Earnings Season Kicks Off With Record Concentration
TREE NEWS reports: Third-quarter earnings season for U.S. equities begins in earnest next week, with S&P 500 earnings per share projected to grow 27% year-over-year — the fastest pace since 2021. But that headline number masks an unusually narrow base: just two companies, memory-chip maker Micron and AI chip leader Nvidia, are expected to contribute roughly one-third of the index’s total earnings growth, while the median S&P 500 constituent is forecast to grow EPS by only 9%, down from 14% in the second quarter.
Micron reported results on September 30, beating expectations with adjusted EPS of $33.42 on revenue of $54.2 billion and an 87% gross margin. Its guidance for next quarter revenue of $60–63 billion also topped consensus. Yet the stock barely moved, because the market had already priced in even stronger growth. The debate has shifted from “how good can it get” to “how long can this momentum last.”
Market Implications: Index Strength, Narrow Breadth
The divergence between the index and the average stock is now extreme. The S&P 500 sits just 2% below its all-time high, yet the median stock is down 17% from its own peak. The average realized correlation among S&P 500 constituents has fallen to 0.06 — the lowest in a decade of data — meaning index performance increasingly hinges on a handful of AI and energy leaders. The top 10 stocks now account for about 40% of index market cap and 37% of expected earnings.
Bank of America’s Michael Hartnett notes that roughly 400 S&P 500 members have fallen below their 50-day moving averages, and about 300 below their 200-day. His “AI Big 10” — the Magnificent Seven plus Broadcom, AMD and Micron — now represents 42% of index market cap, exceeding the concentration seen during the “Nifty Fifty,” the TMT bubble, and the 1920s. Only Japan’s 44% and U.S. railroad stocks’ 63% in 1881 were higher.
Morgan Stanley’s Mike Wilson warns that the divergence between an index near record highs and a large cohort of stocks in bear-market drawdowns “will converge one way or another,” with bond-market volatility likely serving as the ultimate arbiter.
Sector and Thematic Divergence
AI infrastructure is the dominant earnings engine. Excluding hyperscalers, AI infrastructure is expected to contribute 54% of Q3 earnings growth, with hyperscalers adding another 19% — together about three-quarters of the total. Energy, buoyed by higher oil prices amid Strait of Hormuz tensions, is forecast to grow EPS 109%, while information technology grows 64%. Together they account for roughly 80% of quarterly earnings growth. By contrast, consumer discretionary earnings are expected to be flat, and consumer staples are forecast to contract.
Analysts have also trimmed margin expectations for the median S&P 500 company by 11 basis points since the quarter began, with downgrades across every sector except IT.
Hyperscaler Capex: The Hidden Support
Behind the AI earnings boom is relentless hyperscaler capital spending, expected to rise 116% year-over-year in Q3, up from 87% in Q2. Cloud revenue growth among Amazon, Google, Microsoft and Oracle accelerated to 48% in Q2 and is projected to reach 55% in Q3, with combined cloud backlog near $1.7 trillion. Historical patterns suggest consensus capex forecasts consistently lag actual growth — a dynamic that could keep upward pressure on 2027 estimates.
However, this capex surge comes with rising financing needs, much of it debt-funded. Two accounting tailwinds also flattered first-half earnings: mark-to-market gains on private AI equity stakes (about $150 billion in Q2, equal to 12% of S&P 500 EPS) and $69 billion in tariff refunds from the U.S. Treasury (about 6% of pre-tax profits). Neither is expected to repeat at the same scale in Q3.
Key Takeaways for Investors
- Concentration risk is real: With the top 10 stocks driving 68% of earnings growth, any disappointment in Nvidia’s November report or Micron’s follow-through could trigger outsized index moves.
- Breadth is historically narrow: The gap between index highs and median stock drawdowns is unsustainable; watch bond volatility as a potential catalyst for convergence.
- AI capex remains the key variable: Hyperscaler spending and cloud backlogs support the earnings narrative, but debt financing and accounting one-offs warrant scrutiny.
- Consensus may still be too conservative: JPMorgan sees Q3 growth potentially exceeding 29%, and Goldman Sachs targets S&P 500 EPS of $375 in 2026 and 8,000 by year-end 2026.
