Bain: AI needs $6 trillion in annual revenue by 2031 to justify data centre spending

Bain & Company has put a figure on what the AI buildout has to earn to pay for itself: roughly $6 trillion in annual revenue by 2031. According to the consultancy's latest technology report, as reported by The National, sustaining the current pace of data centre investment requires an AI market of that size, and much of the money would have to come from products that do not exist today.

The arithmetic behind the target is straightforward. Bain forecasts that annual spending on AI infrastructure — new facilities, added capacity and upgrades to the installed base of GPUs, memory and networking equipment — could reach about $1.5 trillion by 2031. If capital expenditure holds at roughly a quarter of industry revenue, a ratio Bain describes as ambitious but reasonable based on trends among cloud providers, the industry needs to be generating close to $6 trillion a year to support that level of investment.

Where the revenue is expected to come from

New product development is projected to carry most of the load, at about $4.2 trillion within the next half-decade. Bain's analysts place search, advertising, autonomy and physical AI in that bucket, and point to potential new markets in areas such as drug discovery, mental health and energy generation.

Enterprise productivity is expected to contribute $1 trillion to $1.4 trillion, covering gains in software development, sales, marketing, customer service and IT operations. Consumer-facing services such as subscriptions and advertising add a further $200 billion to $400 billion, a segment Bain considers crucial as providers push AI products to billions of users worldwide.

David Crawford, chairman of Bain's global technology practice and lead author of the report, argued the industry cannot rely on workplace efficiency alone. The debate, he said, is currently fixated on employee productivity, while the economics of AI infrastructure demand trillions in new revenue beyond productivity gains — what he described as a wave of innovation that will dwarf what mobile and cloud unlocked.

Bain also flags absorption speed — the pace at which companies can actually put AI to work — as what it calls the new competitive variable, and says leading AI labs are investing upwards of $9.75 billion in engineering models that help businesses adopt the technology faster.

Data centres doubling in size and cost

The report notes that the scale and cost of AI data centres has been doubling approximately every 12 to 16 months. Meta's Prometheus facility in Ohio illustrates the curve. According to data from research firm Epoch AI cited by The National, the site had a capacity of 600MW at an estimated cost of $24 billion in 2025, and is projected to reach 2GW and $80 billion by 2027, 5GW and up to $175 billion by 2029, and 9GW and $200 billion by 2030.

Money is not the only constraint. Bain lists grid capacity to power new facilities, GPU and component supply, hiring and retaining skilled workers at rates far above historical levels, and public pushback over resource use and noise pollution among the challenges the sector must manage. On the other side of the ledger, several governments — including those of the UAE, Saudi Arabia, the EU, South Korea and the US — are actively supporting data centre growth, framing the facilities as central to technology innovation, economic growth and national sovereignty. Bain adds that bottlenecks in power, semiconductors and other inputs carry large capital needs of their own, opening additional entry points for investors, while sovereign partnerships offer both access and geographic diversification.

Why it matters

The report reframes the central question of the AI investment cycle. It is no longer only whether data centres, chips and power systems can be built fast enough, but whether enough economic value can be created to justify what has already been committed. If the projected $4.2 trillion from new products fails to materialise, the funding model behind the buildout weakens, because enterprise productivity and consumer services combined fall well short of the $6 trillion mark. That gap also defines the industry's next competitive battleground: whoever invents the products that close it captures the returns needed to pay for the infrastructure the whole sector is currently constructing.