THE ECONOMIST: Markets churning fast beneath surface as AI prompts investors to reassess business models

An investor waking from a stupor that began on New Year’s Eve might question whether they had missed anything. The S&P 500 share index of big American firms sits almost exactly where it did at the end of 2025: nearly at a record high. There has been no shortage of geopolitical drama, but little sign of it in the trajectory of the world’s most watched index.

Beneath this calm surface, however, the churn in America’s financial markets has been furious. A panic about what artificial intelligence will do to business models has prompted software firms’ stock prices to tumble: they are more than 30 per cent below a recent peak last year. And it is not just software companies that investors worry will be disrupted by the rise of AI agents and “vibe coding”.

Waves of share-price volatility have rippled through sectors as varied as trucking and commercial real estate.

So far index investors’ losses have been offset by the gains from a gaggle of winners. Just as some sectors have been slammed, investors have clamoured for “HALO” stocks (heavy assets, low obsolescence). Energy and commodities firms have benefited, as have sturdy utilities and companies selling consumer staples.

Within the tech sector, those making the hardware that powers AI have boomed. The share price of Sandisk, which designs and manufactures memory chips, has more than doubled since the start of the year.

Much more reassessment lies ahead, since investors still know little about how AI will eventually reshape businesses. And there is no guarantee that the fragile balance between winners and losers will continue to hold.

The stockmarket’s dispersion—expected differences in movements of individual stocks, measured by the price of options — is higher today than it has been for 98 per cent of the time over the past decade.

This measure tends to jump when investors are rapidly pricing in new economic realities, which usually entails the headline share-price index plunging or soaring. Previous instances include the sharp falls of March 2020, as the COVID lockdowns began, and the tariff-driven panic of last April.

Investors’ reassessment is evident in debt markets, too. In the middle of last year risky “junk” bonds issued by American technology firms yielded some three percentage points more than Treasury bonds. That was about the same spread as for junk bonds more broadly, regardless of the issuer’s sector.

Today investors are demanding more compensation from tech borrowers, such that the bonds yield five percentage points more than Treasuries — a similar spread to that during last year’s tariff panic. The rest of the market has barely budged.

The consequences in markets for leveraged loans and private credit, which are heavily exposed to tech and business-services firms, are likely to be more severe. Matthew Mish of UBS, an investment bank, thinks defaults on such debt could rise by 2.5 and 4 percentage points, respectively, by late 2026. In a scenario where disruption is more aggressive still, the increases in defaults would be twice as large. And private-credit investors are already taking fright.

Many big funds faced outflows late last year. On February 18, Blue Owl permanently restricted redemptions from its private-credit fund aimed at retail investors, to give it time to sell assets. The share prices of several listed private-investment giants, such as Ares, Blackstone and KKR, have tumbled.

It seems unlikely that things will calm down any time soon. Novel applications of AI are bound to emerge, meaning investors will have to continually update their understanding of which business models will be disrupted and how severely. And if AI turns out to be far less transformative than many now believe, that would hammer the share prices of some of the world’s biggest firms, such as Nvidia. Roiling markets could in turn hinder the development of AI.

“If losses spike too quickly and to sufficiently high levels in loan markets, the tightening in credit and financial conditions could be severe,” wrote Mr Mish in a recent research note. If firms find it harder to borrow, they will build the infrastructure needed for AI more slowly.

Investors with broad stockmarket exposure, meanwhile, might see their luck run out. The firms that make the most money from AI could end up being unlisted ones, such as Anthropic or OpenAI. Moreover, no rule specifies that every losing sector must be perfectly balanced by a winning one. With investors reassessing every business model under the sun, do not bet on the stockmarket staying calm on the surface.