Only 37% of Wall Street’s Money Managers Think AI Will Pay Off Within 2 Years — Half Are Still Bullish

Barron’s put two questions to 81 professional investors, and the answers point in different directions. Half are bullish on American stocks over the coming year. Yet only 37% believe today’s AI infrastructure spending will produce positive returns within two years. Another 32% say it won’t, and the rest say it is too early to know.

The bulls still expect the S&P 500 to gain approximately 8% through the end of 2027. Taken together, the numbers describe a market that trusts the economy more than it trusts the AI build-out propping it up. That matters for anyone holding index funds in a 401(k), because the companies spending the most on AI carry the largest weights in those funds.

October Already Showed the Split

The survey arrived after a month that tested its logic. The Invesco QQQ Trust (NASDAQ:QQQ), a proxy for mega-cap tech, fell 6.95% in October. The SPDR S&P 500 ETF Trust (NYSEARCA:SPY) slid just 1.08%. The broad market absorbed a sharp retreat in its biggest technology names.

Pressure on the AI buyers built in the final week. Bloomberg reported that Alphabet (NASDAQ:GOOGL | GOOGL Price Prediction) suffered a record market-value wipeout as AI ate into sales. Reuters found Big Tech is returning its smallest share of cash to shareholders in years because of AI spending. On Monday, Fortune reported on a major credit-rating agency. Citing the risks of the spending boom for Meta Platforms (NASDAQ:META), the agency had put a negative outlook on Meta.

A 5% Treasury Raises the Bar for Every Data Center

Respondents split equally on the biggest near-term risk: 27% named an AI slowdown and 27% named rising interest rates. Those are the same risk seen from two sides. The 10-year Treasury yield hit 5.31% on October 5 and still stood at 4.98% on October 29. The Federal Reserve’s target rate holds at an upper bound of 4.00%. When a government bond pays close to 5%, a data center funded with debt, or with cash that could have gone to buybacks, faces a far higher hurdle. Every month the return slips, the math worsens.

Suppliers Get Paid First, Buyers Wait

AI spending becomes supplier revenue fast. Chipmakers, networking firms and power-equipment makers receive it as soon as the purchase order is approved (we covered seven of these suppliers, from power to cooling, in a free report you can grab here). The hyperscalers doing the buying need years of cloud and software sales to recover it. Suppliers and the broader economy can keep growing. A professional investor can reasonably expect that while doubting Meta or Alphabet earns an adequate return on that capital soon.

That focuses the question for ordinary investors: can the S&P 500 keep rising if AI infrastructure disappoints, and October suggests it can tolerate a mega-cap slump, at least briefly. The buffer is thin, though. University of Michigan consumer sentiment fell to 48.1 in September, a reading in recessionary territory. If AI spending slows while consumers are already gloomy, little remains to carry earnings.

What to Watch Before Year-End

This week brings the midterm elections and a jobs report. The more telling signal is narrower: whether the Nasdaq-100 keeps lagging the S&P 500 while the 10-year yield holds near 5%, and if it does, the market is repricing AI buyers specifically, and the bulls’ 8% forecast rests on everything else. Also keep an eye on how Meta answers the rating agency. Trimming spending would validate the skeptics. Borrowing more to keep building would show the hyperscalers still prize the build-out over balance-sheet comfort. The VIX ended October at 19.42, near the top of its normal range, so traders are already pricing some of that uncertainty.

Contact [email protected] for any questions or corrections.