Forget risk on/risk off. AI on/AI off is now driving the stock market.
Once upon a time, the market was guided by whether investors wanted to take risk or not. The idea was that on some days, traders would become more optimistic about the economy and earnings and buy all kinds of stocks and sell their stodgy bonds. On others, they'd guiltily sell the speculative stuff and return to the homey comforts of bonds, gold, and the like. Risk on. Risk off.
You don't hear that phrase much anymore, because markets no longer work that way. Rocky Fishman at Asym Research points out that of the 11 sectors that make up the S&P 500, five have a zero or negative correlation with the index over the past five months, including healthcare and real estate. Think about that for a second. If I told you the State Street Health Care Select Sector SPDR exchange-traded fund is up 1.2% this week -- it is -- you wouldn't be any closer to knowing that the S&P 500 is falling 0.2%, which it is. (The Dow Jones Industrial Average, which often follows its own course, is also down 0.2% this week.)
It's the same story at the single-stock level. Jonathan Krinsky of BTIG observes that the average stock in the S&P 500 moved in a different direction than the index on 52 of the 135 days so far this year. By comparison, more stocks rose on index-down days or fell on index-up days only 24 times in the first 135 trading days of 2025. The comparable stat never got above 15 in the 2010s.
This trend of more dispersion within the equity market has plenty of interesting implications for the options wonks among us. (In fact, simultaneously buying options on individual stocks and selling options on the index has become a favorite hedge fund trade.) It also explains why the S&P 500 can move so little on days that feel head-spinningly volatile: When half the crew is rowing in the wrong direction, your boat goes nowhere.
And it means risk on/risk off, which once seemed to explain everything, now explains almost nothing. It doesn't take Hercule Poirot to figure out what killed it.
Today, calling artificial intelligence a market theme would be like calling the economy a market theme. Five AI heavyweights -- Nvidia, Microsoft, Amazon.com, Alphabet, and Meta Platforms -- make up about a quarter of the S&P 500. It's not just a U.S. story, either -- just look at the Korean or Taiwanese markets.
That's not to say that investors are waking up every morning and buying any vaguely AI stock they can get their hands on -- that's so 2025. These days, flows change almost daily, as investors feel better or worse about AI. In late June, for instance, chip stocks absolutely tanked, and on not particularly much news. At the same time, software stocks, which have seen their valuations get crushed in an era of AI vibe-coding, rallied. You could call it a momentum reversal, or an intra-tech rotation. But it was really an AI-off move in a year dominated by the AI-on trade. The Nasdaq Composite is on track to close down 1.6% this week, but is up 8.1% for the year.
The questions about artificial intelligence that drove the June selloff remain. Barclays this week highlighted survey data showing that more American adults use AI for personal applications than for work, while less than 14% use it daily on the job -- a figure that's essentially flat over the past two years.
The Barclays economics team puts a positive spin on it, saying that "the evidence is consistent with an economy that remains in the early stages of the adoption cycle." But ChatGPT was released nearly four years ago, and while AI has gotten better since then, the advances haven't been revolutionary. Nor are they expected to be so in the future. For all the talk about creating "artificial general intelligence," or an AI model that beats humans at every task, the goal actually seems much further off now than it did in the halcyon days of 2022.
Meanwhile, the jury is out on how useful AI will be for companies. The Barclays team observes that "industries adopting AI more rapidly do not appear to be experiencing faster productivity growth," which would surely have shocked most 2022-era AI bulls -- as well as the companies spending a gazillion dollars to bring AI tools into the workplace.
If AI proves to be a consumer product only, the AI trade is toast.
Those worries permeated the stock market as the Magnificent Seven started reporting earnings. On Thursday, Alphabet and Tesla shares slumped on concerns about the ever-increasing sums they were spending on AI. If these questions become more pressing as Microsoft and Amazon report this coming week, there are sure to be a few beneficiaries of the AI-off trade -- companies such as Adobe, Cognizant Technology Solutions, and Apple -- that will be ready to rally.
The bigger fear is that the AI trade, which has been running powerfully for years, permanently fades. If that happens, it may be time to resurrect risk-off.
Write to Alex Rosenberg at alex.rosenberg@barrons.com
This content was created by Barron's, which is operated by Dow Jones & Co. Barron's is published independently from Dow Jones Newswires and The Wall Street Journal.
(END) Dow Jones Newswires
July 24, 2026 11:41 ET (15:41 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.
Comments