A Boston-based consultancy has put forward a case for the artificial intelligence sector needing to reach $6 trillion in yearly income by 2031, so as to warrant the investment going toward data centres. Bain and Company makes that argument in its latest technology report series, which came out on Tuesday.
Within the next half-decade, the consulting firm expects revenue from new product development to become the largest contributor to the industry, with an estimated contribution of roughly $4.2 trillion toward funding AI’s global market. To achieve enterprise productivity gains, the figure stands at a need for approximately $1 trillion to $1.4 trillion in revenue. Consumer-focused services — subscriptions and advertising — are expected to contribute somewhere between $200 billion and $400 billion.
Crawford on the Economic Case
“The debate today is fixated on employee productivity,” said David Crawford, chairman of Bain’s global technology practice and lead author of the report. “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.”
According to Bain’s forecast, annual spending on AI infrastructure could reach $1.5 trillion by 2031. The total covers new facilities, increased capacity and upgrades to existing GPUs, memory and networking equipment. Assuming these capital expenditures represent roughly a quarter of industry revenue — “an ambitious but reasonable percentage based on trends among cloud providers” — maintaining this scale of investment would need an AI market of around $6 trillion per year.
The Prometheus Data Centre Example
Bain reported that the scale and expense of AI data centres are rising at a steady pace, with capacity roughly doubling every 12 to 16 months. A facility operated by Meta Platforms, the parent company behind Facebook, sits in Ohio. Its current capacity stands at 600MW, with an estimated cost of $24 billion in 2025. Epoch AI data suggests that figure could climb to as high as 2GW and $80 billion by 2027.
According to San Francisco-based Epoch AI, capacity at Prometheus is expected to hit 5GW by 2029, with a price tag of up to $175 billion. It would then grow to 9GW by 2030, at a cost of $200 billion.
| Year | Capacity (GW) | Cost Estimate |
|---|---|---|
| 2025 | — | $24 billion |
| 2027 | — | $80 billion |
| 2029 | — | $175 billion |
| 2030 | — | $200 billion |
Absorption Speed as a Competitive Variable
Absorption speed, defined as the pace at which companies can put AI to work, has become the “new competitive variable,” according to Bain. Leading AI labs are investing upwards of $9.75 billion in engineering models to help companies assimilate faster.
The report also notes that the industry’s future markets may include drug discovery, mental health and energy generation. New products and uses that don’t exist today will enable new markets and opportunities from abundant intelligence, Bain said.
The Challenges Ahead
Several obstacles must be overcome as data centre construction increases. Among them are expanding grid capacity for powering these facilities, securing GPUs and other infrastructure components, maintaining a qualified workforce, and keeping them “far above historical rates,” along with pressure from public opinion and regulation, including objections to resource use and noise pollution.
Bain noted that several governments are backing the expansion of the AI and data centres sectors, with the UAE, Saudi Arabia, the EU, South Korea and the US among those providing support. These facilities have become essential to technology innovation, economic growth and national independence, the firm said.
“Capital needs for data infrastructure will remain high … bottlenecks in power, semiconductors, and other inputs carry large capital needs of their own, opening additional entry points for investors,” Bain said. “And as sovereign infrastructure becomes a bigger part of national strategies, partnerships offer both a way in and geographic diversification.”
What the Numbers Mean
Forward-looking projections underpin the report far more than actual observed demand does. That $1.5 trillion figure for infrastructure spending falls into that category.
Epoch AI’s estimates show that a single data centre like Prometheus can start with a cost of just $24 billion and rise to $200 billion within a decade. The size of the increase is notable, even though the numbers are only estimates.
A Skeptical View
The core issue — whether sufficient economic value can be generated to warrant the investment — has yet to be resolved. What the report offers is a picture of what the sector must produce, rather than proof that there is enough demand today for the goods and services it envisions.
The authors are placing their wager on a surge of invention that eclipses what mobile and cloud unleashed. Whether it comes to pass remains uncertain.
The trend is unmistakable: a rising tide of data, a growing appetite for processing power, and an increasing price tag. What remains uncertain is whether income can keep pace with these expanding costs.
Source material: “AI needs $6T in annual revenue to justify data centre boom,” thenationalnews.com.
Get the Notebook.
The day's best stories and every fresh verdict, in plain English, in your inbox by seven. One email a day, no more.