The magic of AI is wearing off for investors
The AI fever in markets is breaking. US stocks are on track for one of their worst years in two decades and suddenly the whole topic is on big investors’ lists of “things to worry about” for 2027 rather than a reason to be endlessly cheerful.
It is hard to say what exactly has spoilt the mood, but you can trace it back from the memory chip crashin Asian markets over the summer. Since then, a wide range of important groups of stakeholders have said “no”. Corporate bond investors are sick of being forcefed often questionable tech debt. The general public in the US is sick of supersized data centres in their neighbourhoods. Fund managers are sick of, let’s say, flaky corporate governance standards at Nvidia, which would matter less were it not for the fact that in October this company hit a peak market capitalisation of about $5.7tn — more than twice the size of the entire FTSE 100 index in the UK.
And now this week’s spectacular stock market listing of AI lab Anthropic feels less like a moment of triumph (aside from the chosen few now counting their billions) and more like a last hurrah. Fizzy market debuts are a classic sign of a cyclical top in markets, and a $1.5tn market capitalisation at listing certainly fits the fizzy description. It is time to move into a new gear.
“The market has been obsessed by the core AI topic and in 2027, the focus will shift,” said Vincent Mortier, chief investment officer at Amundi in Paris. “Monetisation will be difficult. It will be harder to justify higher levels of valuations. We are a little bit cautious and we prefer companies that are using AI — in robotics, in pharma, and some very boring industrial companies.”
The many AI slip-ups of 2026 have delivered some of the worst days in the US stock market of the entire year — including a 2 per cent drop in the index on Anthropic’s first day — and now the S&P 500 benchmark index is up a measly 5 per cent in 2026. Without a serious pick-up in the next few weeks, this will be one of the worst years for the index in two decades — only six years have churned out poorer performances, and some of those were in periods of outright crisis.
All of this feels weirdly unfamiliar. One of the defining characteristics of the US stock market since the pandemic has been its extraordinary resilience. Come what may — war, tariffs, more war, you name it — the market has mostly ground higher. But again, the bombastic AI trade goes a long way to explaining that, and not just because tech stocks drive the bus for the overall market.
A lot of it has to do with the technodollar — the newfangled high-tech version of the petrodollar. Think back to the old-economy version of this, where oil is priced in dollars and the proceeds of its sale flood exporting nations, especially in the Middle East. These states churned those proceeds into US government bonds, in part because it’s the biggest, baddest bond market and nothing else is deep and liquid enough to accommodate those billions. The same is true of the chips trade, but this time in Asia.
Chip exporters such as South Korea and Taiwan have been flooded with money from US hyperscalers constructing vast data centres. Deutsche Bank noted this week that Korean semiconductor exports have jumped 200 per cent this year. Its trade surplus just in chips stands at almost $250bn. Indeed, for the first time ever, wrote Mallika Sachdeva at the bank, the global trade in chips had exceeded global trade in oil and gas.
But this time, she said, the riches were being churned not into bonds but into US stocks. This is likely for the same reason that the petrodollar pushed into bonds — the US makes up 64 per cent of the global stock market, based on weights in the MSCI all-country world index. Nothing else is big and liquid enough to accommodate the billions.
So the enormous AI infrastructure build-out pumps money into US stocks, which pump money into AI expansion, which pumps money into US stocks. On and on it goes, until the music stops.
“There is an element of circularity to be conscious of,” wrote Sachdeva at Deutsche Bank. “This could introduce instability in global flows should demand be reassessed.”
Well, analysts are wondering whether that reassessment is indeed coming, not necessarily immediately but in the next year or so. In a presentation this week, Jan Hatzius, chief economist at Goldman Sachs, said he was “optimistic on the economic pay-off to AI” but that he did see “potential downside risk”.
“Some of it is related to the fact that we are building a large capital stock and that large capital stock . . . eventually it’s going to be built out, and then investment stops growing and may start to come down,” he said. “We don’t expect that to happen in 2027 . . . But that is, of course, subject to questions.”
You don’t have to be an AI doomster to see the cracks forming here. But all really exciting disruptive technology goes through phases of innovation, then whizz-bang expansion, then excess, then a sober sorting into winners and losers, and commoditisation. For investors, the magic is wearing off.
Comments
Post a Comment