S&P 500 Breadth Weakens as AI Stocks Drive Record Highs

A record that few stocks are sharing

The S&P 500 closed above 7,800 for the first time this week, and the Nasdaq also set a record close on the same session 2. A closer look shows how narrow the rally has become. According to Dow Jones Market Data, fewer than half of the index's members finished that record day above their 200-day moving average, a common measure of whether a stock is in a longer-term uptrend 2. In practice, the average stock is not taking part in the advance that the headline index suggests.

The gains have been concentrated in companies tied to artificial intelligence. Many other stocks have been held back by higher interest rates and stubborn inflation 2. Dan Russo, investment chief at Potomac Fund Management, told The Wall Street Journal that the large-cap companies with very strong balance sheets are what is holding the market up. He compared the pattern of tech titans rising while most stocks lag to the dynamics of the dot-com era 2. The rally has also moved ahead even as the bond market has been sending warning signs 2.

Concentration at historic extremes

Several separate measures point to the same imbalance. Bank of America estimates that an "AI Big 10" of Apple, Microsoft, Nvidia, Amazon, Alphabet, Meta, Tesla, Broadcom, AMD and Micron makes up close to 40% of the S&P 500. Other ways of measuring concentration put the share of the largest companies even higher 1. Data from Creative Planning shows that Nvidia, Apple and Microsoft alone account for more than 21% of the index. That is the highest three-stock concentration on record. The nearest comparison is the mid-1980s, when IBM, AT&T and ExxonMobil peaked at a combined 13.4% 3.

An FXEmpire analysis argues that concentration at these levels resembles conditions before some of history's worst crashes. It identifies 7,002 on the S&P 500 as a level where a break lower could show that investors have overpaid for future growth 1. That figure is a technical marker, not a forecast. It sits roughly 10% below the current record, which gives a sense of how far the market could fall before a correction would confirm the bubble argument.

The earnings case for the bulls

The strongest counterargument is that profits are concentrated too, not just prices. Goldman Sachs expects S&P 500 earnings per share to grow 27% year over year in the third quarter, the fastest pace since 2021 4. Micron is projected to contribute 19% of that growth and Nvidia 15%, so these two semiconductor companies would account for about a third of the total. The ten largest contributors together would account for 68% 4.

Analysts have revised their expectations in the same direction. Barclays strategist Venu Krishna found that about 40% of this year's upward revisions to 2026 S&P 500 earnings estimates came from semiconductor and hardware companies. For 2027, that share rises to roughly 75% 3. In other words, Wall Street is betting that the AI hardware cycle will extend well into next year.

The same data also shows how weak the rest of the market is. Goldman expects median S&P 500 earnings growth of only 9% in the third quarter, down from 14% in the second quarter 4. The typical company is slowing while a small group speeds up. This is the earnings version of the breadth problem visible in the 200-day average data.

Reading the signals

All four analyses agree on the main facts. Prices, profits and estimate revisions are increasingly concentrated in a small group of AI and chip companies. Where they differ is in how worried to be. The FXEmpire and Potomac views lean toward historical parallels with bubbles 12. The Goldman and Barclays figures suggest the rally is backed by real, fast-growing earnings 34.

On balance, this does not look like a profitless mania along the lines of 1999. Nvidia and Micron are expected to deliver real earnings growth. The main risk is how much depends on them. When a third of the index's quarterly earnings growth and three-quarters of next year's estimate upgrades rest on one industry, a single disappointment could hit the whole market. Possible triggers include a cut in data-center spending, signs of a memory-chip glut, or cautious company guidance. The Yahoo Finance coverage describes this risk as a potential market "air pocket" 3.

Third-quarter earnings season begins next week, and it will be the near-term test 4. If the chipmakers deliver, the concentration may persist. Lasting improvement in breadth, however, probably depends on lower rates and a pickup in earnings for the median company, not more records from the same ten stocks. Until then, investors should treat the index's new highs as a reflection of a handful of companies rather than the market as a whole.