BIS Chair: AI Boom Poses New Risks to Financial Stability
nashnova research
BIS Chair Pablo Hernandez de Cos warned that global AI investment has reached the trillion-dollar scale — increasingly funded by opaque debt rather than earnings — and could trigger systemic fragility if corporate profits fall short.
How much money is pouring into AI?
BIS estimates the five largest tech firms will spend over $1 trillion on AI in 2025–2026.
Industry forecasts go further: global AI investment may swell from roughly $500 billion today to $4 trillion by 2030.
This means → AI is no longer just a technology story; its investment footprint is large enough to move the global economy.
Where is the money coming from — and why is that the real risk?
The core concern: AI expansion is increasingly financed by debt and private credit, not by corporate earnings.
These funding sources are "opaque and interconnected" — outsiders cannot easily see who owes what, or how large the exposures are.
In plain terms = the money is borrowed, not earned, and the borrowing channels are not transparent. If profits disappoint, repayment pressure spreads along chains no one can fully trace.
Has this script played out before?
Hernandez de Cos drew parallels with the railway expansion era and the dot-com bubble.
He stressed he is not predicting the same ending, but said "the scale and speed of the current investment boom, and the high expectations for commercial returns, do warrant vigilance."
This reflects a deeper regulatory worry — not about AI technology itself, but about the fragility created when high valuations + market concentration + opaque financing stack on top of each other.
What does this mean for central banks and jobs?
AI does not change central banks' monetary-policy mandate, but it affects demand, supply, and financial markets simultaneously, making the economic picture harder to read.
On employment, mass displacement has not materialized yet, but early signs of substitution are visible in customer service, coding, and administrative roles. Reskilling is becoming more urgent.
On productivity, generative AI boosts efficiency by 10%–65% in specific tasks such as coding and consulting — but whether that translates into economy-wide productivity growth remains an open question.
Can this boom last — and what decides?
BIS laid out a clear test: the pace at which corporate profits materialize + how fast financing transparency improves.
In plain terms = if companies actually earn money from AI and borrowing becomes more visible, the boom can continue. Otherwise, risk accumulates.
Current estimates suggest AI could lift total-factor productivity growth by roughly 0.5 percentage points per year — whether that materializes is the foundation the entire thesis rests on.
市场有风险，内容仅供研究参考，不构成投资建议。