The market is splitting into two camps: 'AI' vs. 'non-AI'

The S&P 500 no longer reflects the state of the average U.S. stock.

For decades, the S&P 500 has been the preferred benchmark for monitoring the performance of the U.S. stock market. But these days, it is behaving less like a barometer of market health and more like a funhouse mirror.

Over the past few months, the performance of the premier U.S. equity index has grown increasingly disconnected from that of the average U.S.-traded stock.

The changing relationship became a big story in markets in September, so much so that it inspired Goldman Sachs derivatives strategist Brian Garrett to point out in commentary shared with clients over the weekend that the S&P 500 SPX was "no longer behaving like the clearing price for risk" in the U.S. market. To put it more plainly: The index is no longer representing the condition of the broader market - it barely reflects what is happening with its own members.

For example, the median S&P 500 stock was down roughly 17% from its 252-day high, according to a recent analysis from Warren Pies at 3Fourteen Research. But the S&P 500 finished Monday just 0.3% below its most recent record close in August. This largely reflects the growing influence of a handful of mega-capitalization companies, including the "Magnificent Seven," a group of some of the largest U.S.-publicly traded companies that saw big gains earlier in this bull market.

Taken together, members of the Magnificent Seven - which comprises Microsoft (MSFT), Apple (AAPL), Nvidia (NVDA), Alphabet (GOOGL), Meta Platforms (META), Amazon.com (AMZN) and Tesla (TSLA) - finished Monday with a market capitalization just shy of $25 trillion, - a new record high, according to Dow Jones Market Data.

Due in large part to the growing influence of artificial intelligence, the S&P 500 has lately seemed untethered from virtually every other risky asset, aside from itself. Whether it be other equity indexes or even credit spreads - another barometer of investors' risk appetite - America's favorite index is becoming increasingly unmoored from these longstanding historical relationships, Garrett said.

As of Monday, the rolling three-month correlation between the S&P 500, which is market-capitalization weighted, and the S&P 500 equal-weight index RSP was drifting back toward its lowest levels in recent memory, according to Dow Jones Market Data.

For investors who own fewer shares of the market's winners than the benchmark, weak breadth can of course be a problem. As for whether or not it ultimately leads to declines in the broader index, history doesn't offer a clear verdict. There is certainly salient evidence to the contrary: During the dot-com boom, the New York Stock Exchange advance-decline line peaked in 1998, data showed.

That didn't stop the S&P 500 and Nasdaq composite (COMP) from continuing higher for another two years, as the tech-driven bull market carried on.

Others, including Goldman's Garrett, believe the hollowing out of the market is a risk worth monitoring. That was his conclusion, influenced by conversations with people "in the trenches" and internal data showing hedge funds had cut back on their net leverage, suggesting that Wall Street was less inclined to be heavily long the market.

Even Ari Wald, head of technical analysis at Oppenheimer & Co., said in commentary published over the weekend that the collapse in the historical relationship between the S&P 500 and the average stock was an indication of "late-cycle behavior."

"It paints the picture of this later equity cycle environment, where correlations are lower, and we have these credit spreads widening out as well," Wald told MarketWatch.

Wald noted that the number of Russell 3000 stocks trading above their 200-day moving average had recently turned sharply lower. Technical analysts use the 200-day moving average to measure a stock or index's long-term trend. Patterns like this are typically seen only when major indexes like the S&P 500 and the Russell 3000 RUA are struggling. Instead, both were recently trading near record territory.

That might seem alarming, but Wald said it isn't necessarily a cause for concern - at least not right now.

As long as the AI trade continues, that should be enough to drive indexes like the S&P 500 and Nasdaq higher.

"Ideally, you want everything to be working. But at least right now the right things are working."

AI vs non-AI

The U.S. stock market isn't so much fracturing as it is splitting into two factions.

It's increasingly becoming a story of "AI" vs. "non-AI," said Julian Emanuel, chief equity and quantitative strategist at Evercore ISI. And not just in U.S. stocks; AI's impact on the investment landscape has started to seep into other markets like credit and emerging markets.

"Neither bonds nor EM equities are a diversifier from the dominant 'AI Revolution' theme," Emanuel said.

For months, Emanuel and his team have been tracking the number of stocks trading with "negative beta" to the S&P 500 - a number that has boomed lately. They found that number has recently hit its highest level on record.

A stock with negative beta routinely trades opposite the broader market. But rather than this being a risk, Emanuel said he sees opportunity. These stocks provide an inherent buffer for a diversified portfolio. When AI begins to wobble, investors have tended to shift money into other corners of the market, helping mitigate losses in the S&P 500. That helped the index avoid a larger monthly loss in July, even as the AI theme came under pressure.

"We expect the portfolio diversifying properties of negative beta stocks to remain meaningful into 2027."

Stock market breadth started to thin in mid-August, as Treasury yields pushed higher. However, it has made something of a comeback over the past two trading sessions. Of the 11 S&P 500 sectors, 10 finished higher on Monday for a second session in a row.

The last time at least 10 of the S&P 500's 11 sectors finished higher two days in a row was April 14, 2025. Back then, stocks were setting up for a historic comeback after nearly skidding into bear-market territory following the White House's initial "liberation day" tariff rollout.

Chelsea Ng and Michael DeStefano contributed

-Joseph Adinolfi