Market Views: boom or burden?

19 October 2026

Jonathan Sparks,

UK Chief Investment Officer, HSBC Private Bank and Wealth

It is impossible to escape news of AI driving one of the largest investment booms in recent history. But without tangible returns, the money trees fuelling this bumper crop will not endlessly bear fruit.

While the jury – and the court of public opinion – is still out on whether the benefit of this new technology ultimately justifies the scale of expenditure, for investors the focus is on whether, and when, adequate returns will justify the cost.

Debt at the current scale inevitably attracts attention, but history suggests that high levels of borrowing alone are a poor guide to what happens next. More important are where that debt sits, who is providing the capital, how concentrated the investment has become and, above all, what returns it eventually produces.

That makes today’s AI boom less a story about an imminent debt bomb going off and more a test of whether capital is being deployed faster than the technology it is funding can deliver financially. 

There are good reasons not to overstate comparisons of today with previous crises, including the 2008 Great Financial Crisis. The broadest measure of US non-financial private corporate debt today is around 140% of GDP, well below the 171% reached in 2008. US Household debt – a primary catalyst of the GFC – has today substantially declined from 97% of GDP at its peak in 2008, to around 66% today. Banks today are also better capitalised, whereas in 2008 they were much more vulnerable to losses.

Those differences matter because they suggest today's vulnerability is not broad-based borrowing across households and businesses in the same way as during previous crises.

A more interesting concern is the concentration and circular financing of AI. Large amounts of capital are flowing into a relatively concentrated ecosystem, while the ultimate financial returns remain difficult to judge. Instead of questioning whether there is too much debt, investors should ask what it is financing and when the cash flow will appear. 

The current corporate earnings season will be an important test of whether AI-related spending is beginning to translate into profits. And this creates an intriguing tension. Cheaper open-source models could commoditise elements of AI and challenge some assumptions about future returns. Equally, this could accelerate adoption, making AI more accessible and expanding the overall market.

The debate today shouldn’t be reduced to just the size of the investment bill. A successful technology can still produce lacklustre returns that don’t justify the initial outlay. 

Markets are already wrestling with that uncertainty. Recent volatility has weakened investor sentiment and pushed some highly AI-exposed stocks sharply lower. I would also argue that markets have become too aggressive in pricing in assumed interest rate rises – lower bond yields could improve sentiment and ease borrowing costs, taking some of the heat out of the market.

That doesn’t eliminate any AI financing risks. But it does demonstrate why investors need to distinguish between an expensive investment theme and individual assets where expectations may have already significantly fallen.

So how to assess the opportunities today? Be disciplined and straightforward: look at who is borrowing, how much leverage is being taken on, and whether realistic cash flows can support spending.

AI may well justify the investment boom, but being right about the technology does not guarantee being right about every investment used to finance it. And just because the money tree may bear fruit, doesn’t mean you won’t pick a bad apple from time to time.