Could The AI Bubble Burst?
The artificial intelligence boom has attracted extraordinary levels of investment, transformed technology markets and created expectations of a new era of productivity, but a warning from the world’s leading central bank organisation suggests that the consequences could be felt far beyond the technology industry if financial returns fail to match the hype.
Why Are Economists Worried About An AI Bubble?
The Bank for International Settlements (BIS), often described as the central bank for central banks, has compared the current AI investment boom with previous periods of technological excitement, including Britain’s railway mania, the electrification boom of the 1920s and the dotcom era.
Its concern is not that AI itself lacks value, but that genuine technological breakthroughs can still attract far more investment than their eventual commercial returns justify. As the BIS explains, previous investment booms “all shared one common trait: a genuine technological breakthrough that attracted capital in excess of what commercial returns could ultimately justify.”
The distinction is quite important because the internet survived the dotcom crash and went on to change almost every part of business and society, yet the collapse of the investment bubble still caused investors to lose enormous sums, companies to disappear, investment to dry up and the wider economy to suffer. The concern now seems to be that AI could follow a similar path, with the technology itself continuing to develop and transform businesses even as excessive investment is unwound and the financial consequences spread far beyond the companies directly involved.
The Scale Of The AI Investment Race
The sums being invested in AI infrastructure are difficult to overstate. Major technology companies are spending hundreds of billions of dollars on data centres, chips, servers, power infrastructure and the enormous computing capacity needed to train and operate increasingly advanced AI systems.
The BIS has warned that competitive pressure could encourage companies to overcommit to projects whose eventual returns remain uncertain. It says disappointment could “trigger a sudden pullback in financing and turn the capex boom into a protracted investment bust with potential knock-on effects on the financial conditions.”
This is where the risk extends beyond Silicon Valley. For example, AI infrastructure involves a wide ecosystem of chipmakers, construction companies, engineering contractors, energy suppliers, data-centre operators, property developers and financial institutions.
If hyperscalers suddenly cut spending, the effects could travel through those supply chains. Companies that have expanded, borrowed money or recruited heavily in anticipation of continued AI investment could find themselves exposed.
Would An AI Crash Cause A Recession?
Nobody really knows conclusively whether there is an AI bubble or when any correction might happen. However, the BIS does seem to be quite concerned that the combination of high valuations, growing debt and complex private financing arrangements could make a reversal more damaging than a simple fall in technology shares.
A sharp fall in investment could, for example, affect construction, employment and business confidence. Falling technology valuations could hit pension funds and investment portfolios, while losses in private credit markets could make lenders more cautious.
For businesses outside the AI sector, this could mean tighter access to finance, delayed investment decisions and weaker demand from customers. The precise effect would depend on the scale of any correction, but the important point is that the AI economy is becoming connected to many more parts of the wider economy.
A Different AI Risk Is Emerging In Financial Markets
As if that doesn’t sound bad enough, there is another potential problem on the horizon. Sarah Breeden, Deputy Governor for Financial Stability at the Bank of England, has warned that autonomous AI agents could amplify financial instability.
The threat is based on the fact that AI systems trained in similar ways and pursuing similar objectives could respond to market stress at the same time. This means that rather than human traders reacting at different speeds and reaching different conclusions, large numbers of autonomous systems could actually sell, buy or withdraw simultaneously.
Breeden has warned that agentic AI could “amplify volatility in stress”, while regulators are considering whether measures such as circuit breakers or “kill switches” may eventually be needed.
This doesn’t mean AI will cause a financial crash, but it does mean that financial regulators are now treating AI risk as a serious financial stability issue rather than a theoretical technology debate.
The Productivity Question At The Heart Of The Boom
Ultimately, the sustainability of the AI boom depends on whether the technology produces enough economic value to justify the investment behind it.
There is some real evidence of benefits so far. For example, research published by the National Bureau of Economic Research found that generative AI assistance increased productivity in customer support work, with particularly strong gains among less experienced workers.
The difficulty is turning individual efficiency gains into sustained, organisation-wide financial returns. For example, businesses may save an employee 30 minutes on a task, but that does not automatically translate into higher profits, greater output or lower costs.
This gap between impressive demonstrations and measurable business value is one of the most important issues facing AI adoption.
What Does This Mean For Your Business?
For most businesses, the sensible response is neither to ignore AI nor to chase it at any cost. AI projects should really be treated like any other business investment. For example, start with a genuine problem, establish what success looks like, measure the result and avoid building critical processes around tools simply because they are fashionable.
Businesses should also understand their exposure. A company may not invest directly in AI shares, yet could depend on customers, suppliers or sectors whose fortunes are closely linked to the AI investment boom.
If the bubble does burst, AI itself will not disappear. The stronger technologies, suppliers and applications are likely to survive, while weaker business models and speculative projects are exposed.
For business leaders, perhaps the most useful lesson from previous technology booms is that transformative technology and financial excess can exist at the same time, which means the challenge is to benefit from the opportunities AI creates without becoming dangerously dependent on the investment boom surrounding it.
