AI Investment Boom Could Mirror Dotcom Bust as BIS Warns of $1 Trillion Spending Spree

AI investment boom could turn to bust, warns BIS paper

This content has been selected, created and edited by the Finextra editorial team based upon its relevance and interest to our community. The five largest hyperscalers are set to spend more than $1 trillion combined on AI-related capital expenditure across 2025 and 2026, making the build-out one of the largest technology-driven investment booms in US history. In its recently published Annual Economic Report, the BIS raised concerns about the levels of investment. Now it has shown its workings in a new paper on the subject that raises the spectre of previous boom-bust cycles, from the canal and railway manias of the 19th century to the dotcom boom of the 1990s. The paper models the AI boom as a race between firms in a winner-take-most environment. Firms competing for a few dominant positions over-commit resources and the over-investment leaves the sector exposed to revenue disappointment that could turn boom into bust, warns the BIS, adding that, the larger the boom, the deeper the eventual bust. The race to commit early through debt and circular financing also makes a bust more likely. Calibrated to balance sheet and deal data, the model points to over-investment of around 1.5 times the efficient level, rising to around three times where demand is less elastic. And, a network analysis shows that stress in one firm could cascade to others through chains of financial exposures.

BIS Warns That AI Debt Could Turn Boom to Bust | PYMNTS.com

"The AI build-out ranks among the largest technology-driven investment booms in U.S. history," Phurichai Rungcharoenkitkul, an economist for the bank (BIS), wrote in a report published Tuesday (July 14). "Its scale, reliance on debt and circular equity ties raise questions about the boom's sustainability and financial stability." Rungcharoenkitkul added that the study examined a "dynamic contest" in which companies vying for a "few dominant positions" over-commit resources, leaving the industry exposed to "revenue disappointment" that could transform the AI boom into bust. "The larger the boom, the deeper the eventual bust. The race to commit early through debt and circular financing also makes a bust more likely," the study said. The report likened the boom in AI investments to other periods of heavy spending, such as the "canal mania" in the U.S. during the 1830s, a similar drive to build railroads in England a decade later, and the 1990s dotcom boom. All of those booms "ended in sharp corrections" that had far-reaching economic implications, with the scale and speed of AI investment indicating the impact this time could be larger. "The potential demand for AI services is clearly vast and could justify a substantial expansion in computational power," Rungcharoenkitkul wrote. "Yet relative to its pre-boom trough, the current buildout is on track to outgrow every previous episode only three years in. To the extent that contest motives also shape investment decisions this time around, the tendency toward excessive commitment raises questions about the sustainability of the current boom." Meanwhile, a report last week from Bloomberg News said that the five companies spending the most on AI data centers in the U.S. had doubled their debt load in the past five years to fund their efforts. In all, Alphabet, Amazon, Meta, Microsoft and Oracle added about $350 billion to their debt obligations, marking a change for the software industry, in which companies tend to bring in high margins without much routine capital spending. PYMNTS reported in January that focus had moved from how much Big Tech was spending on AI to whether those investments were paying off in terms of durable growth and profitability. "The sharp increase in capital expenditures this year, largely driven by investments in data centers, chips and AI infrastructure, intensified scrutiny of margins, cash flow and the pace at which AI-driven products can be monetized," PYMNTS wrote.