AI may be one of the most significant economic transformations of the coming decades, boosting company productivity and opening up new markets. Even so, investors can lose money in companies connected to it.

These two ideas may seem contradictory, but they are not.

That was the point Bank of England Governor Andrew Bailey emphasized when he warned that markets could face shocks linked to the AI investment boom. Bailey does not deny the technology’s potential; he believes it could boost growth. But he cautions against assuming that every company benefiting from today’s AI wave will be a winner in the future.

A successful technology does not guarantee a successful stock

A stock’s price depends not only on how good a company is today, but also on how much profit investors expect it to make in the future.

That is where the problem arises when expectations are too high.

If the market expects a company to double its profits for years, its share price may rise to a level where merely delivering good results is not enough; the company must deliver results better than the already lofty expectations.

That is why an entire industry can succeed while some of the companies in it see their share prices fall.

Something similar happened during earlier stages of the internet’s development. The technology itself truly changed the world, but many companies once considered leaders of the wave later disappeared or lost their standing.

So there are three different things:

The success of the technology, the success of the company, and the return investors get for the price they paid.

The first can happen without the second or third.

When does a boom become a financial risk?

The risk grows when investment is no longer funded solely by companies’ own money, but increasingly relies on debt.

The Bank of England noted that global debt issuance linked to AI reached about $450 billion by early September 2026—more than double the 2025 total—according to estimates the Bank attributed to Morgan Stanley. Analysts also expect a large share of future data-centre spending to be financed through debt markets and private credit.

Debt changes the equation.

If a company buys billions of dollars’ worth of equipment with its own money and returns fall short of expectations, the company bears an investment loss.

But if it borrows to fund the project, it still has to service the debt even if revenue falls short of expectations.

At that point, the problem can spread beyond the price of a single stock to banks, bonds, credit funds, and the investors who financed these projects.

Why do valuations matter?

The higher an asset’s price rises on the basis of distant future expectations, the more sensitive it becomes to any change in those expectations.

If investors believe AI will rapidly boost productivity, they may accept high valuations for companies that stand to benefit from it.

But if adoption takes longer, revenue is lower, or the costs of energy, data centres, and financing rise, investors may reassess what these companies are worth.

That does not mean the industry will collapse.

It simply means that the old price assumed a more optimistic future than the one the market now expects.

What concerns the central bank?

The central bank is not concerned only when an investor loses money on a tech stock.

Its concern begins when risks become interconnected enough to affect the financial system.

Today’s AI spending is linked to tech companies, chipmakers, data centres, and energy, as well as bond markets, private credit, and banks.

That is why the Bank of England’s Financial Policy Committee warned that the growth in AI-related financing is expanding the number of investors and markets exposed to a sudden reassessment of expectations.

Real success comes after the spending

The most important question in the years ahead will not simply be:

How many billions of dollars were spent on AI?

It will be:

How much economic value did those billions generate?

If data centres, chips, and models deliver higher productivity and new revenue, they may justify much of today’s optimism.

But if valuations and debt outpace companies’ ability to generate returns, we could see a correction even as AI itself continues to transform the economy.

That is the paradox: the technology may truly be the next revolution, while some investors may have paid too much, too soon, to be part of it.