The central bank of central banks draws uncomfortable parallels between today's AI investment boom and the dot-com bubble, flagging over $1 trillion in hyperscaler spending as a potential pressure point.
The Bank for International Settlements just dropped its Annual Economic Report 2026, and buried inside the usual central-banker prose is a warning that should make crypto and traditional investors alike sit up straighter. The BIS argues that if confidence in AI investments sours, the fallout won’t stay contained to tech stocks. It could cascade into corporate credit markets, choke off financing for smaller firms, and trigger a downturn that moves faster than anything we’ve seen before.
The reason it might move faster? A massive chunk of AI-related debt is flowing through private credit channels and hedge funds, not traditional banks. Those less-regulated pipelines are great when money is abundant and sentiment is bullish. They become a problem when the music stops.
A trillion dollars and a lot of leverage #
Here’s the scale of what the BIS is looking at. The five largest hyperscalers are projected to spend over $1 trillion on AI-focused capital expenditure during 2025 and 2026 combined. That’s not revenue. That’s spending, much of it financed with debt rather than internal cash flows.
The BIS report identifies AI investment sustainability as one of four major “pressure points” for the global economy. Credit spreads for certain AI-related issuers have already started to widen, which is the bond market’s quiet way of saying “we’re getting nervous.”
BIS regional representative Zhang Tao warned that the extensive use of hedge funds and private credit in AI financing could lead to a faster unwinding of market values during any downturn.
The report draws explicit parallels to the dot-com bubble and even the 19th-century railway boom, both periods when transformative technology attracted capital far in excess of near-term returns.
The dot-com parallel cuts both ways #
The BIS comparison to the dot-com era is instructive, but it’s worth noting what happened after that bubble popped. The internet didn’t go away. Amazon survived. Google launched. The technology was real. The problem was that too much capital arrived too early and with too much leverage.
The same logic probably applies to AI. The technology is genuinely transformative. But when you have over $1 trillion flowing into capital expenditure over a two-year window, funded significantly by debt, the math has to eventually work. Disappointing returns on that spending wouldn’t just hurt tech stocks. The BIS warns it could exacerbate inflationary pressures and fiscal strains across the broader economy.
What to watch from here #
The widening of credit spreads for AI-related issuers is the canary in the coal mine. If those spreads continue to blow out, it signals that fixed-income investors are demanding higher compensation for holding AI-related debt, which makes new borrowing more expensive and starts a negative feedback loop.
Watch the private credit space closely. Unlike public bond markets, private credit doesn’t have real-time pricing transparency. The BIS is essentially saying the plumbing of AI financing looks uncomfortably similar to the kinds of structures that amplified previous crises.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our