As AI-related investment becomes the dominant factor on Wall Street and in the US economy more questions are being raised about its viability and whether its very mode of expansion, as AI firms become increasingly dependent on each other, is creating a bubble which must inevitably collapse.

The impact of AI investment, particularly in the build-out of data centres, is not easy to directly measure but it is estimated to be more than one-third of the GDP growth of 2.2 percent reported in the second quarter.

Another major component is the expansion of consumer spending. But a significant portion of this increase comes from the upper portion of the income scale because of the increase in wealth resulting from the rise of the stock market boosted by AI and other high-tech stocks.

Summarising this situation, in remarks reported by the LA Times, Michael Pearce, chief US economist at Oxford Economics, said: “The economy is increasingly reliant on AI gains and the corresponding wealth effects boosting higher-income households’ spending power to fuel recent growth. The economy remains sensitive to a sudden reversal of optimism on AI.”

A recent study by economist Stijn van Nieuwerburgh, issued by the Brookings Institution, concluded that investment in data centres and AI-related infrastructure is projected to total $10.3 trillion from 2025 to 2032.

Reporting on these findings last month, the Wall Street Journal commented: “That is a staggering 3.6 percent of gross domestic product a year, on average. Never before has the US economy been so dependent on the build-out of a single industry.”

The article noted that while the investment was transforming every corner of the economy and “minting new billionaires,” it was also “creating significant risk, as much of it is built on debt” and an “abrupt slowdown could ignite shock waves throughout the US economy.”

The dependence of the economy on AI companies is mirrored in the stock market. The S&P 500 index reached a new record high this week. But it is being propelled by a small range of AI stocks and the trillions of dollars poured into its development. Many sectors of the market have not experienced a surge.

According to a report in the Financial Times (FT) Goldman Sachs has forecast that AI infrastructure stocks will contribute more than half of the growth in earnings per share of companies in the S&P 500 in the third quarter with just two of the chips group, Micron and Nvidia, to account for more than a third of the entire earnings growth of the index.

It noted that since the most recent low-point of the S&P 500 in mid-September only a third of stocks had risen while the other two-thirds had fallen further “amid fears of further US interest rate increases” following the decision by the Federal Reserve to lift its rate for the first time in three years at its meeting last month.

There is considerable commentary that the continued rise in market indexes, albeit at a slower rate, shows that the effects of rising interest rates, both by the Fed and in the bond markets, are being shaken off.

But according to Max Kettner, a strategist at HSBC, in comments to the FT: “Rate-sensitive stocks… have really suffered. There’s damage under the hood.”

Maija Veitmane, global head of equity strategy at the investment management firm State Street. said she expected “companies with weaker fundamentals to crack under higher interest rates” and that the sell-off of bonds, which has seen borrowing costs rise, had started to feed through to American companies.

So far, the AI companies have not been impacted but their financing has become increasingly reliant on debt amid concerns that their use of circular deals presents dangers to financial stability.

This prospect was raised in a report prepared by staff at the Bank for International Settlements earlier this month.

It noted that investments in AI had reached an “extraordinary scale” and that the spending of hundreds of billions of dollars by AI firms had become “macroeconomically significant in many economies.”

However, it continued, a “meaningful share of this capital flows through firms that serve simultaneously as suppliers of inputs and investors in firms that buy those inputs. When financing and commercial relationships overlap, a ‘circular’ investment structure emerges.”

The report identified three types of circularity: A tech firm that receives both financing and business from another tech firm; suppliers of tech products which finance customers to stimulate demand for them; and, flowing the other way, customers who finance suppliers to ensure they have access to their products.

Examining the activities of some 1246 AI firms between 2021 and 2025, the report’s authors found that more than half of incoming investments came from other AI firms and within that group half had some commercial relationship with the result that “a significant portion of financing and commercial relationships in AI are self-referencing.”

They warned that while circular arrangements may address contracting problems at the level of the firm, their highly intertwined and capital-intensive nature can “translate into macroeconomic risks” on top of those created by over-investment and debt financing of the AI boom.

Circularity also made it more difficult to assess the real profitability of an AI firm because “when a supplier finances a customer, part of the supplier’s revenue growth reflects its own capital investment, rather than organic final demand.”

This issue had arisen in the telecommunications boom of the late 1990s when part of the vendors’ sales growth was being funded by the vendors themselves so that when operators’ revenues failed to materialise or slowed, they could neither repay the loans nor sustain purchases of new equipment so that the vendors suffered loss of sales and financial losses.

“Such dynamics may play out in AI if revenue growth and end user demand fall short of firms’ expectations. Amid heightened AI-driven global equity valuations this could drive significant financial market volatility.”

Emphasising this risk, the report said that an investor that is also a supplier in a circular arrangement risks being hit twice because a shock to the customer reduces both the value of the investment and future product revenues.

Furthermore, because circular deals are concentrated in a small number of very large firms adverse shocks could be transmitted simultaneously through commercial and financial channels amplifying the risk of contagion, that is, the passage of a crisis well beyond its point of origin.

“These risks,” it concluded, “may be magnified by the growing use of private credit and special purpose vehicles [the activities of which are kept off the balance sheet] to finance AI infrastructure, which can create hidden leverage and interconnected exposures that amplify financial stress during downturns.”