AI’s “Exuberance” Problem: $3Tn Spend, $1.5Tn in Risky Debt

A wave of “debt-fueled exuberance” is defining the $3 trillion AI datacenter boom, and analysts are warning it “will backfire.” The problem is a $1.5tn funding gap that is being filled by “private credit,” a “shadow banking” sector that has “raised the alarm” at the Bank of England.
This $1.5tn in “speculative” funding is being used for “unproven” projects, often “without their own customers.” Analysts warn that lenders, caught in the AI hype, are “improperly assessing the risks” of funding “very quickly depreciating assets.” A hedge fund founder calculated these assets depreciate twice as fast as revenue, a recipe for default.
This “exuberance” is happening despite mounting evidence that the business case for AI is not yet proven. An MIT study found 95% of corporate AI pilots are yielding “zero return,” directly contradicting the “lofty revenue expectations” justifying the $3tn spend.
Even industry watchdogs are skeptical. The Uptime Institute, which rates datacenters, stated “many” of the projects announced “with a fanfare” are “speculative” and “will never be built.”
While “hyperscalers” (Google, Microsoft) are spending their own cash, this $1.5tn influx of “speculative” debt “could end up representing structural risk to the overall global economy” if, or when, the “exuberance” fades and the bills come due.