AI faces $6 trillion test to justify data centers, Bain says

Bain & Co. estimates the global AI industry must generate $6 trillion in annual revenue by 2031 to justify the massive investments in data centers worldwide. While existing AI services could contribute up to $1.8 trillion, a further $4.2 trillion in new revenue will be needed, according to the firm's annual global technology report released on Tuesday.

The shortfall is likely to stem from emerging areas like autonomous machines, robotics, drug discovery, mental health, and energy generation. Bain's lead author, David Crawford, emphasized the need for "a wave of innovation that will dwarf what mobile and cloud unlocked" to meet this demand. The report warns that AI infrastructure is being developed ahead of the demand curve, and sustaining this growth could require adding approximately 1% to the annual global GDP growth rate.

Major tech companies including Microsoft, Google, Amazon, Meta, and Oracle are investing billions in data centers to keep pace with AI's computational demands. Data center sizes and costs are doubling roughly every 12 to 16 months, driven by rising chip prices from Nvidia and SK Hynix, as well as increased networking equipment and other components.

The report arrives amid growing concerns about the uncertain returns on AI investments. Critics argue that the interconnected web of dependencies between technology manufacturers and AI developers is fueling overly optimistic expectations that necessitate even larger financial commitments. While current discussions often focus on AI's potential to boost employee productivity, Bain argues that the economics of AI infrastructure will require trillions in new revenue beyond productivity gains.

The consultancy projects data-center spending to reach $5 trillion to $6.5 trillion by 2030, adding at least 150 gigawatts of capacity, which will put additional strain on countries' energy resources. Annual spending on AI infrastructure, encompassing data centers, computing capacity, and upgrades in accelerators and memory chips, could hit $1.5 trillion by 2031.

Data center developers are already facing shortages in transformers, water, and power supplies, and have faced significant opposition, with $68 billion worth of projects blocked or delayed in the U.S. in the June quarter alone.