The artificial intelligence sector would have to generate annual revenue of $6 trillion by 2031 to sustain the level of infrastructure investment being made in the industry, according to a new report from Bain & Co.

To put that figure in perspective, $6 trillion exceeds the annual gross domestic product of Germany, the world’s third-largest economy.

The infrastructure investment to support AI continues to grow, with capital expenditures possibly reaching $780 billion this year, nearly five times the level of just three years ago. This unprecedented build-out includes billions flowing into chips, data centers, networks and power systems.

Advertisement

Bain & Co. estimates that the size and cost of AI data centers are doubling every 12 to 16 months. The largest AI data centers today are approaching the capacity of a gigawatt of power. That is enough electricity to power 876,000 U.S. homes for a year, according to the Carbon Collective website.

By 2027, many data centers will likely approach the capacity for 2 gigawatts of power. By 2030 it is estimated that the largest AI campuses will cost about $200 billion each and have the power capacity of 9 gigawatts, which is enough power for nearly 8 million homes for a year.

While the focus has been on the challenge of building AI capacity, the real question may be whether enough economic value can be created to justify the investments currently being made, asserts the business consulting firm.

The report contends that by 2031, annual spending on AI infrastructure could reach $1.5 trillion. The total would include new data centers, expansion of the infrastructure and ongoing upgrades to the installed base of computer and networking equipment.

With the assumption that capital expenditures amount to about 25% of industry revenue, sustaining this level of investment would require an AI market of about $6 trillion annually. As astronomical as that figure may sound, some of that revenue is already taking shape.

Bringing innovation to industry

Bain & Co. estimates that consumer AI products, supported with subscriptions and advertising, could potentially generate an estimated $200 billion to $400 billion by 2031. At the same time, corporate adoption of AI could add $1 trillion to $1.4 trillion in revenue “as AI delivers meaningful productivity gains to enterprises across software development, sales, marketing, customer service and IT operations.”

The consumer and enterprise AI market could potentially account for $1.2 trillion to $1.8 trillion. That leaves about $4.2 trillion of new revenue that would need to be generated to meet the $6 trillion market goal that the consulting firm estimates would be necessary to fund the current and future build-out.

Some possible sources of new value include expanding search and advertising, which the report estimates could generate $100 billion to $200 billion in additional revenue. Developing AI to be an autonomous operator of automobiles, trucks, drones and other equipment could create a $400 billion market.

Another revenue source is referred to as “physical AI.” This category, which is estimated to be worth $900 billion, includes advanced AI models that can create realistic simulations and “digital twins” of physical processes, helping companies improve productivity, test modifications and accelerate research and development, possibly cutting costs by 10%. AI-powered robotics could be used in everything from mining and manufacturing to surgery, reducing manufacturing costs, increasing yields and speeding up construction.

Even if the report’s rosiest projections are used, that would still leave the AI market somewhere around $2.7 trillion short of the stated $6 trillion goal. Bain & Co. expects new applications in the future will help close that financial gap.

One possible area AI could be deployed is helping to develop new medical breakthroughs that will cure diseases and improve healthcare. AI could also perhaps accelerate progress in neuroscience and fusion energy or develop next-generation batteries and semiconductors. But more innovations and inventions will be required.

“The industry needs a wave of application innovation comparable with what mobile and cloud [technologies] unlocked, not just productivity gains on existing workflows,” Bain & Co. writes. “The infrastructure is being built ahead of the demand curve, and funding it sustainably will require adding approximately 1% to the annual global GDP growth rate. The question is whether the applications arrive in time to pay for it.”