertising on AI products — are projected to contribute between $200 billion and $400 billion. That is the segment the industry has been pushing hardest, as providers race to put AI-powered products in front of billions of users, and even its optimistic ceiling falls far short of what the infrastructure demands.
"New products and uses that don't exist today will enable new markets and opportunities from abundant intelligence — these may include drug discovery, mental health and energy generation," Bain said.
'A Wave of Innovation That Will Dwarf Mobile and Cloud'
David Crawford, chairman of Bain's global technology practice and the report's lead author, said the debate over AI's economics has been too narrowly focused.
"The debate today is fixated on employee productivity," Crawford said. "The economics of AI infrastructure demand trillions in new revenue beyond productivity gains. What the industry needs is a wave of innovation that will dwarf what mobile and cloud unlocked."
The report also highlights a new metric that Bain says has become the industry's competitive variable: absorption speed — the pace at which companies can actually put AI to work. Bain notes that leading AI labs are investing upwards of $9.75 billion in engineering models specifically designed to help enterprise customers assimilate the technology faster.
Data Centers Doubling in Cost Every 12 to 16 Months
The scale of the infrastructure race behind these numbers is accelerating on its own curve. Bain found that the size and cost of AI data centers is doubling approximately every 12 to 16 months.
The report cites Epoch AI data on Meta's Prometheus data center in Ohio as an example. The facility currently has a capacity of 600 megawatts at an estimated cost of $24 billion in 2025. By 2027, that is projected to jump to as much as 2 gigawatts and $80 billion. By 2029, capacity is expected to reach 5 gigawatts at a cost of up to $175 billion, and by 2030 the site would balloon to 9 gigawatts at roughly $200 billion.
Those cost curves explain why the revenue question is becoming urgent, and why Bain argues that the industry cannot simply outbuild its way to sustainability. Every doubling of capacity raises the revenue bar that future products must clear. The report also flags a series of constraints that grow harder as the buildout continues: securing grid capacity to power the facilities, procuring GPUs and other infrastructure components, and finding the skilled workforce to run them.
The Question Hanging Over the Boom
The report arrives at a delicate moment for the sector. Hyperscalers, sovereign wealth funds and private credit lenders have all escalated their exposure to AI infrastructure, and Wall Street analysts have spent much of 2026 debating whether revenue growth can keep pace with capital expenditure.
Bain's answer is not that the buildout is doomed, but that it is unfinished: the industry has proven it can spend at unprecedented scale, and now has roughly five years to invent the products, markets and consumer habits that would make $6 trillion in annual revenue real.
If that innovation wave materializes, the infrastructure bet pays for itself. If it does not, Bain's math implies a painful correction — trillions of dollars of data centers chasing a market that never grew into them.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
